Qhimm.com Forums
Project forums => Team Avalanche => Topic started by: Mayo Master on 2012-10-22 07:00:05
-
Hi everyone,
With Ulpian having made a thread for the remake of sector 6, I'm going for Sector 5, given that's where I've made my first field screen, and that's where I'm going to make the next.
Anyway, I wanted to post about my latest progress, even though Ulpian's work makes me feel very humble :-[
(http://imageshack.us/a/img513/2764/myfirstlily.png) (http://imageshack.us/photo/my-images/513/myfirstlily.png/)
It's only the beginning of a very long process. Thankfully, playing around with lattices in Blender allows for getting variations of the flower position fairly easily. I'm aware the pic can be improved (petal textures, stigma and stamen to be resized, etc.) though I'll end up with a whole population of these.
Some random facts I learned while making this flower:
- the flower is essentially based on lilium longiflorum, a variety of lilies originating from Japan. It's also called Easter lily. I've read this flower was a symbol of purity and resurrection.
- oddly enough (basing my source on Advent Children), even though the flower bud is a lily, the stalk and the leaves are not. Real lily leaves grow in successive whorls, and have straight parallel veins, while the flower in AC are crenate, growing in an alternate fashion.
- there's a whole market dedicated to moulds of lily petal veins, which are meant for shaping sugar crusts for pastries :?
Let me know if you have some advice, even if it seems very early on.
-
Hi everyone,
...even though Ulpian's work makes me feel very humble :-[
Haha! Maybe you could share any knowledge to reach perfection together ;) Nice work on the plant too...
-
Well thanks for the props MM :) no reason to feel humble in any case, I think you're doing a pretty good job yourself.
So what field scene are you going for exactly? Loved the trivia on the Easter lily. The model could, IMHO, benefit from some of the following improvements:
- A more oval-shaped and overall curvier form for the petals. The curved tip here looks fine, but the rest of the petal seems a bit too straight and rectangular-shaped.
- I believe petals and leaves are a great opportunity to use material translucence. Doing so might help making them less plastic-looking.
(I'm basing those suggestions on this picture:)
(http://www.mobot.org/gardeninghelp/images/low/thmb/W210-0901020.jpg)
The rest (especially the pot and dirt) seems fairly impeccable to me.
-
Thanks! I came across the same reference picture during my search, this one is particularly good. I also took some screenshots of Advent Children.
Concerning the shape of the petal (and leaves), with the lattices I'll be able to finely tune how they're bent. I realize that in my image, the petals are folded along a bell shape (a bit like a tulip), while the real deal is more trumpet-shaped. Concerning the petals themselves, I might have to refine that indeed. At first I wanted to deal with shapes by making a UV transparency map using a real petal, but it didn't work that well (namely the edges of the petal were not smooth - like an aliasing problem). Looking at your picture closely, it also seems like the petals do not all have the same shape : one triplet (with 120 degrees angle between each) is larger, and the second triplet looks more narrow. I'll verify that on other reference pictures. I'm also a bit worried about how much subsurf I can get to it, since I'll have to make a whole population of those (I'll be working on eals_1) and my computer is an old, unoptimized laptop.
Otherwise, I did read that translucency was a key factor in making plants look realistic, not plastic (though I haven't investigated a lot on that yet). So I did crank up the translucency, but it didn't seem to make much of a difference. I wonder if the light source itself has also a part to play in it, and I haven't bothered with that yet.
-
Hey guys,
I'm having some problem with my Blender modelling, I don't know if anyone can help me out. And I've not seen anyone around encountering this kind of issue!
It's very basic really. I started to make a picket fence, but when rendering, I end with really weird glitching texture/shadows on the pickets. I made the pickets completely plain (in terms of texture) to try and locate the problem. The odd part: if I take a camera up close, the pickets look completely normal.
I also noticed the weird glitching shadows are more apparent once I turn "ray shadow" on, but for that scene, I can't live without ray shadows.
Here are the pickets up close:
(http://imageshack.us/a/img703/3811/wiptestfenceclose.th.png) (http://imageshack.us/photo/my-images/703/wiptestfenceclose.png/)
And here they are, glitching with my render camera.
(http://imageshack.us/a/img198/6048/wiptestfencefar.th.png) (http://imageshack.us/photo/my-images/198/wiptestfencefar.png/)
Any ideas? :|
PS: on my lighting settings, cranking up the "samples" or the softness of shadows doesn't do a thing.
-
What shaders are you using?
-
What shaders are you using?
For "debugging purposes", I switched to a very simple material (all textures removed). I'm at work right now so I can't access my Blender files and tell you exactly, but it must be something along "diffuse white intensity 0.9", "specularity white intensity 0.1", no translucency, no reflectivity. Simplifying the material (or even changing to another material) didn't seem to have any effect on these glitched shadows.
-
This kind of thing is usually caused by polygons being too close to eachother, try making the pickets thicker
-
This kind of thing is usually caused by polygons being too close to eachother, try making the pickets thicker
Thanks, I will try that when I'm back home. Though I'm a bit surprised, because my pickets - although thin - have non-negligible thickness (as far as I recall, the thickness to height ratio should be around 1:20). Verdict in 6 hours :P
-
yes, but when highly zoomed out it can mess up.
-
yes, but when highly zoomed out it can mess up.
Okies. I just read something about z-fighting which concurs. It was mentioned that z-fighting was emphasized when clipping distances where far apart, which is definitely the case in my scene. I will try to tune that too (I can definitely increase the distance of front clipping).
UPDATE: Bingo! :) It was indeed an issue of z-fighting. I modified the clipping distances of the camera so that it would only encompass the visible area, and it worked. No need to retouch the picket thickness.
-
Ok guys, I have an interesting issue to deal with. I guess you could call that one a problem of consistency.
Here's the thing: I'm currently working on eals_1:
(http://img713.imageshack.us/img713/3069/sgarden.png) (http://imageshack.us/photo/my-images/713/sgarden.png/)
Now, the issue is about the house itself. I thought, for accuracy's sake, that it would be best to refer to the field screens of the house in order to model it. The one below gives an accurate picture of the layout:
(http://img11.imageshack.us/img11/9052/groundfloor1.png) (http://imageshack.us/photo/my-images/11/groundfloor1.png/)
Now, if I try to replicate this layout on eals_1, I end up with something which is really off (the model is the wireframe, overlaying the reference image):
(http://imageshack.us/a/img99/8911/wipealshouseissue.png)
As you can see the section on the flank of the house (the kitchen, actually) is really misaligned.
So... is there a guideline about this kind of problem? There are different approaches:
- Draw the outside regardless of the inside layout, even if it means a distorted geometry, relying only on the original reference image.
- Revise the drawing to account for consistency between the inside and the outside of the house.
I'm partial to the second option, but I know the first approach has been somewhat prevalent in this project so far. Some advice (particularly from the mods) would be much appreciated.
I also think this kind of issue is bound to happen on numerous occasions. The most glaring example will be on the modelling of the church (the outside drawing is completely inconsistent in the original game, the versions of AC and CC are much better in that respect).
-
Keep the distorted geometry, just have em look like the originals.
-
This is just a graphical upgrade. Fixing inconsistent world geometry would require altered walkmeshes, which are outside the scope of Team Avalanche.
-
Keep the distorted geometry, just have em look like the originals.
Got it.
This is just a graphical upgrade. Fixing inconsistent world geometry would require altered walkmeshes, which are outside the scope of Team Avalanche.
Indeed the walkmesh would have to be altered if I had to fix the geometry in this particular case, so I understand the guideline. :)
-
if entering the house simply loads a new area (the inside of the house), then the outside and inside models don't in any way need to be the same mesh.
so it makes sense to have one for the inside, and another for the outside, making sure they all look like the originals.
-
I know that the inside and the outside don't need to be on the same mesh, since they're loaded in different environments. It may be a personal taste, but I greatly appreciate when different aspects of the same object are rendered consistently. I just feel weird about having to make a non-euclidian house while there's no story-wise reason for defying the laws of physics. Besides:
- the inside of the house has a nice regular geometry which is architecturally sound.
- the outside is actually quite distorted, and from this viewpoint I wouldn't like living in there. Frankly, on that one, it's not 90's mangas which are inspirational, but the works of E.C. Escher!
- Having to align several objects with each having distorted geometry is a real pain, and modelling with messed up face normals is not fun. I'm slow at modelling because I'm a noob, but this time I'm being extra-slow.
Anyway, I'm almost done with it (textures notwithstanding), currently into modelling the inclined hexagonal tower in the back. I feel like I may build another model of the house "as it should be" and store it somewhere, for my own sake :P
-
Those 'defying laws of physics' geometry is pretty consistent throughout FF7
and is part of its style. Just look at the world map, chibi towns and forests.
In field, chibi field models and regular sized furniture.
-
Those 'defying laws of physics' geometry is pretty consistent throughout FF7
and is part of its style. Just look at the world map, chibi towns and forests.
In field, chibi field models and regular sized furniture.
I believe the inconsistency I'm mentioning about the house belongs to the category of errors and mistakes that were in no way intentional, or part of an "art style". Back then, the guys had eventually to release a game and were bound by the timeframe of the project, so if minor errors were made then "who cares?" (in particular, many of these errors are absolutely not obvious. Who noticed the house was distorted before I did?). You can actually find a bunch of these throughout the game (I already found some in my previous work on "the man in the pipe" scene), the following list reports a few of those in FMVs namely. http://www.ff7citadel.com/secrets/s_errors.shtml (http://www.ff7citadel.com/secrets/s_errors.shtml)
Sincerely one could have ensured inside/outside physical consistency while keeping the same art style.
-
...You can actually find a bunch of these throughout the game (I already found some in my previous work on "the man in the pipe" scene), the following list reports a few of those in FMVs namely. http://www.ff7citadel.com/secrets/s_errors.shtml (http://www.ff7citadel.com/secrets/s_errors.shtml)
Sincerely one could have ensured inside/outside physical consistency while keeping the same art style.
lol, how I know that feeling, just the starting field MDSTIN has already some issues, notice the stairs at the end of the stage where you go up left...
It is definitely a challenge to find the correct solution for those, I'd say, try it out and see what the difference is, if it is not noticeable, go with it.
And with noticeable I mean, it's going to be a pain when the walkmesh has to be altered, I believe no-one tried doing that...yet.
but of course, this is just my two cents.
8)
-
Posting my progress on eals_1, as I think I'm done with the modelling of the house (and many other things) - textures notwithstanding. Hope you guys enjoy it (click for high-res).
(http://img16.imageshack.us/img16/8846/wipeals1.png) (http://imageshack.us/photo/my-images/16/wipeals1.png/)
If the house looks like the original and does not seem distorted, it means I've done my job well. I hope you'll find it that way :P
-
house Looks great, just a little comment on the cliff.. it seems a little unnatural.. don't know how I should explain this.. the lines are kinda too straight and regular giving this feeling like I'm looking at a man-made structure rather than a natural cliff
-
I think that after it is textured it will not be that visible. I think it's looking great right now. If it looks unnatural after texturing though just try and move vertices here and there and I think you should be fine.
-
yeah, I guess it's cause of lack of textures, this looks great can't wait till it's completed, good job!
-
Just to clarify the status of this scene for you guys:
- In terms of adding objects, I have one pillar in the background to make, and a small one in the foregrounds, all the way to the left.
- What remains to be modified: the water, the ground, the cliffs, and making populations of grass and flowers. Grass and flower models are almost done. The cliffs will be done using displacement shaders, it's going to look cool. However, for doing that, I'll have to subsurf the cliffs like grazy, and I don't know if it will lead my computer to agony.
- I also have to add the waterfalls, which will be done by fluid simulation. Again this is going to be very computationally intensive.
In a way, I'd like to make all the stuff that does not put too much of a strain on the computer first, and do the cliffs and the waterfalls last.
-
I'd do the water via matte painting. just get some appropriately shape waterfall/water resources and shoop it up at the appropriate resolution. Fluid dynamics will was yer render time.
-
I'd do the water via matte painting. just get some appropriately shape waterfall/water resources and shoop it up at the appropriate resolution. Fluid dynamics will was yer render time.
I do want to try to make it with Fuild Dynamics. My computer may not like it, but I do! You can call it a professional quirk :D
And if it just renders slowly, I'll lauch the renderings before going to bed :P
If it doesn't seem technically feasible, then I'll consider matte painting.
-
its also a matter of getting the fluid to simulate the way you like it with the right lighting and glow. If there were animated frames to it, i'd suggest doing this, but for a single frame, the fastest and most accurate (to original) results will come from a matte painting.
-
I've been doing quite a bit of texturing, so it's time for a progress update.
So, here's a side-by-side comparison, at the original resolution.
(http://img844.imageshack.us/img844/773/wipeals.png) (http://imageshack.us/photo/my-images/844/wipeals.png/)
... and below is a close-up on the house, whose textures are practically done (click for high-res).
(http://img145.imageshack.us/img145/3802/wipealshouse.png) (http://imageshack.us/photo/my-images/145/wipealshouse.png/)
I hope you'll enjoy.
-
Nice work [thumbs up] :)
note that the light rays are on a sprite layer
(http://img338.imageshack.us/img338/9919/eals1.png) (http://imageshack.us/photo/my-images/338/eals1.png/)
8)
-
Very nice
-
Nice work [thumbs up] :)
note that the light rays are on a sprite layer
I didn't know that! I guess I wasn't working with the proper reference image... I had made an effort to cast shadows on the scene to bring contrast to the rays (I set a shadow from the plate, plus soft shadows from clouds), probably I'll have to change that.
However, how would you recommend to render the rays? Do I have to make a separate HD render of them on their own?
Very nice
Thanks!
-
However, how would you recommend to render the rays? Do I have to make a separate HD render of them on their own?
Thanks!
I think a light layer in photoshop might do the trick, or if Blender is able to, you could render out different passes.
some info (http://www.blender.org/development/release-logs/blender-243/render-passes/)
Actually, there are more layers (in this scene the light-rays are animated by flipping the layers) others have objects you walk behind.
you can see it here:
(http://a.imageshack.us/img691/7692/bg2w.png)
8)
-
Rendering different passes should be totally doable in Blender. However, I'm not at all familiar with compositing and working with nodes. I have put books about it on my Xmas wishlist, so hopefully I should be able to do all that later on :D
The necessity of rendering several layers (namely for objects you walk behind) wasn't in my mind, so thanks for reminding me that. As a matter of fact, there are many objects in my scene which are concerned by this: the picket fence, flower patches, the water tank... Probably I'll need some guidance as to what should be rendered on which layer to be featured in the game. But I don't think I'm at that point yet.
-
To find out what the layers are, you can use Palmer and extract all the different layers.
What I did to create the following scene:
(http://img607.imageshack.us/img607/66/mds7pb2sample.png)
is to render the layer with object ID's (which I've put in manually) which gave me this result:
(http://imageshack.us/a/img850/3671/mds7pb22048025objectmxs.png)
that makes it easier to select the different parts, and replicate the layers from Palmer in photoshop.
Apart from that, it's always possible to render just the different parts, but the catch there is the shadows and reflections you might mis.
8)
-
To find out what the layers are, you can use Palmer and extract all the different layers.
The problem is that I don't own a PC version of FFVII, I only have a Playstation version. Probably I'll need to have someone do that for me :-[
Apart from that, it's always possible to render just the different parts, but the catch there is the shadows and reflections you might mis.
Shadows wouldn't be a problem. I could save another blender file where I'd set the materials as "cast shadows only" of the relevant objects (I did that already for my "man in the pipe" scene, where the whole pipe is modelled but half of it is "cast shadows only"). Reflections may be trickier to deal with.
-
The problem is that I don't own a PC version of FFVII, I only have a Playstation version. Probably I'll need to have someone do that for me :-[
For the psx version you can use 7mimic to see the layers, but I don't think you can extract them seperately with 7 mimic.
8)
[edit]
But don't worry about the layers for now, modeling everything is still the way to go for now, just look at the layers in 7mimic so you'll know which parts you should model, the rest will come when the scenes are finished.
8-)
-
Okay, I've spent some time on texturing the cliffs of my scene, and I think it's time for a new update... (click for high-res)
(http://img141.imageshack.us/img141/3751/wipeals2.png) (http://imageshack.us/photo/my-images/141/wipeals2.png/)
I hope you enjoy
Basically I would have to texture the ground, the water, make populations of flowers, and make a waterfall.
However, I'm currently encountering some technical difficulties... well, mostly, I'm hitting the maximum computing capacities of my poor old laptop. For the cliffs, I would have like to essentially make them with displacement shaders on heavily subsusrfaced objects, but with that method I'm just crashing (I would easily get past 5 million faces, which is a breaking point for my laptop). I have to considerably lower the amount of subsurf (from 6 levels of subsurf down to 4), and rely a lot on bumpmapping (which, alas, don't show detail on completely shaded areas, while displacement would). It reminds me of the old days when I was having fun with Terragen. However, if the Blender renderer can digest the amount of faces, rendering itself is not too much of a problem (this WIP took about 15 mins to render). Nevertheless, this causes some concern, as adding a population of flowers and the waterfall will put more strain on my limited computing resources. I'm actually wondering if I can finish it at all with my current hardware. I've dropped the idea of making waterfall with fluid simulation, though I hope to make it via smoke simulation instead.
EDIT: Well, fortunately I was able to partly solve my problems, I rearranged the mesh by adding some loop cuts and decreased the subsurf, which actually gives me better results with fewer faces. I'll post some progress on that later.
-
Really awesome man, after all those himble stuff you said at the beginning, LoL you're pretty good with this .. good luck! hope you finish this soon.
-
Progress update: Flower power!!! (click for high-res)
(http://img405.imageshack.us/img405/9937/wipflowers.png) (http://imageshack.us/photo/my-images/405/wipflowers.png/)
Now, there are a few things I've yet to tune about these flowers, namely:
- More randomization on the distribution/orientation of the flowers (I've to tune the grass distribution as well).
- Maybe put the stems and leaves to a darker shade of green, and fine tune the translucency.
I hope you like it though.
-
Thank you! This is amazing indeed, the level of detail and everything! Is there a way to put a slight animation on those flowers like there is a light breeze from the waterfall? and some steam cause by it around the waterfall area?
-
Thank you! This is amazing indeed, the level of detail and everything! Is there a way to put a slight animation on those flowers like there is a light breeze from the waterfall? and some steam cause by it around the waterfall area?
Well, the waterfall and the steam are definitely on my to-do list, it's pretty much the last big chunk I have to do. As for animating the flowers under a slight breeze, it's doable in theory via soft-body simulation. In practice, I don't think my computer can handle it - to be honest I'm pushing my machine pretty hard already. I will have to leave it at that (at least, until I get a new computer :P ).
-
If you need extra CPU/GPU power if you tell me how to get it done i can do it. 4.5ghz 3570k and a 570 basically just idling at this point haha.
-
If you need extra CPU/GPU power if you tell me how to get it done i can do it. 4.5ghz 3570k and a 570 basically just idling at this point haha.
Thanks for the offer! I'm not sure I'll need this for now though. First I want to get a still image right, animating would be a bonus (besides, even if I sort of know how that soft body simulation should work, I'd still need to learn how to actually do it before I could pass the information to you). We'll see about that later. Still, if I plan to make some crazy mountain landscapes for a later scene, or if I do want to make the waterfall using the fluid simulation for this one, I'll know where to ask ;)
-
I think Mayo's issue is more related to manipulating his awesomely big and beautiful model within blender than
how long it takes him to render it. Running a remote virtual machine from your pc to his laptop might help but so could
running blender from a lightweight linux distro. But I really dont know what im talking about.
I really just wanted to say that field screen looks freakin beautiful
-
It's really a beautiful upgrade to this scene!!! So detailed! With a slight animation it would be a wonderful addition to its experience...
-
I think Mayo's issue is more related to manipulating his awesomely big and beautiful model within blender than
how long it takes him to render it. Running a remote virtual machine from your pc to his laptop might help but so could
running blender from a lightweight linux distro. But I really dont know what im talking about.
I really just wanted to say that field screen looks freakin beautiful
well damn.. Alright i got $50 to donate for a new vid card, Who's with me?! :D
-
it's looking really good man, but i'm curious as to why you're going with tesselation + displacement, over low poly + bump map? you could recieve much the same result as you have now.
to be clear, not questioning your results, just your method =]
-
Thanks a lot for your positive feedback, guys!
I think Mayo's issue is more related to manipulating his awesomely big and beautiful model within blender than
how long it takes him to render it. Running a remote virtual machine from your pc to his laptop might help but so could
running blender from a lightweight linux distro. But I really dont know what im talking about.
I really just wanted to say that field screen looks freakin beautiful
At this point, I actually have both issues. As I manipulate my scene, I can work my way around by hiding the flower layers, and show it only for renders. Still, it's getting pretty slow. Then, yesterday I went for a full render with all the flower fields, it took about 1 hour to render. When I'm to including water transparencies and reflections, and then the waterfalls, it's gonna take time...
well damn.. Alright i got $50 to donate for a new vid card, Who's with me?! :D
;D Well, that's what I got for having purchased a laptop to browse the web and use MS office :P I plan on investing for a new desktop PC sometime next year though.
it's looking really good man, but i'm curious as to why you're going with tesselation + displacement, over low poly + bump map? you could recieve much the same result as you have now.
to be clear, not questioning your results, just your method =]
I actually used a combination of both. At first I wanted to do everything with displacement, but my laptop could not. So, I went for using displacements for large scale deformations, and use bumpmaps for small scale details. Bumpmap details are actually more apparent on this picture. I did not discard the large scale displacements, because I believe they are essential for 2 aspects:
- without the large scale displacements, the vertical ridges are too... "vertical". I needed some irregularities on those (check the side of the rock cylinders in high res, for instance). The displacements on the vertical ridges (and also on the back cliff) are also very important for the way the rocks cast shadows on themselves. Bumpmaps cannot do those things.
- It turned out that, in the completely shaded areas of the cliffs, my bumpmaps almost did not show. Without displacements, the shaded areas of the cliffs look very flat, and compared to the areas exposed to light, it was just looking really off. In a way, the displacements were almost the only thing that make the cliff look bumpy when it's completely shaded.
I think that relying on bumpmaps only could have worked at lower resolution, but as we go for pretty high-res, we enter a level of details where the displacements are necessary. I could say I learned that during my previous scene (at first I modelled a corrugated metal sheet with bumpmaps, but I considerably improved my results when I meshed the corrugations instead).
-
Bad render times? Turn of raytracing in the shading panel and use buffer shadows.
-
I'm adding the water at the moment (no waterfall yet). It's actually much trickier than I anticipated, for combining the sense of depth, the transparency, and the reflectivity of the water all at once. And I think this sort of materials requires that I work with raytracing. For that matter, including the actual index of refraction of water makes some funny effects.
On another note, I had a scare this morning with my laptop playing dead for 2 hours. Once it was back on, I had the urge to back up all my stuff on an external hard drive, so if I fry my laptop all won't be lost :P Probably I'll end up buying this new PC earlier than I originally thought ::)
-
I can donate 10 euros to your project. You can make exchange if you wish. Just send me a bank account that's compatible with a Greek bank.
-
I can donate 10 euros to your project. You can make exchange if you wish. Just send me a bank account that's compatible with a Greek bank.
I appreciate the thought. That won't be necessary though :)
By the way, I'm afraid that completing the scene will get delayed - even though I finished the lake area yesterday (only the waterfalls are missing). My laptop is failing, most likely I'll have to get it repaired, which may take some time. If it is beyond repair, I have everything backed up, so we're safe on that respect.
-
Glad to hear that it is all backed up. Your work is to fantastic to lose! Hope you can get your rig fixed up man!
-
Hi guys,
I have some news, and some questions...
As far as the EALS scene is concerned, my laptop reached its limits for handling particle systems (flowers and grass) plus a smoke simulation (waterfall). Currently, with my hardware, there's no more work I can do on it. So, hellbringer616 is kindly helping me out by providing his computing resources in order to carry out the smoke simulation, then the render. There should be more update about this by the end of next week.
Thanks to SpooX almighty walkmesh/camera exporter, I started to work on a new scene (mds5_i, item store in the sector 5 slums). Here's what I made this far (many elements are not subsurfed yet, so don't worry about the polygonal look of cylindrical features).
(http://img703.imageshack.us/img703/8217/wipmds5i.png) (http://imageshack.us/photo/my-images/703/wipmds5i.png/)
Most likely I'll be able to get to texturing fairly soon. Now, for texturing, I have some questions where I hope you'll be able to help me out. Considering the original picture (below), as you can see many elements have kanji writing on them, but the low resolution makes them very difficult to read, so if I could have some input on that, it would be great. My girlfriend helped me to decipher some (she has some basics in Japanese, my own knowledge amounts to nothing in this area), it will be good to pool our research notes.
(http://img.photobucket.com/albums/v287/Akabi/Itemshop.png)
- On the main pipe, there are 2 kanjis (first guess, from a chinese friend: 'The first kanji on the pipe is 迚, which means "very" or "Some way or another". When used with a negative verb form, it means "cannot/unable to do something". I can't make out what the second kanji says, or what that is between 迚 and the second kanji. For now, the closest I could find are 敦 (drink, liquid) and 飲 (industry).')
- On the cardboard boxes, there are 2 kanjis, no guess.
- On the shelf, there's a sort of green bowl/cup with 1 kanji, most likely it's 茶, which means "tea".
- On the shelf, there's a big brown pot, with at least 1 kanji (best guess: 漠 (vague)... probably there's a second we cannot read).
- On the doormat, there are katakana & hira (アイテム の)which seem to read "of item(s)", and below some kanji (from chinese friend: 'On the carpet, it looks like the second kanji is 良 (good). First one could either be 馬 or 烏. (horse or squid)'
- On the poster, there are katakana in red, no guess so far.
Any help will be much appreciated. By the way, I'm expecting to find similar issues on a bunch of scene (japanese writings practically unreadable), so if anyone has knowledge in this department, that would be awesome.
-
The writing on the jar reads “China” (漢), possibly the first character of “Chinese Medicine” (漢方).
The writing on the pipe is “Welcome” (迎歓). While “Welcome” would normally be written 歓迎, the the characters are read from right to left, as is done traditionally in Japan. Thanks to Qhimm for reminding me of this.
The welcome mat reads “Torishima Item Shop” (アイテムの鳥島屋).
A lot has already been deciphered....
Hint:Look Here (http://www.glitterberri.com/final-fantasy-vii/map-analysis/)
8)
-
Very nice! I didn't know so much had been already done. Thanks for the tip :)
UPDATE: modelling for that scene is practically done, so here's how it looks like (50% of final picture size)
(http://img716.imageshack.us/img716/8217/wipmds5i.png) (http://imageshack.us/photo/my-images/716/wipmds5i.png/)
For the cardoard boxes, I'll texture them (most likely with UV textures) before duplicating them all over the place. I was thinking of using a cloth simulation to let the poster fall into place (there are a bunch of rivets underneath). There are also a couple of sheathings I want to add around some cables.
Onward to lighting and texturing!
I hope you guys like it so far.
-
Can we get another 100 of you guys? Then we'd really be rockin' 8-)
-
The only I can say is beautiful!!!
I really want walk in the Aeris house :B
-
Can we get another 100 of you guys? Then we'd really be rockin' 8-)
Thanks for the compliment! Though as Ulpian pointed out, we're really understaffed.
The only I can say is beautiful!!!
I really want walk in the Aeris house :B
Thanks! As for walking inside the house, it will be possible eventually ;)
-
The progress across the threads is amazing... Although I can't model, I'm happy to run renders and such for people if needed. At least I can help like that!
I'm actually attempting to recruit some game-design grads into doing work for the projects, but the comment that I most get is "If only I had more time"! Any ideas for combating this remark? :\
-
Thanks! Well, maybe you can also check outside graph-design grads... I mean, I hope it doesn't show too much, but my experience with 3d modelling software amounts to 4-5 months (I actually pretty much started because I wanted to help Team Avalanche!), and my background is not that of an artist.
-
Mayo Master your work is outstanding! I can't believe for one minute that you only have 4-5 months experience in this!
Do you think you could explain how you became so good so fast? As a newbie I have attempted to line up blender cameras to field images and to be honest it seems impossible :(. Also getting the scale to be correct seems to be extremely difficult too, do you have any tips/methods?
Finally how are you getting the textures/colours to be so close to the original? Are you using GIMP to make textures from scratch? What rendering engine are you using, internal or cycles?
Sorry for so many questions, can't wait to see the final render of the house field :D.
-
Some people have natural talent in something, even though they just begin. Or they learn fast, like I do. Sometimes real life helps (concerned people will most likely understand or guess what I'm saying, so I will not bother to explain more).
-
Some people have natural talent in something, even though they just begin. Or they learn fast, like I do. Sometimes real life helps (concerned people will most likely understand or guess what I'm saying, so I will not bother to explain more).
You might say my education and job required that I'm a fast learner, so that certainly helped. Also, from my point of view, I would believe that people familiar with CAD software coming from engineering or architecture backgrounds could transition to "artistic 3d modelling" pretty fast.
Mayo Master your work is outstanding! I can't believe for one minute that you only have 4-5 months experience in this!
Do you think you could explain how you became so good so fast? As a newbie I have attempted to line up blender cameras to field images and to be honest it seems impossible :(. Also getting the scale to be correct seems to be extremely difficult too, do you have any tips/methods?
Finally how are you getting the textures/colours to be so close to the original? Are you using GIMP to make textures from scratch? What rendering engine are you using, internal or cycles?
Sorry for so many questions, can't wait to see the final render of the house field :D.
Let me try to answer all these questions... but for reference, you can also check out how I progressed on my very first scene (it's in the "Field scene screens (non bombing mission) gallery" thread), you'll see that there's a lot more fumbling around.
I would say that, at first, you have to be patient about getting the basics right. Make baby steps before you try to run on your own. Initially, omega_res_novae pointed me to the following wikibook, which helped a lot in the beginning.
http://en.wikibooks.org/wiki/Blender_3D:_Noob_to_Pro (http://en.wikibooks.org/wiki/Blender_3D:_Noob_to_Pro)
In the aforementioned wikibook, I personally went through the first two units entirely before starting to model on my own, even though it itched to do it before that. Then I really started with very simple objects (a beer can, a crumpled piece of paper, a trash can), and when I decided to model my first scene, I picked one that looked simple. As a noob, you have to be able to take satisfaction from completing very basic modelling, while being humble before the gap that separates you from true 3d artists (I think it's healthy to keep that mindset). As you take on gradually more complicated stuff, there are two sources that can be very helpful:
- Tutorials around the net. There are LOTS of those. I personally favor the tutorials from Andrew Price at Blenderguru.com, I like his teaching style.
- Real life. More explanations here: after I started to do 3d modelling, I would almost say that I took a completely new interest in just "observing the world" (yeah, I was feeling a bit nuts sometime). For example, I could stare at a tile for 10 mins to study how it was made, how the light was interacting with it, what affected specularity and reflections, etc. You can get a lot of information just from observing everyday's life environment.
Now, to your more specifically technical questions:
- Aligning the field screen and the camera IS a real pain. I managed to do it manually for my first scene, it took ages. To do it manually, generally the idea is to draw two parallel lines along each axis to get the alignment right, but it is very difficult. Which is why SpooX's walkmesh and camera data exporter is such a powerful tool. I strongly recommend you use it. If you don't have PSX discs to extract the data, ask me what field you'd like to work on and I can send you the Blender file, much like SpooX did for me while he was building his script. All that's left to do is adjusting the focal distance, but with the right alignment it's not difficult to do.
- About texturing: that's actually the thing I find the hardest. In my first "man in the pipe" scene, I must have redone the texturing of each object more than 15 times. I think one difficulty is that your colors are a combined consequence of both your texture colors and your lighting. To dissociate these components, I try to get a particular color reference. For example, in the scene I am currently working on (mds5_i), I know the top of the medkit has a white diffuse color (all RGB at 255), and the red cross on the side has the RGB of 255, 0, 0, in terms of texturing. Your lighting is what will make the difference between these texture colors and their RGB values on your rendered image. I find it easier to set the lighting from reference colors in the scene, and then work on the remaining textures. Of course it gets more complicated when your scene has several light sources (something I did not anticipate when I started "the man in the pipe"). Now for texturing, because my artistic skills are rather limited (I'm not unable to draw with paintshop or GIMP, but it's not my strong suit), I tend to prefer working with a combination of texture images (cgtextures.com may be your friend) and procedural textures. For example, the cliffs on Aerith's house outdoor scene are entirely made with procedurals. Also, combining the textures using "Multiply" instead of "Mix" can give pretty good results (I used that a lot for the roof of the house), in a way you can have a texture pattern with any color. As for color matching, if your light setting is correct, then it's a tedious looping process about going back and forth between rendering your model, compare it with original picture, see the discrepancies, adjust the texture colors in your model, re-render and so forth. At least that's how I do it. Don't forget that texturing isn't just about setting a diffuse colors, something like bumpmaping is almost as important. But if you're good with GIMP, more power to you because you can have much more accuracy with UV mapping.
- Finally, I simply use Blender's internal render engine, we don't need anything more fancy than that. Or at least, the improvements I can make to a scene are much more depending on my own improvements as a 3d modeller, than on the use of a fancy render engine. Besides, I don't use cycles, mostly because cycles is meant to help your rendering processes via GPU usage, and my graphic card sucks.
I hope all that helps :)
-
Thanks for your response it helps greatly. Already know some basics but I'll work through the book you linked and start doing that guys tutorials. It sounds like the observation is a vital point which I didn't appreciate until now :).
Interesting that you are using blender internal - I know cycles gives much more realistic lighting. But looking closely at the original field scene it has very black and sharp shadows which is not something that cycles would generate as this is not true to real life as the light is not refracting as much as it should be. I guess back when FF7 was actually created the rendering tech at the time may have been something similar to blender internal in relation to lighting, I've never really noticed until now.
If you could send me any walk mesh/field image to try I'd love to give it a go. It might be a long time and take many versions before I can come up with anything that's release quality though, I guess that will be part of the challenge :). How many hours per week do you think you spend using blender? I guess you also spend a lot of time outside of blender thinking of how things could be done when looking at real life objects?
Edit: I can't send PM's:
"You are not allowed to send personal messages." :'(
I'd like to try the reactor 5 tunnel :).
-
You may send PM's once you reach 5 posts. The limitation is designed to prevent spam PM's.
-
How many hours per week do you think you spend using blender? I guess you also spend a lot of time outside of blender thinking of how things could be done when looking at real life objects?
It's hard to tell, because it's highly variable (I had weeks with 2h on Blender, some others where I went crazy and spent about 30). I'd say about a dozen hours on average. As for the time spent in contemplation of real life objects, well, it's more like I see things with a new eye (or wondering "how would I model that?") when I do mundane things (walking from home to work, doing the shopping at the local store, washing the dishes), it makes life more interesting in a way :P Unless there's something very specific I'm looking for (in which case, the internet is also filled with reference pictures), I wouldn't go out specifically with the purpose of "observing real life in order to apply the knowledge in 3d modelling". Besides, it's currently -30 C in my place, so I wouldn't go out at all :P
-
As promised, some progress update on my part. I'm being a bit slow (though I was fairly busy these days), but I hope my hours in contemplation of brushed steel pipes have paid off :P
Even if it's not related, I anecdotally found out that the convoluted and obnoxious key-card system of the Shin-Ra top floors is actually pretty realistic :o
Anyway, enough rambling... (click for high-res)
(http://img843.imageshack.us/img843/8217/wipmds5i.png) (http://imageshack.us/photo/my-images/843/wipmds5i.png/)
At this point, I consider the textures of the cylindrical pipes (with the obvious exception of the painted "welcome" kanji) to be done, as well as the medkit, the yellow-black cable box, and the red pipes on the top. The main light source is done, but the ambient lighting will be adjusted last. Everything else is WIP.
I hope you like it so far.
-
Awesome work. I personally prefer the japanese writing on the pipe compared to the english writing though.
Keep up the great work!
-
Update... boxes!
I tried to write (I had some help for this) stuff that made sense on the carboard boxes, if someone fluent in Japanese can confirm whether it's fine or I have to revise it, that'd be nice (the writings correspond to "Shin-Ra", "Shin-Ra Pharmaceuticals", "Antidote", and "Potion", or at least that's what I aimed at).
(http://img571.imageshack.us/img571/8217/wipmds5i.png) (http://imageshack.us/photo/my-images/571/wipmds5i.png/)
I'm done with all the pipes and the wall (though I may have to adjust the location of the steel plate joints of the wall), the last was a real pain to texture so I hope it's fine for you as it is now. I'm wondering if adding a touch of dirt on the carboard textures would improve anything. I think the most tedious part of texturing is done at this point.
I hope you like it.
-
Everything looks too clean lol.. so of course i say to make the textures look more used up rustic and what not.
Though that would take up a pretty large amount of hrs of tedious work... Id say do it!
... But actually this time i have a solution. You should look up how to implement a Dirt Map into your scenes with blender.
I actually looked up one way...
http://blenderartists.org/forum/showthread.php?183082-Dirt-maps-creator
Pretty simple process that will make a huge difference in aging your scene
-
... But actually this time i have a solution. You should look up how to implement a Dirt Map into your scenes with blender.
I actually looked up one way...
http://blenderartists.org/forum/showthread.php?183082-Dirt-maps-creator
Pretty simple process that will make a huge difference in aging your scene
I can look into that if it can improve my texturing. However, with my inexperience, the "pretty simple process" is far from being straightforward (I've never done any texture baking before, I don't even know how to do that). I hope it will be worthwhile.
Everything looks too clean lol.. so of course i say to make the textures look more used up rustic and what not.
Can you be more specific? When I look at the original picture it doesn't seem that dirty to me. This may be the slums, but the guy has to have a decent place to set up a business. On another note, I have made quite a bit of real-life observations about how brushed steel pipes should look, and where not stained they retained all their specularity and reflectivity.
-
Nah not bashing on the metal work of the textures in general. Just trying to make a case for you to use a dirtmap shader on the scene. It will give the overall scene a more natural subtle "lived in" feeling.
-
Ok, after much sweat and blood and many Ambient-Occlusion baked dirtmaps, a new update that I would consider quasi-final (click for high-res):
(http://img39.imageshack.us/img39/8217/wipmds5i.png) (http://imageshack.us/photo/my-images/39/wipmds5i.png/)
So, unless I have insightful advice, I consider everything to be done except the painted kanjis (i.e. the "welcome" writing on the pipe and the "chinese medicine" writing on the big brown jar) and the posters (on the wall and on the door). For these elements, I will rely on the help of my girlfriend who has better 2d drawing skills than me, so that will come later. However the UVs are already set, I just need to update the texture images and it's done.
Let me know if that looks fine to you.
-
Looks great to me
-
I'm with sl1982. Great job Mayo Master! I just think it would benefit from softer shadows, as if the light was a bit more diffuse.
-
I'm with sl1982. Great job Mayo Master! I just think it would benefit from softer shadows, as if the light was a bit more diffuse.
Well, there can be some sort of open debate about this. On the one hand, as you mention, softer shadows may give a more realistic feeling to that scene, because indoor lighting/shadowing tends to do that. On the other hand, the original picture has somewhat hard shadows.
Since the general guideline was to remain faithful to the original looks, so far I chose to have only very slightly softened shadows. I could play around with soft shadows settings and submit that to a vote.
EDIT: So here are 4 tests with different settings for softshadows. The top-left picture is the version previsouly submitted (with soft shadows setting to 6), the others are with stronger shoft shadow settings (48, 86, 192). Click for a higher resolutions, the scenes were rendered here at 50% of the size.
(http://img28.imageshack.us/img28/5413/testsoftshadows.png) (http://imageshack.us/photo/my-images/28/testsoftshadows.png/)
While the strongest soft shadow settings feels more realistic, some shadows are so softened that they completely disappear. I think namely of the shadow of the rectangular pipe that is casted on the rotating beacon, in the top corner of the room. With that in mind, I would rather choose soft shadow settings between 48 and 96, but I'm open to suggestions.
One other thing I am considering, though it is separate from the lighting/shadowing: I'm considering using a displacement to model the cracks in the wooden floor (bumpmaps are used so far) to have even more realistic shadows. I think it would improve the area around the doormat. However, I will have to subsurf the floor so much that I don't know if my computer can take it.
-
^^^^
SS48 and 96 are both very good. I do like the shadows casting in the corner more on 48 but I do like the softer shadows on 96 with the Welcome pipe. I couldn't complain with either one being in game.
-
While the strongest soft shadow settings feels more realistic, some shadows are so softened that they completely disappear. I think namely of the shadow of the rectangular pipe that is casted on the rotating beacon, in the top corner of the room. With that in mind, I would rather choose soft shadow settings between 48 and 96, but I'm open to suggestions.
One other thing I am considering, though it is separate from the lighting/shadowing: I'm considering using a displacement to model the cracks in the wooden floor (bumpmaps are used so far) to have even more realistic shadows. I think it would improve the area around the doormat. However, I will have to subsurf the floor so much that I don't know if my computer can take it.
Doesn't Blender support radiosity and photons like POVray does? That was how I always rendered soft shadows. What radiosity does it create 'area' lighting based on the reflective properties of surfaces. Photons allows light to refract through a surface (think window glass light, lens affects and prisms). If you have light sources set in the scene there should be some way to create the area light sources to soften the shadows.
That's what I would do but ... blender was always a bit weird to work with for me, a bit counter intuitive so I always had a difficult time with fiddling with the settings. Drove me nuts "what I can't enter the exact size I have to use this screwy drag the edge thng, where's the snap... argh" hehehe.
That kind of stuff.
Cyb
-
As far as I know, the softness of shadows in Blender is independant from the materials properties. In the material properties you control diffuse color, specularity, reflectivity and transparency. The softness of the shadows depends on the settings of the lights. Also, in this scene I believe the light is emitted from a "point" object, given how each object cast their shadow depending on their location.
-
I like SS96 best (although SS48 looks great too). On the other hand, I'm not sure that recreating the bumps on the floor would have a desiderable effect. I think it's fine the way it is.
-
I think I'll be agreeing with the trends so far - SS96 is best, SS48 is also really nice.
Are you going to disperse the shadows any further like they were in the original, or keep them well defined?
-
SS96
-
Definitely agree with SS96
Also damn good job on the modeling and texturing of that room
-
SS96 is looking good. However, if you just want shadows to be soft you can turn of the shadows of your current lamps.
Move a set of Spot lamps in a place where you want the shadows to be and check shadow only.
If you you set them to use shadow buffers you can tweak the softness seperately and it will render much, much faster than using 96 raytracing samples on lamps if you just want soft shadows.
-
While it is pretty evident that SS96 has the best shadows, it still needs some
dirt in the pipes and the boxes and almost everywhere. Well done anyway! :)
-
Thanks for the feedback! I will go for SS 96 setting then.
SS96 is looking good. However, if you just want shadows to be soft you can turn of the shadows of your current lamps.
Move a set of Spot lamps in a place where you want the shadows to be and check shadow only.
If you you set them to use shadow buffers you can tweak the softness seperately and it will render much, much faster than using 96 raytracing samples on lamps if you just want soft shadows.
Thanks for that rendering tip. I don't think I'll use that for this scene though. As a matter of fact, the setting here corresponds to the soft shadow size, not the number of samples. The number of samples here is set to 6, and this amount is enough to give good results while having decent render speed.
While it is pretty evident that SS96 has the best shadows, it still needs some
dirt in the pipes and the boxes and almost everywhere. Well done anyway! :)
I disagree with you about the amount of dirt. Hoping that I can convince you, I'd like to mention that I already added significant amount of dirt on floor, pipes and walls, and I have studied how brushed steel pipes get dirty. The place is a shop, and, in my opinion, should be well-kept. You may think it should be dirtier just because we're in the slums, but I can redirect you to counter examples (for one, Aerith's house is in the slums, and it's not dirty). Concerning the boxes and jars, I didn't add dirt because it wouldn't be relevant for these items to have collected much dirt if one assumes the cardboard boxes come from this morning's shipping, and the jars are frequently used.
-
Haha! Nice explanation!Famboism against fanboism lol
-
96 looks fine with regards to the shadows.
As for the dirt, tone down the colors a bit, but you might solve that by a color correction after rendering.
I must agree, you will find not much dust in a shop, concider the owner, even in the slums, would be proud to have his own little shop. I don't think one should concider that fanboism, but logic needed to build the scene.
Nice going Mayo, keep up the good work (and please get rid of those full blown RGB posters :P)
8-)
Ps I'm still in the process to get my scenes back up and running....
-
Make sure not to confuse dirt/grime with just being dark. The place is sure to suffer from poor lighting. I think the current light set up is rather bright, especially when combined with the dummy textures.
Low lighting, combined with some yellowish lights will set the mood a lot more, though I don't believe you've gotten to the final lighting passes anyways.
-
Make sure not to confuse dirt/grime with just being dark. The place is sure to suffer from poor lighting. I think the current light set up is rather bright, especially when combined with the dummy textures.
As for the dirt, tone down the colors a bit, but you might solve that by a color correction after rendering.
I believe I got the light settings right. The original scene isn't poorly lit, actually. For each texture I made, I repeated a tedious looping process of rendering, compare RGB values with original picture, adjust texture colors, re-render and so forth. I may not be on an exact matching (for that matter, the color scheme of the original game as only 32 values for RGB as opposed to 256, so exact color matching is irrelevant), but I think I come close enough - you can make the comparison of RGB values yourself if you like.
-
You people realize the more you nitpick this the longer it will take before you will see it released? It is a fine looking render
-
You people realize the more you nitpick this the longer it will take before you will see it released? It is a fine looking render
I don't want to seem lazy, but I couldn't agree more with that statement. It matters a lot that we get things done, because there are some 600+ scenes in total. I'm aware of my own limitations as a 3d modeller, and scenes are always perfectible, but we have to get to a point when we're happy with what we get and move on. There are things I might have wanted to redo on the first scene I modelled, but I have to settle with it. Given my personal productivity (which hopefully will also improve with time), I should be aiming at completing at least one scene per month. If we have the ambition of carrying out this project to the end within a reasonable amount of time (still, about 6 years), we would need 10 times that productivity from the whole team, and unfortunately we're much fewer than 10 people on this job (a shame Ulpian seems to have vanished, really).
-
Well as long as you keep a backup of your source files then they can alway be revisited later if need be. It makes more sense to get some good scenes out now, especially since someone will always have something to gripe about with them.
-
I sure keep a backup of the source files, no worries.
By the way, for the anecdote, did you know that the way the scene is laid out is completely unpractical for the shop clerk? Actually, one can hardly use the till because the square pipe blocks half the way to the table! I wonder if that could be the sort of things people might want to revisit later, as well.
-
Haha! This is so entertaining! Wonderful work though anyway! My personal preference is scenes to become as most realistic as possible ;)
-
I hope you like where this is going (they're rendered at 25% of their actual sizes)...
(http://img6.imageshack.us/img6/2503/wipealin1.png) (http://imageshack.us/photo/my-images/6/wipealin1.png/)
(http://img812.imageshack.us/img812/3848/wipealin12.png) (http://imageshack.us/photo/my-images/812/wipealin12.png/)
Two remarks:
- Making all these rocks is a real trial of patience. So tedious. And still so much to do.
- The original design of the stairs is actually really flawed. According to regular building codes for stairs (check stairs on wikipedia for reference), the standard pitch angle (more or less the angle of the ramp with the horizontal) of stairs is about 30 deg, and the maximum angle is about 45 deg. Here, the pitch angle is 56 deg, which makes this construction something between stairs and a ladder. Another flaw is that, assuming the floor is about 2.30 m high, the number of steps on the original picture would yield a rise height (the height of each step) above 40 cm, which is clearly wrong (try to climb that!). Thus, I took the liberty of arranging slight modifications so that the stairs would be somewhat usable. The walkmesh being what is is, I maintained the pitch angle to 56 deg. I increased the number of steps so that the thread depths would be ergonomic (i.e. about 9 inches or 23 cm minimum) while decreasing the rise of each steps. The best I could do was to design the steps with roughly 34 cm of elevation (decreasing the rise more than that would result in unsufficient thread depths). It's still very steep, but much more usable, while fitting the walkmesh just like the original. I hope you're fine with these revisions.
-
you really can overdo things....stairs, regulations, next you're going to stamp an ISO number on the door... :-P
as for the stones on the floor, go either with a displacement map, or even a modifier (http://wiki.blender.org/index.php/Doc:2.6/Manual/Modifiers/Deform/Displace), in that way you can paint the stones and be done with it. I would love to see some wires of the current scene.
Nice start though, I see you refined the stair-pillars.
8-)
-
You may say that it's a professional quirk that I want things to be functional :-P That being said, I think my modelling is improved by looking at actual references, so why not check out the real deal?
Concerning the stones in the floor: using a displacement shader was a possibility that I examined in the first place. However, I couldn't find a proper displacement map which would remain faithful to the floor pattern of the original picture. If I were to draw the map myself with Paintshop, it would take pretty much the same time as making the actual mesh. Besides, making the mesh of the rock is less costly in terms of poly counts than subsufing the floor to the level required for a displacement shader, I think. So I eventually decided to go for making a mesh of the floor. In a sense, I feel like the time I am spending making this mesh will be compensated by the fact that texturing it will be much simpler that way.
By the way, for reference, here are the wires, though they show some levels of subsurf (3 levels of subsurf on stair pillars and studs, 2 levels of subsurf on the rocks). As far as the unsubsurfed mesh is concerned, the stair pillars and studs are based on 16 sided cylinders, and each rock is actually based on 12 faces.
(http://img571.imageshack.us/img571/2783/ealinwires.png) (http://imageshack.us/photo/my-images/571/ealinwires.png/)
-
You may say that it's a professional quirk that I want things to be functional :-P That being said, I think my modelling is improved by looking at actual references, so why not check out the real deal?
Well I prefer things be something close to right, it reminds me of a discussion someone had on Uncharted 3 about a section of the game where you crossed a ship in a storm. The original idea was too make the entire scene a repeated animation and fake the storm and that would be good. However the programmer said "why not make it a real wave simulation it's not like we don't have the power and more importantly it would not be predictable and thus more realistic".
Concerning the stones in the floor: using a displacement shader was a possibility that I examined in the first place. However, I couldn't find a proper displacement map which would remain faithful to the floor pattern of the original picture. If I were to draw the map myself with Paintshop, it would take pretty much the same time as making the actual mesh. Besides, making the mesh of the rock is less costly in terms of poly counts than subsufing the floor to the level required for a displacement shader, I think. So I eventually decided to go for making a mesh of the floor. In a sense, I feel like the time I am spending making this mesh will be compensated by the fact that texturing it will be much simpler that way.
I concure actually, I prefer to do that because textures and faking it don't always work in all cases. I have been doing optical simulations using polycarbonate material index of refractions and it iritates the crap out of me that I can't vary the IOR based on the wavelength of light. Their are a number of interesting distortions one gets in materials that aren't obvious, for example IOR is non linear (and so plastic prismatic behavior is different than diamond glass or flourine based materials). I like things to look real not 'good enough' LOL. So if you actually create a flour that can be shaded in all circumstances it will save you time screwing around with other fiddling details such as stretching the uv map to look right in a fixed direction. I find it harder to twiddle textures than make shapes.
By the way, for reference, here are the wires, though they show some levels of subsurf (3 levels of subsurf on stair pillars and studs, 2 levels of subsurf on the rocks). As far as the unsubsurfed mesh is concerned, the stair pillars and studs are based on 16 sided cylinders, and each rock is actually based on 12 faces.
I've a dumb question, I know blender supports python scripting etc. Isn't their surface brushes available to 'brush' on rock surfaces in fitted patterns? It seems like a common feature / need too me. I have seen a lot of games use surface normal textures or height fields to perform the same affect as individual rocks. It's very similiar to using terrain making software.
Anyhow looks too me you are doing fine with what your working on and enjoying it. That's the important part, don't do stuff you can't enjoy.
Cyb
-
Well I prefer things be something close to right, it reminds me of a discussion someone had on Uncharted 3 about a section of the game where you crossed a ship in a storm. The original idea was too make the entire scene a repeated animation and fake the storm and that would be good. However the programmer said "why not make it a real wave simulation it's not like we don't have the power and more importantly it would not be predictable and thus more realistic".
I've seen that interview too :)
I concure actually, I prefer to do that because textures and faking it don't always work in all cases. I have been doing optical simulations using polycarbonate material index of refractions and it iritates the crap out of me that I can't vary the IOR based on the wavelength of light.
You're doing optics with bulletproof glass?
Their are a number of interesting distortions one gets in materials that aren't obvious, for example IOR is non linear (and so plastic prismatic behavior is different than diamond glass or flourine based materials). I like things to look real not 'good enough' LOL. So if you actually create a flour that can be shaded in all circumstances it will save you time screwing around with other fiddling details such as stretching the uv map to look right in a fixed direction. I find it harder to twiddle textures than make shapes.
Well, the nice thing of displacement shaders is that they don't fake shadowing and interaction with lights, they act as an actual transformation of the underlying mesh. However, the underlying mesh needs to be highly subsurfed and regular (rather based on squares than stretched rectangles, for instance) for them to work properly. In any case, I find texturing harder than building meshes too. On my recent "Item shop" scene (mds5_i), it took me actually 5 days to make the objects, and 4 weeks to texture them.
I've a dumb question, I know blender supports python scripting etc. Isn't their surface brushes available to 'brush' on rock surfaces in fitted patterns? It seems like a common feature / need too me. I have seen a lot of games use surface normal textures or height fields to perform the same affect as individual rocks. It's very similiar to using terrain making software.
Actually, I'm not exactly sure what you mean. At first I was wondering if you were speaking of a texture paint mode (and it's in Blender, I used it quite a bit to make grime textures on mds5_i), though as you mentioned height fields I don't know if you were speaking of displacement shaders (or tesselation) instead. And displacement shaders are a key element in terrain making software (I encountered them first when I fiddled with Terragen).
Anyhow looks too me you are doing fine with what your working on and enjoying it. That's the important part, don't do stuff you can't enjoy.
Well, making all these rocks is a bit of a chore, but I'll be happy when it's done. It's a bit the "I am washing the dishes" moment of 3D modelling, it needs to be done before you can have something more enjoyable.
-
Progress update... huzzah! I made all the rocks ;D
(http://img11.imageshack.us/img11/3848/wipealin12.png) (http://imageshack.us/photo/my-images/11/wipealin12.png/)
It's rendered at 50% of the size.
-
Sweet deal! Is any of this stuff on the repo yet?
-
On the repo, I only updated my completed scene (mds5_dk) and the blender files related to the outdoor scene of Aerith's house (eals_1) because I need hellbringer_616 to help me with the smoke simulation I want for modelling the waterfall (which requires lots of RAM and processing power). I still have things to modify about that one (you may say it's 95% done).
I haven't uploaded the item shop yet (mds5_i), I'll do that when it's completed. I only need a few textures (for the 3 posters) which my girlfriend will draw (hopefully within the next 2 weeks).
As for Aerith's house (indoors), I only started, though by finishing the rock floor I have already done a big tedious chunk of work. I am planning on modelling the entire interior in just one file. The ground floor model will give ealin_1 and ealin_12, the top floor is ealin_2. However, I am having considerable trouble to have proper camera alignment on ealin_2 (even with SpooX's almighty ShinraTool).
Is there anything else in particular you would like to get on the repo? (WIP renders? Blender files?)
-
Are you planning on cutting up the pieces to put in game or do you need someone else to do that? If someone else is doing they cutting they are going to need to blender files to render masks.
-
When are you guys going to show pictures of your Jessie, Biggs, Wedge, and First Ray models?
-
Sorry on the major delay in the smoke sim, I've been pulling 65 hour work weeks and have had zero time for anything. I'll see if i can't fire it up tonight or tomorrow
-
Sorry on the major delay in the smoke sim, I've been pulling 65 hour work weeks and have had zero time for anything. I'll see if i can't fire it up tonight or tomorrow
Thanks for letting me know. I'll be greatful if you can run these tests, but don't worry about the delays and take some time for a rest :)
As for the current scene: the wooden furniture is done except for the table, now I'll have to fill all these shelves and cupboards. Hopefully I should be able to finish Elmyra's CRT TV today (textures and all), I'll send a picture when that's done.
EDIT: ... and DONE!
(http://img543.imageshack.us/img543/9075/wipealintv.png) (http://imageshack.us/photo/my-images/543/wipealintv.png/)
I hope you like it!
-
I have the sinking feeling that nobody's on the forum to check out whatever trash I might upload :P
(http://img824.imageshack.us/img824/5139/kitchensink2.png) (http://imageshack.us/photo/my-images/824/kitchensink2.png/)
(idiotic pun intended)
-
I dig through your trash all the time! Nice sink man
-
very punny of you!
Actually the CRT visual receiving system looks pretty darn good no aliasing patterns visable on the speaker section.
The sink looks ... like a sink! OH MY ... anyhow :D
Visually they aren't stunning but then again they weren't in the originalm so it's par for the course. They do add realism.
I always found it funny how many CRT's and stuff that Square threw in things, even though flat panels were known to be the wave of the future way back in 1989 (seriously). History is always interesting.
back to some fun erstwhile.
Cyb
-
Hey guys,
After some various delays, I finally managed to get a new render of my ealin_1 scene. It's still WIP of course (the furniture needs to be filled with miscellaneous stuff), but I hope it will also serve to get this forum out of the limbo.
Click for full res...
(http://img689.imageshack.us/img689/3848/wipealin12.png) (http://imageshack.us/photo/my-images/689/wipealin12.png/)
I hope you enjoy.
-
Everything's great now! I would prefer sharper and more realistic textures though if this was possible... Thank you for supporting this amazing game!
-
Everything's great now! I would prefer sharper and more realistic textures though if this was possible... Thank you for supporting this amazing game!
Well... I did the best I could. I'm not sure what you mean by "sharper and more realistic textures" (there are already a number of dirtmaps/ambient occlusion textures)... is there any particular item or object in this scene you're concerned with?
-
Hey guys,
After some various delays, I finally managed to get a new render of my ealin_1 scene. It's still WIP of course (the furniture needs to be filled with miscellaneous stuff), but I hope it will also serve to get this forum out of the limbo.
Looks pretty good too me. I could nit pick but that doesn't help. (For example not enough particulate matter how can a house be the clean, kind of sillyness).
The question is do YOU like it (LOL). I don't remember seeing the original and it might take me a bit to get FF7 up and working on the PC.
I can't say where my original playstation is, and it might take a bit to get PCSX working under linux (if I can get bloody Gentoo too boot again).
Anyhow good enough I say. If it has the right color and light balance. You aren't going for photo realistic so ... anyhow.
Cyb
-
All the original fields can be seen via the File Reconstruction Sheet (https://docs.google.com/spreadsheet/ccc?key=0AhqH9L8_QubjdEstb1JrRjV6NFFzOHBtSG8wVG93MWc#gid=0), in the "LGP files" tab, at the "Field LGP" section (the scene I'm working on here is ealin_1). That can give you a good reference.
When setting the lights and the textures I always go back and forth between my renders and the original image to adjust the light and color balances, so I think I'm fine in that respect. Overall I'm quite happy with what I'm doing, given my limitations (in terms of talent, experience, time, hardware resources, and texture source files). Besides, I'm far from being as productive as I would like, so I can't really afford to model details such as stains of burnt grease on the side of the stove. I could dedicate more time on details if we were a hundred modellers on the project, but alas we are probably 4-5 people on the job, at best.
-
Thank you Mayo master! I appreciate all your hard effort you have already put on this. I admit the ambient occlusion idea was such a great thought! It livens up the mood a bit on this scene. By realistic I mean more depth to the texture. I get kind of a glossy feel when I watch this image. These are some reference pictures:
(http://img547.imageshack.us/img547/8189/zack4.th.jpg) (http://img547.imageshack.us/i/zack4.jpg/)
(http://img42.imageshack.us/img42/606/rzde01f6dbdc6914.th.png) (http://img42.imageshack.us/i/rzde01f6dbdc6914.png/)
(http://img255.imageshack.us/img255/8236/rzde01bd1a71fc14.th.png) (http://img255.imageshack.us/i/rzde01bd1a71fc14.png/)
(http://img341.imageshack.us/img341/5443/rzde01a4c14bb114.th.png) (http://img341.imageshack.us/i/rzde01a4c14bb114.png/)
(http://img90.imageshack.us/img90/5207/rzde013688839914.th.png) (http://img90.imageshack.us/i/rzde013688839914.png/)
(http://img26.imageshack.us/img26/686/rzde0114190e5e14.th.png) (http://img26.imageshack.us/i/rzde0114190e5e14.png/)
(http://img189.imageshack.us/img189/6307/rzde01307e7aca14.th.png) (http://img189.imageshack.us/i/rzde01307e7aca14.png/)
(http://img811.imageshack.us/img811/1333/rzde01096e6df014.th.png) (http://img811.imageshack.us/i/rzde01096e6df014.png/)
(http://img11.imageshack.us/img11/792/rzde0116f35d9214.th.png) (http://img11.imageshack.us/i/rzde0116f35d9214.png/)
(http://img29.imageshack.us/img29/6180/rzde019d8e98cd14.th.png) (http://img29.imageshack.us/i/rzde019d8e98cd14.png/)
(http://img819.imageshack.us/img819/9647/rzde019c5049eb14.th.png) (http://img819.imageshack.us/i/rzde019c5049eb14.png/)
(http://img89.imageshack.us/img89/2429/rzde016f0fa37f14.th.png) (http://img89.imageshack.us/i/rzde016f0fa37f14.png/)
(http://img690.imageshack.us/img690/589/rzde015c22826514.th.png) (http://img690.imageshack.us/i/rzde015c22826514.png/)
*textures created by the user "Insane Slug" (http://insaneslugstexturepacks.blogspot.co.uk/) for a re-texture project of "Zelda - The twilight princess"
-
Ok, I think I see your point. Honestly, I believe we could debate endlessly about textures in that respect, and every modeller might have a different idea about it. Nevertheless, let me explain a bit the take I had on texturing the furniture. Most of the texture images you showed are rather old worn wood textures, like if they had been exposed to the weather outside. However, when I see the furniture in Elmyra's home, I would rather have the idea that it is made of varnished wood, and that the house is well-kept and cleaned regularly. That's really the impression I get when I look at the original image. Thus I wanted a feel of "even if it may be somewhat old, it still looks rather new" in those textures. That's why I picked them in these styles, and that I had a very very light touch when it came to add dirt shaders. In fact, I did also add small scratches on many wood textures, but since they were made at a rather realistic scale you can hardly see them even in the high res picture.
One example describes that fairly well: the piece of furniture under the TV. It's a wooden piece of furniture with glass doors with magnet locks: those things were very popular around the end of the 90's (i.e. very modern if you compare that with the release date of the original game). Thus, it would make sense to make that piece of furniture look "new and modern" like if it's straight from your IKEA store, and not made with old worn wood. In a way, my take on Elmyra's home is about having an old house that has been renovated with modern equipment. With that in mind, I think the style of textures I picked is more appropriate.
-
Small update on various stuff...
Thanks to my girlfriend I finally have this poster texture for the mds5_i scene...
(http://img90.imageshack.us/img90/8246/wiposter.png) (http://imageshack.us/photo/my-images/90/wiposter.png/)
Now, in ealin_1, Elmyra is cooking some tamagoyaki...
(http://img690.imageshack.us/img690/9333/wipkitchen.png) (http://imageshack.us/photo/my-images/690/wipkitchen.png/)
I hope you like it!
-
Well done "girlfriend"! :P :P I want some "tamagoyaki"!
-
This is awesome work!!!
Contact me if you want me to help out with some things.
You make our dreams come true!
Keep up the great work!
-
Kitchen cabinet of Elmyra's home (ealin_12 camera):
(http://img138.imageshack.us/img138/1203/wipkitchencloset.png) (http://imageshack.us/photo/my-images/138/wipkitchencloset.png/)
-
Yuri!
-
IKEA...
-
IKEA...
An Ikea bookcase would not have the durability to hold that many books, it must be akin to a higher quality brand
I am very excited to play this field in the game, it is coming together very nicely. Hopefully the textures of the character models will blend well with the colors of the backgrounds
-
An Ikea bookcase would not have the durability to hold that many books, it must be akin to a higher quality brand
Oh Gosh! How could anyone deny that... haha! I have spotted one that looks alike in the stock section though lol!
-
Updating...
(http://img827.imageshack.us/img827/1870/wipviewtabletv3.png) (http://imageshack.us/photo/my-images/827/wipviewtabletv3.png/)
Actually, this is only a 50% size of a render I made... but when I tried to upload the full res render at imageshack, it said "file too big for upload" :?
Anyway, I have 3 tasks to carry out for the scene to be completed:
- Make all the flowers (and that's going to take time, I think there are about 5-6 different flower species in the scene)
- Make the painting (near the bottom of the stairs)
- Make final adjustments/retouches on textures/lighting
-
I like the table sheet texture ;)
-
There's more! Here's a whole update on the ealin_1 scene:
(http://img89.imageshack.us/img89/550/wipealin1new50epc.png) (http://imageshack.us/photo/my-images/89/wipealin1new50epc.png/)
So... actually I did render the image at full resolution, but again I had troubles to upload it. So what you have here is a 50% size of the full-res that I rendered. As you can see I mostly need to make the flowers. There's also something weird going on with the glass of the kitchen cabinet, where it looks completely black... I wonder what kind of reflection led to that (and I hope to correct it at some point). There may be a few things to retouch.
By the way, I'm also coming across a certain issue which will be recurrent in many scenes, and that's something that was pointed out by Timu and Ulpian in the Sector 6 thread (about the fence bordering the playground): a bunch of objects in the original scenes (and in this one in particular) have very awkward scales. For instance, the teacups in ealin_1 would be the size of bowls in the original image, while the flowers have ridiculous proportions (they would be some 50 cm across). They were made so in the original scenes because the very low resolution was a cause for enlarging these objects in proportions that would make them recognizable by the viewer. I believe we would need to make a few corrections in that department. Still, I would like the opinion of the mods in that respect, namely to figure if we should do that kind of changes systematically. Modellers would have more work because we would need to figure what should be the proper scale of the objects involved in the scene (as opposed to following the original image blindly).
-
So... actually I did render the image at full resolution, but again I had troubles to upload it. So what you have here is a 50% size of the full-res that I rendered. As you can see I mostly need to make the flowers. There's also something weird going on with the glass of the kitchen cabinet, where it looks completely black... I wonder what kind of reflection led to that (and I hope to correct it at some point). There may be a few things to retouch.
By the way, I'm also coming across a certain issue which will be recurrent in many scenes, and that's something that was pointed out by Timu and Ulpian in the Sector 6 thread (about the fence bordering the playground): a bunch of objects in the original scenes (and in this one in particular) have very awkward scales. For instance, the teacups in ealin_1 would be the size of bowls in the original image, while the flowers have ridiculous proportions (they would be some 50 cm across). They were made so in the original scenes because the very low resolution was a cause for enlarging these objects in proportions that would make them recognizable by the viewer. I believe we would need to make a few corrections in that department. Still, I would like the opinion of the mods in that respect, namely to figure if we should do that kind of changes systematically. Modellers would have more work because we would need to figure what should be the proper scale of the objects involved in the scene (as opposed to following the original image blindly).
You have made an important point, a lot of the things they did were to make sure the details weren't invisable at the 320x224 game resolution on the playstation. As far as I know correcting for scale as needed is a good idea. However I cannot speak for your team. There are a lot of objects that were too small to render without having higher resolution (obvious statement).
That being said, most of these are interior field locations. Although the ones that are at that beach town aren't all "indoors".
Cyb
-
As far as I know correcting for scale as needed is a good idea. However I cannot speak for your team.
Let's hear what the mods have to say, then :)
-
This is a lovely scene! Actually I'm live streaming Aeris theme right now in a lovely remix haha! What a coincidence!
Jeremy Robson - Valse Aeris (Flowers Blooming in the Church ~ Aerith's Theme
-
Hi guys,
Just showing what I've been up to this weekend...
(http://img15.imageshack.us/img15/8936/wipgladiolus.png) (http://imageshack.us/photo/my-images/15/wipgladiolus.png/)
... making the flowers is so time-consuming! I still have 3 and a half flower bunches to do...
-
Remaking the scale is fine since they will be identifiable at target resolution. I dont think anyone would have issues with this. It is mainly just for cosmetic things anyways as the structural stuff would have the follow the original scale.
-
Thanks for the little details! They are useful in livening up the scene a bit and... nice wall texture :)
Wood is oak right?
-
Thanks for the little details! They are useful in livening up the scene a bit and... nice wall texture :)
Wood is oak right?
Hmmm it could be clear pine or poplar ... no looks to me to be clear pine oak I believe has a slightly different pattern, this is more poplar maple or clear pine. The number of wood patterns is nigh unto impossible to enumerate.
Hi guys,
Just showing what I've been up to this weekend...
... making the flowers is so time-consuming! I still have 3 and a half flower bunches to do...
I remember doing a christmas tree once, I made a single needle ... and then I had to use a dither pattern and spirally wind the needles around each branch, The same deal with the branches branching and thickness adjustment. It ended up being a study in fractals (mutters) and 'the soccer' formulea of form generation. Their are some nifty plant and flower generators in the web that made my little program ... look bad (LOL).
Anyhow I digress (and I wonder where the image of that tree is...). Plants are difficult if you don't use a parametric generation system with some sort of control system.
Cyb
-
Remaking the scale is fine since they will be identifiable at target resolution. I dont think anyone would have issues with this. It is mainly just for cosmetic things anyways as the structural stuff would have the follow the original scale.
Thanks of the heads-up. I'll make some revisions and revisit the other scenes as well (I recall that in the mds5_i item shop scene, there was a tea cup the size of a salad bowl ??? )
Thanks for the little details! They are useful in livening up the scene a bit and... nice wall texture :)
Wood is oak right?
Thanks. The wall is essentially made with procedurals, with a very slight dirt / AO shader. As for the wood, I actually don't know what kind it is :P . I used that texture (http://cgtextures.com/texview.php?id=14121&PHPSESSID=h350vo1t3a3eai9e1mtntrlsq2 (http://cgtextures.com/texview.php?id=14121&PHPSESSID=h350vo1t3a3eai9e1mtntrlsq2)) as a basis, which I combined with another base color mixed with a cloudy color pattern.
Hmmm it could be clear pine or poplar ... no looks to me to be clear pine oak I believe has a slightly different pattern, this is more poplar maple or clear pine. The number of wood patterns is nigh unto impossible to enumerate.
I remember doing a christmas tree once, I made a single needle ... and then I had to use a dither pattern and spirally wind the needles around each branch, The same deal with the branches branching and thickness adjustment. It ended up being a study in fractals (mutters) and 'the soccer' formulea of form generation. Their are some nifty plant and flower generators in the web that made my little program ... look bad (LOL).
Anyhow I digress (and I wonder where the image of that tree is...). Plants are difficult if you don't use a parametric generation system with some sort of control system.
I think I now have a good method for making the plants, though it's not parametric. Using reference pictures to cut out leaves and petals and make UVs is the first main step, then the very nifty tool in the process is lattice deformation. However, what start to take an awful lot of time is due to the limitations of my laptop: the scene has so many elements that now there's always a bit of lag between what I do and when Blender responds (like 0.3 seconds lag). It gets pretty irritating. Hopefully I should be able to salvage a "barely used" 4 years old PC very soon, which should be still much faster than what I have right now.
-
Updating... only one more bunch of flowers to go!
(http://img209.imageshack.us/img209/1048/wiphibiscus.png) (http://imageshack.us/photo/my-images/209/wiphibiscus.png/)
Talking about the scaling issue, it's actually fortunate that hibiscus makes such big flowers (more than 20 cm in diameter). They're very fitting as the table centrepiece, given the huge flowers in the original scene.
Otherwise, I wonder if I should upload a "make a flower" tutorial at this rate. The scene features flowers similar to lilies, marigolds, pansies, gladiolus, hibiscus, and I'll finish with some potted penstemon.
-
Updating... only one more bunch of flowers to go!
Talking about the scaling issue, it's actually fortunate that hibiscus makes such big flowers (more than 20 cm in diameter). They're very fitting as the table centrepiece, given the huge flowers in the original scene.
Otherwise, I wonder if I should upload a "make a flower" tutorial at this rate. The scene features flowers similar to lilies, marigolds, pansies, gladiolus, hibiscus, and I'll finish with some potted penstemon.
Consider the number of scenese with flowers, the church, the ancient forest, that's just off the top of my head. I can't remember about the chocobo farm but ... somehow.
I would say "please do so" but don't kill yourself doing it :D
Give it a try, sometimes by trying to do a tutorial you can vastly improve what you do yourself. (my case anyhow)
Cyb
-
(http://www.frogforum.net/attachments/tree-frogs/24502d1329586823-got-new-red-eye-intro-tips-farnsworth-good-news.jpg)
Ealin_1 and Ealin_12 are practically done, everything is modelled. I may have to retouch a few textures, rescale a few items and change a bit the lighting, but I'll do that later, when I have a more powerful computer. Finishing this scene was a real pain because my poor old laptop was short on memory. I could hardly do anything on my blender file in the end. In fact, I had to use another PC to make the renders. In the meantime, I hope you'll enjoy these (they're full res):
ealin1
(http://imageshack.us/a/img838/8595/ealin1done.png)
ealin12
(http://imageshack.us/a/img541/6194/ealin12done.png)
Feel free to make comments and suggestions, so that I could improve this scene later on. Mostly, I think I'll need to adjust the light setting: I had to use a lot of ambient lighting because of how the external walls are lit (and ambient light is the only light source for these), but it makes most of the scene with too diffuse lighting, and it lacks of contrast in my opinion. I'll also have to revise a few textures (I made the painting texture in 5 mins, I'll have to come up with an improved version, for one thing).
Also, let me know if it's already worth uploading to the file repository. This version may not be final, but I think it's good enough to start with.
Last, I'm really considering making a flower tutorial thread, though I'd like the mods' opinion if these forums are a good place to present 3d modelling tutorials before I'd do something like this.
-
Looks amazing.
Are any of these in game yet? Or is there some sort of conversation that's needed?
Lastly! I will soon have time to help you finish the waterfall
-
And that's a more than satisfyingly completed scene! Love the contrast!!! And texturing!
-
Can't wait to have it playable :P
-
Hey guys,
If you think these scenes are good enough to be released, I would appreciate if I could get some guidance in order to render the separate layers. SpooX mentioned earlier they can be retrieved with Palmer on the PC version, but I only have PS1 disks. Thanks!
-
Hey guys,
If you think these scenes are good enough to be released, I would appreciate if I could get some guidance in order to render the separate layers. SpooX mentioned earlier they can be retrieved with Palmer on the PC version, but I only have PS1 disks. Thanks!
7mimic I think ( look here (http://gamehacking.org/download#section15)) (thanks lazy bastard for the link) will give you the layer information for the playstation. It could also mess with the animation sequences. Also note that layers where animations where supposed to be or covered by something sometimes had green sections behind them. That's where 7mimic can be VERY useful.
I know a tool exists to create / modify new field location image data (it's been a long time). I would suggest you start with a 'small' render so it takes less time and you can see what problems you have with the encoding. Pallet data might be somewhat of an issue. I do remember their being a tool for it but for the live of me I cannot remember.
Stephen
-
7mimic is a fine tool to use, I use it to extract the backgrounds, however to get the different layers, you do need Palmer, you also need it to construct the layers again to be able to put it in game (mod dir).
Since the psx version is bound to hardware, it makes no sense in trying to get updated images inside, exept if we have some briljant genius who can port the whole game to the ps3....
8)
-
7mimic is a fine tool to use, I use it to extract the backgrounds, however to get the different layers, you do need Palmer, you also need it to construct the layers again to be able to put it in game (mod dir).
Since the psx version is bound to hardware, it makes no sense in trying to get updated images inside, exept if we have some briljant genius who can port the whole game to the ps3....
8)
Really it would make more sense to use Q-gears XML controlled media data. At least the format data can be tested. Hmmm I'll see oif Akari has some input on using it to test background data quickly.
Not sure how well the backgrounds work off hand in Q-gears.
The advantage of Q-gears is its "portable" to most anything. Flexible, and it mostly works (mostly). :D I almost compiled Q-gears but right now I have my system limping badly so one thing at a time.
I thought their was another tool than Palmer. as I said it's been a long long time.
-
Portable to everything?
Oh..no FFVII PC rewritten in Java.
-
Portable to everything?
Oh..no FFVII PC rewritten in Java.
If it has a LUA interpretor in it and you decide to rewrite all that C/C++ code sure. :D
Cyb
-
1.) Ambient/ Omni/ World/ Hemi (whatever lol) washes away a lot of details, I think this is what you mean by low contrast.
2.) Walls and Floor look to flat and just like shades of 2-3 colours.3.) Wood texture looks so the same on every object.
3.) Porcelane and lamp and flowers look very good. They are the most detailed objects but take the least space.
4.) Towel on table looks good, carpet looks flat.
-
Lighting: Kill the hemi and use fill lights ( turn of spec) to get rid of completely black shadows.
Or use a proper rendering engine like Cycles. Will get you much better looking shadows, glass and reflections. And hours to wait until a render is finished ;)
Texturing: Can´t really give any specific tips other than having simple UV´s help a lot. In Cycles you can resolve a lot in the shader directly, maybe take a look?
It is a nice looking scene. Would be a shame if you dont push it to the next level.
-
Thanks a lot for your feedback :) That will be helpful when I am able to revise this scene. Currently, I have annoying limitations because my laptop is so old and slow. As the scene gets complicated and the file grows bigger, many things become impossible. Mostly:
- I can't do any proper render anymore, even with a downsized render. So, to test all the lighting/textures (which I've mostly done by going back and forth between the original image and the rendered pic), it becomes unfeasible.
- I'm limited on polys, so I can't make enough subsurf to apply displacement on most of my textures.
Anyway...
1. Yes, that's what I meant by "low contrast", everything looks flat. The problem with ambient lighting is that it cancels bumpmaps (which are revealed by a preferential direction of the lighting). If I don't use ambient lighting, then I'm wondering how I am going to light the external side of the walls. Or do I have to combine 2 separate renders with different lighting conditions?
2. I relied on bumpmapping for the walls and the floor. I could ramp it up, and while decreasing ambient lighting, hopefully it will look less flat. As for the colors, maybe it needs more shades on the flagstones (do you recommend the use of a photo texture?), however I'm not sure I need more shades on the walls (after all, I would think the walls are painted), unless you meant to apply a more elaborate dirt map or something. As for wood textures, I did try to have some variations there, I can try to have other texture images to get in there.
3. Thanks. The vases were easy, while I went creative with the porcelain. I used special flower fonts to make the UVs :) The flowers took me some time, but I ended up with a decent method. I will make that "make a flower" tutorial some day :P
4. I'd like to make the carpet less flat, and the one in your scene would be very much like what I'd like to achieve. Did you use displacement on your carpet, or were bumpmaps enough? I used only bumpmaps, however once again the ambient lighting killed them.
Then... at this point I've stuck to Blender internal rendering engine. I've never used cycles - I don't think my current laptop can handle it (the GPU of my laptop is very poor). However I think it's worth looking into it, I'll do that.
-
1.) I suspect that you have relatively low RAM count if the scene doesn´t render at all. Having little CPU power is a problem as well, but it only forces you to wait longer.
Things you can do to get higher poly count: Disable raytracing ( might be impossible for a lot of scenes but you seem to use buffer shadows anyway) or use Cycles.
Having all even quads on the object you want to tesellate helps a lot as well ( I suspect it will only be the floor and maybe the walls).
Using .jpgs instead of .tgas or .pngs can help but they have less quality, however I find that hard to notice.
2.) I find the walls look better than the floor. Didn´t you model the cobblestones by hand? They absolutely don´t stand out at all. Maybe extracting them further out, and mixing in a grunge map in on the concrete already does the job.
That Omni light makes bumpmaps pretty useless, unfortunately. Is there AO in the scene, btw? It looks like it is either of, or set to add. Setting it to multiply will also give you more depth.
4.) They are strands.
Seriously, give Cycles more than a try. It handles heavy geometry much better, and the overall quality is just superior. With BI in scenes like that all you can do is is throw around a bunch of lights and literally hope for the best.
You don´t need a powerful GPU to use it. In fact I am using it with a an HD4250, and that thing is probably the slowest GPU you can think of. Just make sure to use a small window ( not fullscreen) if you switch to realtime preview.
-
Fabulous work there mate.
-
...
1.) Yeah, my laptop doesn't have a lot of RAM. That was a big limitation whenever I wanted to do some physics (could handle cloth sim, but not much smoke and fluid sim). Lighting creating soft shadows were also crashing. As for raytracing, I tend to prefer that over buffer shadows in order to render transparent materials, as raytracing enables me to do it with actual refraction indexes. Then... come to think of it, I don't think I have a single tesselated object on that scene. I didn't even try, else I know my laptop wouldn't have handled the necessary amount of polys.
Hopefully I should get a new PC (actually a rebuilt 4 years old pc) soon, which should help me with these problems.
2.) I did model the coblestones by hand. Though, after a fashion, I wanted them to be somewhat flat, thinking of practical matters (no one would like to have a cobblestoned floor in house, it'd be a pain to clean!). Still, as you mentioned, they look way to flat. The thing is: I originally calibrated the amount of bumpmap based on the results with hard-shadow lighting (because my laptop couldn't handle soft shadows), and that setting was ill adapted to soft shadow lighting (the bumps need to be magnified considerably, like x5, to give decent results). I forgot to change that in my previous renders. Then, I could also extrude them a little bit more. As for AO, actually you're probably right. I did put AO in the scene, but I don't know whether it's "add" or "multiply", I'll have to check. That last point was something I was not aware of, so thanks for the tip!
3.) There's no 3.)
4.) They're strands? Do you mean it as a Materials setting, or did you make hair strands based on a particle system?
Fabulous work there mate.
Thanks! I hope I can give you some motivation for continuing your awesome Midgar model, or make a field screen :)
-
Thanks! I hope I can give you some motivation for continuing your awesome Midgar model, or make a field screen :)
Sorry buddy, bad news i'm afraid. As I was going through all my stuff that I was able to salvage during a format it appears as though I backed up the wrong Midgar file I am pretty far back now but that's okay, it just means i've made room to perfect what needs to be.
-
Hi folks,
Just to show that I've started to work on mds5_w. I actually got a "new" computer, in fact I tried to repurpose a PC that was being disused in my work... unfortunately, it turned out that I have only a chipset acting as a graphic card, which is terrible. It has a bit more CPU and RAM than my laptop, but the GPU is so poor that it creates aliasing problems on the renders.
Anyway, back on the mds5_w scene (weapon shop in the slums, set up inside the what remains of a bus): I only have "dummy" textures at the moment, I didn't fiddle with that very much. I made some very basic lighting. I will have to look into that when I work with the "cycles" rendering engine - except that I don't think it's an option with my current GPU. Even though I think I started fine, there are a few issues on that scene that I'd like to discuss... so first, here's my current scene:
(http://img515.imageshack.us/img515/3563/wipmds5w.png) (http://imageshack.us/photo/my-images/515/wipmds5w.png/)
The original picture can be viewed at: http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_w.png (http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_w.png).
So... there are mostly 2 things that have been bothering me in that scene. First is the seat on the left hand side, near the back wall of the bus, where one of the shop clerk sits. This particular seat is actually considerably smaller that all the other seats on the scene. I thought of resizing it so that it matches the others, but it looks like the walkmesh in that area corresponds to the small height of the seat. So...
a/ I can carry on with a seat whose size matches the other, but we'll have to test how that'd look in game
b/ I keep the seat tiny, and we'll have to assume that a kid's seat is being used in that place.
The other part that concerns me is much worse: the bus driver seat, dashboard, etc. In the original picture, it's actually a complete nonsense. The seat is misaligned with the cabin, the height of the seat and its position would make it impossible to see the road when driving, it's a mess really. I made some corrections already in my version (the seat is closer to the dashboard, higher, and the steering wheel is more vertical, like in an actual bus), but if that's fine with you, I'd like to redesign that part quite a bit. I think the original picture from the outside of the bus (http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_1.png (http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_1.png)) could be used as some reference (in addition to some photos of real buses). A redesign of this section would not be a problem as far as the walkmesh is concerned.
Well, actually there's a third thing that may need to be corrected. There's some sort of metallic handrail+grid between the bus driver seat and a passenger seat behind it. In the original picture, this handrail/grid is in the middle of the walkmesh ??? I'm of a mind to put it on the edge of the walkmesh (moving it towards the front of the bus), especially since redesigning the bus driver seat section will move this part more towards the front of the bus, as well.
So... that's about it for the moment, some feedback would be appreciated.
-
It wouldn't make any sense to have a seat that is that dramatically different in size. Could you substitute it for another kind of seat that would logically be that size?
Also, although there is a walkmesh in that area, there isn't any characters actually walking in that area, correct? So would it matter as long as it looked correct with the sitting NPC?
-
Just to show that I've started to work on mds5_w. I actually got a "new" computer, in fact I tried to repurpose a PC that was being disused in my work... unfortunately, it turned out that I have only a chipset acting as a graphic card, which is terrible. It has a bit more CPU and RAM than my laptop, but the GPU is so poor that it creates aliasing problems on the renders.
you could use that old baby for different purposes as well, I just set up a NAS with Nas4free, with a low powered PC to have all my files in a central place and available from any computer I might be working on. :)
So... there are mostly 2 things that have been bothering me in that scene. First is the seat on the left hand side, near the back wall of the bus, where one of the shop clerk sits. This particular seat is actually considerably smaller that all the other seats on the scene. I thought of resizing it so that it matches the others, but it looks like the walkmesh in that area corresponds to the small height of the seat. So...
a/ I can carry on with a seat whose size matches the other, but we'll have to test how that'd look in game
b/ I keep the seat tiny, and we'll have to assume that a kid's seat is being used in that place.
could be a child-seat, I wouldn't worrie about the size difference.
The other part that concerns me is much worse: the bus driver seat, dashboard, etc. In the original picture, it's actually a complete nonsense. The seat is misaligned with the cabin, the height of the seat and its position would make it impossible to see the road when driving, it's a mess really...
Well, actually there's a third thing that may need to be corrected. There's some sort of metallic handrail+grid between the bus driver seat and a passenger seat behind it. In the original picture, this handrail/grid is in the middle of the walkmesh ???
About the other things....considering we are in the slums, where everything around is being used, trash is dropped there from the plates, so it is most unlikely the bus will ever drive away. A driver seat might be adjustable, the rail might be moved to a different place... who knows. I wouldn't worrie about it all too much.
Use the force your instincts Luke
8-)
-
Also, although there is a walkmesh in that area, there isn't any characters actually walking in that area, correct? So would it matter as long as it looked correct with the sitting NPC?
could be a child-seat, I wouldn't worrie about the size difference.
I believe one of the shop clerk actually sits there. I think I'll run with the assumption that it's a child seat and match the original image on that, then.
Use the force your instincts Luke
My instincts tell me that the original designer might have had too much sake on the night before handling that scene :P That, and the fact that the outdoor scene of the slums and this one were handled by two different artists. I think I'll re-arrange that section around the driver's place to something I like better.
As for this pc... it would have little use as a data storage unit, it's hard drive is about 100 Gb, while I also have a 1 Tb external hard drive. I think I'll just man up and invest into a good PC and be rid of most limitations (I'm thinking something along the lines of a i7 core with 12 Gb ram and an okay GPU, that should do it and remain below $1000).
-
As for this pc... it would have little use as a data storage unit, it's hard drive is about 100 Gb, while I also have a 1 Tb external hard drive. I think I'll just man up and invest into a good PC and be rid of most limitations (I'm thinking something along the lines of a i7 core with 12 Gb ram and an okay GPU, that should do it and remain below $1000).
May I recommend the one I bought from Cyber Power PC, except use a different seller for your purchase:
INVOICE NO. 00XXXXXX
====================
Invoice date: 2/28/2013
=================
ORDER DETAIL
============
Order date: 2/16/2013
Arrival date: 3/05/2013
Ship via: GROUND
ITEM DESCRIPTION QTY UNIT PRICE PRICE
=============== ===================================================== === ========== ==========
ID-INFO 380548 CFGI3Z 1 0.00 0.00
MODEL1 C SERIES 1 0.00 0.00
SW-152-154 AMD CRYSIS 3 + BIOSHOCK INFINITE 1 0.00 0.00
KB-159-101 AZZA MULTIMEDIA/ INTERNET USB KEYBOARD 1 0.00 0.00
MO-125-101 AZZA OPTICAL USB GAMING MOUSE 1 0.00 0.00
CU-208-223 INTEL I7-3770 3.4 GHZ 8M LGA1155 RETAIL 1 649.00 649.00
HD-504-701 64GB ADATA SP900 SATA III 6.0Gb/s SSD 1 11.00 11.00
HD-403-303 2TB SATA III 6GB/S 7200RPM 64MB CACHE 3.5" HD 1 112.00 112.00
RM-317-115 8 GB DDR3 PC3 PC1333 2 44.00 88.00
FA-WATER-101 ASETEK 510LC 120MM WATERCOOLER 1 0.00 0.00
FA-104-116 CASE FAN 120 MM 1 0.00 0.00
CS-157-519 BLACK THERMALTAKE COMMANDER MID TOWER NO POWER 1 -14.00 -14.00
FD-ASS-107 BLACK 3.5 MOUNTING KIT 1 0.00 0.00
CD-146-101 BLACK SAMSUNG 24X DVDRW 1 0.00 0.00
PS-121-109 CORSAIR CMPSU-750TX V2 POWER SUPPLY 1 71.00 71.00
VC-221-101 AMD RADEON 7970 3GB DDR5 PCI-E 3.0 1 402.00 402.00
SW-170-105 WINDOWS 7 PROFESSIONAL LICENSE 1 135.00 135.00
SW-170-106 WINDOWS 7 PROFESSIONAL COA 1 0.00 0.00
SW-170-114 WINDOWS 7 PROFESSIONAL SP1 64-BIT 1 0.00 0.00
BOX1 SYSTEM BOX AND FOAM 1 0.00 0.00
SERVICE-202 POWER SUPPLY GASKET 1 5.00 5.00
SERVICE-104 ANTI-VIBRATION FAN MOUNTS 1 9.00 9.00
MISC ONBOARD 7.1 SOUND 1 0.00 0.00
DISC SPRING0410 (5% Discount) 1 -75.45 -75.45
MB-399-110 GIGABYTE GA-Z77-HD3 INTEL Z77 CROSSFIRE SATA 3 USB 3. 1 41.00 41.00
0
=============== ===================================================== === ========== ==========
Subtotal 1,433.55
Shipping 75.00
Shipping Discount 0.00
E-WASTE Fee 0.00
Tax 0% 0.00
Total 1,508.55 USD
GUARANTEE POLICY
================
30-days money back guarantee.
No refund after 30 days.
For parts, there is a 15% restocking fee.
Shipping and handling fee is non-refundable.
Visit our website or call us for details on our policy.
-
May I recommend the one I bought from Cyber Power PC, except use a different seller for your purchase:
INVOICE NO. 00XXXXXX
====================
Invoice date: 2/28/2013
=================
ORDER DETAIL
============
Order date: 2/16/2013
Arrival date: 3/05/2013
Ship via: GROUND
ITEM DESCRIPTION QTY UNIT PRICE PRICE
=============== ===================================================== === ========== ==========
ID-INFO 380548 CFGI3Z 1 0.00 0.00
MODEL1 C SERIES 1 0.00 0.00
SW-152-154 AMD CRYSIS 3 + BIOSHOCK INFINITE 1 0.00 0.00
KB-159-101 AZZA MULTIMEDIA/ INTERNET USB KEYBOARD 1 0.00 0.00
MO-125-101 AZZA OPTICAL USB GAMING MOUSE 1 0.00 0.00
CU-208-223 INTEL I7-3770 3.4 GHZ 8M LGA1155 RETAIL 1 649.00 649.00
HD-504-701 64GB ADATA SP900 SATA III 6.0Gb/s SSD 1 11.00 11.00
HD-403-303 2TB SATA III 6GB/S 7200RPM 64MB CACHE 3.5" HD 1 112.00 112.00
RM-317-115 8 GB DDR3 PC3 PC1333 2 44.00 88.00
FA-WATER-101 ASETEK 510LC 120MM WATERCOOLER 1 0.00 0.00
FA-104-116 CASE FAN 120 MM 1 0.00 0.00
CS-157-519 BLACK THERMALTAKE COMMANDER MID TOWER NO POWER 1 -14.00 -14.00
FD-ASS-107 BLACK 3.5 MOUNTING KIT 1 0.00 0.00
CD-146-101 BLACK SAMSUNG 24X DVDRW 1 0.00 0.00
PS-121-109 CORSAIR CMPSU-750TX V2 POWER SUPPLY 1 71.00 71.00
VC-221-101 AMD RADEON 7970 3GB DDR5 PCI-E 3.0 1 402.00 402.00
SW-170-105 WINDOWS 7 PROFESSIONAL LICENSE 1 135.00 135.00
SW-170-106 WINDOWS 7 PROFESSIONAL COA 1 0.00 0.00
SW-170-114 WINDOWS 7 PROFESSIONAL SP1 64-BIT 1 0.00 0.00
BOX1 SYSTEM BOX AND FOAM 1 0.00 0.00
SERVICE-202 POWER SUPPLY GASKET 1 5.00 5.00
SERVICE-104 ANTI-VIBRATION FAN MOUNTS 1 9.00 9.00
MISC ONBOARD 7.1 SOUND 1 0.00 0.00
DISC SPRING0410 (5% Discount) 1 -75.45 -75.45
MB-399-110 GIGABYTE GA-Z77-HD3 INTEL Z77 CROSSFIRE SATA 3 USB 3. 1 41.00 41.00
0
=============== ===================================================== === ========== ==========
Subtotal 1,433.55
Shipping 75.00
Shipping Discount 0.00
E-WASTE Fee 0.00
Tax 0% 0.00
Total 1,508.55 USD
GUARANTEE POLICY
================
30-days money back guarantee.
No refund after 30 days.
For parts, there is a 15% restocking fee.
Shipping and handling fee is non-refundable.
Visit our website or call us for details on our policy.
8GB Ram? get rid of that 2 tera drive and just reuse an old HDD use the SSD for boot and long term programs. Go AMD and save yourself about 500 dollars since the graphics card is AMD. Also make sure the graphics card is on a card with a good heat-sink and fan. XFX has a dual fan setup. The Sapphires are cheaper but they get hot real quick and for the price difference I'd rather have the XFX. For rendering and editing I would grab an 8 core FX as they just dropped to about 170 bucks and you can run them at about 4.0 (which is lower than the quad-core, but the power gets distributed much better. grab at least 16gb ram, preferably like 32 to help out on render times.
Thats a quick breakdown, I just had to post considering that build posted before me is overpriced and underpowered, and that card is not worth 650 dollars. I'd also gram something with a 27in tall case to improve airflow but thats just me. If I had the know-how I could probl render a lot of shit for you guys and never put my PC under much stress.
I used to 3d model but have sinced lost my programs and I get frustrated with blender because I don't understand it lol.
if anyone ever needs anything done and can send me what program to use and what to do, I can try to lend some PC power, just pm me
-
The original picture can be viewed at: http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_w.png (http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_w.png).
So... there are mostly 2 things that have been bothering me in that scene. First is the seat on the left hand side, near the back wall of the bus, where one of the shop clerk sits. This particular seat is actually considerably smaller that all the other seats on the scene. I thought of resizing it so that it matches the others, but it looks like the walkmesh in that area corresponds to the small height of the seat. So...
a/ I can carry on with a seat whose size matches the other, but we'll have to test how that'd look in game
b/ I keep the seat tiny, and we'll have to assume that a kid's seat is being used in that place.
I suggest replacing the seat with an object like a box or something (the seat isn't used anyhow so... I wouldn't care if it looked like a box instead). You could make it a baby seat for a car instead for example :D
The other part that concerns me is much worse: the bus driver seat, dashboard, etc. In the original picture, it's actually a complete nonsense. The seat is misaligned with the cabin, the height of the seat and its position would make it impossible to see the road when driving, it's a mess really. I made some corrections already in my version (the seat is closer to the dashboard, higher, and the steering wheel is more vertical, like in an actual bus), but if that's fine with you, I'd like to redesign that part quite a bit. I think the original picture from the outside of the bus (http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_1.png (http://mirror.ff7data.com.ar/ff7/data/field/flevel.lgp/mds5_1.png)) could be used as some reference (in addition to some photos of real buses). A redesign of this section would not be a problem as far as the walkmesh is concerned.
Yeah it always puzzled me about that chair. Also no broken glass visible inside (heh).
Well, actually there's a third thing that may need to be corrected. There's some sort of metallic handrail+grid between the bus driver seat and a passenger seat behind it. In the original picture, this handrail/grid is in the middle of the walkmesh ??? I'm of a mind to put it on the edge of the walkmesh (moving it towards the front of the bus), especially since redesigning the bus driver seat section will move this part more towards the front of the bus, as well.
Well walking through a handrail does seem odd no? :D
So... that's about it for the moment, some feedback would be appreciated.
You mentioned an old machine and the graphics system on it. You may wish to uprade with a display card (I did on this machine cause the built in one was a 'wtf' do I need this for kind of thing) and disable to iinfernal err internal one. I have 8 gigs on this OLD machine (yep it's 3 years old) it's a quad core and runs OK dokey. I've been looking at upgrades recently myself, but I don't really want to talk about what my next machine will be (bigger is all I can say). All the machines at work have the built in display system disabled and have NVIDIA qudro cards in them because of the huge difference it makes in just doing simple tasks. Running Xp with 2G ram and a quadro card makes my work a lot easier (I do beat the snot out of the machine too I run several applications and compile stuff simultaneously).
Anyhow if you check what bus you have change your Vid card installed instead. That is a good upgrade you don't need a gaming monster card but a good display card IS priceless :D I rarely play games on machines. Save maybe oblivion (which is mostly because I can't mod it on the <censored> PS3 system).
Cyb
-
I am not taking part in this project so a bit off topic
but I still use a dual Pentium III 1260mhz that can take up to 4GB RAM.
-
I believe that in the original the seat is lower and the wheel has a different angle.
Also the square window has that slide thingy on the middle (2 glasses).
-
I believe that in the original the seat is lower and the wheel has a different angle.
Also the square window has that slide thingy on the middle (2 glasses).
About the drivers' seat and wheel: yes you are right, in the original picture the seat is lower, the the wheel is set at a different angle. I started to change that deliberately because when you model it, the original combination seat/wheel/dashboard is completely messed up (basically, a driver wouldn't be able to see the road). Modifying that area wouldn't be a problem as far as the walkmesh is concern, so I'll do that.
About the window: again, you are right. Since I posted the picture, I already modified that window quite a bit and arranged it with 2 overlapping sliding panels, which will fit the original description and be realistic.
As a side note, I also started to work on another field screen from the sewers (colne_be3), using Cycles as a rendering engine. I hope to post some pictures soon.
-
when you model it, the original combination seat/wheel/dashboard is completely messed up (basically, a driver wouldn't be able to see the road).
Maybe thats why its a weapon shop now
-
Great to hear Mayo Master.
I can only hope you find motivation to finish every scenario you start with, you have great skills.
-
Thanks for the encouragement :)
Well, the motivation is there, don't worry about that. I only wish I had more time and a better computer (hell, I don't have the time to get a better computer ::) ).
-
Hi folks,
Just letting you know: I've converted my "old" HD version of mds5_dk with Palmer and uploaded that to the WIP repo. It's 2 .png files that can be found under /Field Scenes/mds5_dk/. To make it work, they need to be placed in /mods/<modpath>/field/mds5_dk/. You're welcome to give it a try.
As far as this particular scene is concerned, I'll do an overhaul later as I plan to redo the lighting and texturing with the Cycles render. That's why I didn't upload it as a "Final" version.
-
Getting started on a new one, while I'm running the fluid sim for colne_b1...
(http://imageshack.us/a/img32/8754/8dbd.png)
It's only 50% res here, no point in making a super-detailed render if I got the scene only 3% done :P
-
Hi everyone and happy new year,
I've resumed working on my scenes during the vacations. I only hope to be able to maintain some kind of pace once the holidays are over.
Anyway, here's a progress update on mds5_1, I hope you enjoy it.
(http://imageshack.us/a/img856/9889/1vle.png)
-
Happy New Year Mayo! Keep it up!
-
Happy new year everybody!
Mayo that's looking really good, cant wait to see the lighting from the monitor and buildings put into that!
-
Happy new year, that scene is looking good so far. Is there any news on aeriths house?
-
Thanks guys!
By the way, I wanted to mention that the roofs in this scene are made with overlays of corrugated sheets, which have been modelled one by one - they're not just painted textures. You cannot see that at the resolution imageshack displays, but it shows in the original (the picture size is about 2500 x 2000). I've just modelled one of the flood lights, I'm also looking forward to seeing how the huge TV screen is going to light up the scene.
About Aerith's house: I intend to revisit all the scenes of the house (exterior and interior) with the Cycles rendering engine. It will require re-texturing everything, but I think it is worth it. The same goes for the other scenes of the sector 5 slums I made with the old internal rendering engine. In addition, I will layer them to have them playable in game. In a way, I would like to eventually come up with a complete package of scenes relevant to this area of the game. I just hope my regular work will leave me enough time for doing that :P
-
Tears of joy!
-
Ok, posting a progress update here. I've been working on the mess in the background, plus the roof tiles. I'll do some revisions on the textures later. I have yet to add lots of rubble, piles of junk and random pipes. I hope you enjoy (It's a 50% size render).
(http://imageshack.com/a/img585/5091/6jbh.png)
-
Your renders are always promising! Keep it up and thank you for your professional approach!
-
Hi folks,
Just posting a progress update here. I'm pretty much done with the modelling of the background (not the texturing though). The Blender file is getting pretty heavy. The render is full res. Enjoy!
(http://imageshack.com/a/img577/5488/cpv3.png)
-
Progress update...
(http://imageshack.com/a/img38/638/jowe.png)
-
Hi folks!
For my big scene of the Sector 5 slums, I have yet to make a number of elements which are actually involved in other scenes (shops, man in pipe, bus...). Thus, I thought it was a good time for me to revise these other scenes, namely to be able to render them with Cycles.
I am almost done with mds5_dk, aka "the man in pipe" or "this guy are sick" scene. Here's an update on this one:
(http://imageshack.com/a/img607/8171/k3tj.png)
Compared with the initial version using Blender's internal render, there are lots of improvements. Everything is retextured, and I can use indirect lighting, objects as light source (like the TV screen), etc. I'm happy it didn't take me too long to revisit that one. I have a couple of things yet to tweak: namely I want to make light interactions with the outer shell of the pipe, and I have to fix the lighting (not enough illumination as a whole, and I have to find a way to reduce noise - I used a meshed light bulb as a light source instead of using a point light source, which seems to create lots of noise). I may have to redo a smoke simulation, as well.
I hope you like it.
-
Now that's a nice scene!!! Excellent!
-
Your updates never fail to impress.
-
With the superbowl on, the last picture looks so ironic! :D
-
Thanks guys! Good to see you're still checking this place. TA has felt pretty lonely, of late.
-
I'm always watching... that sounded creepy. I am incredibly impressed by your artwork and love watching it progress. I would probably would have tried my hand at blender by now if it weren't for the obfuscating terminology and time/system constraints
-
Looking way cooler than before.
-
Hi folks!
I wanted to post an update on the big scene of the Sector 5 Slums, mds5_1. I did manage to import mds5_dk in it at the same scale, it's looking pretty cool. In addition to that, I modelled the outside of the rocket booster, and did quite a bit of texturing. I have also started to fiddle with the lighting, I could use some input there. The trick is: I don't believe there's a point in replicating the original lighting scene, because the original is off (in terms of colors and light sources). An example: you have some lights located inside chimneys. My aim is rather to convey the "feel" of the original, but with more realistic lighting.
Here's my light setup: I made a quick "dummy Midgar model" on a separate file, and made a render at daytime while setting the camera in a location corresponding to the Sector 5 slums. I took that render as an Enviro light source. The strength of this light source is fairly dim (this light source alone would only make a picture with RGB values around 10 or less). In addition to that I made an array of lights as if there was an array of lamps below Midgar's plate. That is providing most of the light in the scene. I intend to make the rest of the lighting with physical light sources (meaning: street lights, floodlights which are actually in the scene). I thought of inserting a Plate object way up, to create more light bounces, but it seems to have little effect on luminosity while introducing a lot of noise. If you believe I'm on the right track, or if you have suggestions, let me know.
Anyway, here's the image (full res at 200 samples)
(http://i.imgur.com/A3tB97m.jpg)
By the way, I'm not happy imageshack doesn't host stuff for free anymore :(
-
Nice work Mayo.
it's wonderfull to see the inside of the rocket from the other scene :)
With regards to lighting, Midgar is a funny place...
It's always dark and cloudy, no 'sunny' or daylight picture exists, except the one with Red looking from a distance..
So if you look at the top plate, it is already very dark. You can imagine when having a plate above you, at about 50 meters,
there's not a lot of light left. I would say use any trick at your disposal to balance between reality and fantasy. That would
mean, use any shortcut to lower noise and rendering-times ::)
As for imageshack, I switched over a long time ago to flickr, works fine for me.
Anyway, I hope you don't feel too lonely here, as I know, many are watching us.... :P
Keep it up, and enjoy as you go.
8-)
-
Nice work Mayo.
it's wonderfull to see the inside of the rocket from the other scene :)
With regards to lighting, Midgar is a funny place...
It's always dark and cloudy, no 'sunny' or daylight picture exists, except the one with Red looking from a distance..
So if you look at the top plate, it is already very dark. You can imagine when having a plate above you, at about 50 meters,
there's not a lot of light left. I would say use any trick at your disposal to balance between reality and fantasy. That would
mean, use any shortcut to lower noise and rendering-times ::)
As for imageshack, I switched over a long time ago to flickr, works fine for me.
Anyway, I hope you don't feel too lonely here, as I know, many are watching us.... :P
Keep it up, and enjoy as you go.
8-)
About seeing the inside of the rocket: when remaking these scenes, I really wanted to have consistency between the inside and the outside of the locations. I think that seeing the inside of places you can go to from the outside helps the immersion, and creating a believable environment. It's something I'd like to be systematic. Similarly, I'll import the other shops into the big scene. The weapon shop made inside the bus will be challenging, because the inside doesn't match the outside at all (in the original scenes). I'll have to make a bunch of revisions. The downside of this method is the size of the outdoor scene, which begins to be considerably huge. I think my object count is around 5000, and the Blender file size around 150 Mb :o (and I have no particle system!) I also intend to carry out something similar for Aerith's house.
About Midgar's lighting: it is definitely odd, and personally I tended to believe that all the events taking place on the top of the plate where going on during night time. There's pretty much no instance of Midgar getting daylight, even atop the plate. The only time you actually see it is when watching Advent Children :P That is, to the exception of Aerith's house and the Church. In my dream-remake, there would be variations in light to account for the passing of time, and getting all the Midgar scenes atop the Plate could get a very different feel if done during daytime.
As for the lighting of my current scene, based on what you mentioned, I believe I'm on the right track.
Oh, and I opened an imgur account :P
-
Anyway, I hope you don't feel too lonely here, as I know, many are watching us.... :P
Haha, well you guys do do some pretty awesome work. I regret that I haven't had the time to properly learn Blender and do some work myself. I wanted to last semester and ended up putting in extra time learning After Effects for a class.
That scene is looking awesome. If anybody doesn't agree wholeheartedly with your consistency idea, that last image should sell 'em.
As for imageshack, that's super lame. I had just used it last week and thought it was still free, but I just checked, and apparently I'm on a free trial that's just about up :(.
-
I regret that I haven't had the time to properly learn Blender and do some work myself. I wanted to last semester and ended up putting in extra time learning After Effects for a class.
Well, just so you know: we have some 650 scenes left to do, so it's never too late to start learning and help us :)
-
Hi folks!
Time for a progress update... namely, all the houses are textured now.
(http://i.imgur.com/B9MBCPX.jpg)
What remains to be done for that scene:
- Make all the broken asphalt
- Make the bus and tents (but that's making a whole new scene)
- Add debris and trash here and there
- Incorporate other scenes (item shop to be retextured, materia shop and random home to be made from scratch).
- Fine tune the lighting (I think I have the good locations for the lights, I have to tune the intensity a little).
I'll be probably moving to the other scenes to be incorporated in this one.
I hope you enjoy!
-
Great stuff, I look forward to being able to run around and check out the houses and shops here and really get to see the work put into this.
Well, just so you know: we have some 650 scenes left to do, so it's never too late to start learning and help us :)
Good to know lol. I'm trying to learn a bunch of stuff in Maya, mostly objects and weapons right now and character models soon. I'll see what I can do to contribute when I'm ready for something as advanced as this. At least do a few to really learn landscapes. People really underestimate how intricate landscape modeling is.
How are you guys going about textures? Are you working off templates or something to try to maintain consistency or is each artist doing their own thing?
-
This is just awesome, great work!
Cost aside, is there a clear cut advantage to using maya over blender or vice versa? The workflow's appear to look totally different, I assume the rendering engines are totally different as well
-
Asking me, Blender can be less intensive on your computer (depending on what you're doing at least) easier to pick up and learn, and of course is free. Maya is more difficult to start out with and learn all the tools, but it's incredibly powerful and intuitive. It also has plugins for many other applications like photoshop and zbrush, so transferring your work back and forth is very easy. After starting with Maya I couldn't imagine using anything else. You can get the student edition for free if you have a student email account, doesn't matter what school or major.
-
Good to know lol. I'm trying to learn a bunch of stuff in Maya, mostly objects and weapons right now and character models soon. I'll see what I can do to contribute when I'm ready for something as advanced as this. At least do a few to really learn landscapes. People really underestimate how intricate landscape modeling is.
Well, whenever you feel like modelling field screens, we'll be happy to see your contribution. When that time comes, I recommend you check out my shortlist of "Scenes for newbies" that I drafted in the "beginner eager to learn" thread. They're mostly small indoor scenes, I would not advise taking a landscape scene right away. SpooX and I know that nature modelling is really challenging. Besides, it's also among the most demanding in terms of computational resources (particles systems everywhere and crazy poly counts, among other things).
How are you guys going about textures? Are you working off templates or something to try to maintain consistency or is each artist doing their own thing?
So far, everybody has been doing their own textures. However, there's a bit of consistency because most of our textures come from cgtextures.com. That being said, SpooX and I were just speaking of sharing resources: it would be indeed much better if we pool textures and maps. At the moment we have no organization whatsoever regarding this topic, maybe it's time to get started. The file repository may be a good platform for sharing. In addition, we could benefit from the efforts of 2D artists who wouldn't be involved in 3d modelling.
-
Hi folks,
Time for a progress update. I've spent some time re-texturing mds5_i (item shop of Sector 5 slums) so that it would be re-rendered using Cycles. I've done most of it, as you can see on the picture (I left out the objects that were not yet textured):
(http://i.imgur.com/7Rfu6wN.png)
Things left to do: a piece of furniture on the bottom left corner, some jars, a medkit, a shelf with various objects on it, and a bunch of cardboard boxes.
Additionally, I've done two other things on that scene:
- I revised the till, namely rescaled it. Originally it was way too big, compared with the size of the characters. I had to revise all the cables attached to it. I also moved it because the way the till was set wasn't easily accessible.
- I made the import of that scene in the main outdoor scene of the Sector 5 slums. It can be seen when opening the door of the item shop (not much is seen beyond some wood boards and the doormat, actually).
-
Looks really pretty! ;)
-
Hi folks,
I'm pretty much done with my re-texturing of the item shop. Here's the almost final render (it's only with 50 samples, I'll ramp that up considerably for final renders), I should be able to layer it and make it playable soon.
(http://i.imgur.com/Raz1EYD.png)
I hope you like it!
-
I like it!
Looks very nice :)
Though that pipe's shiny line part kinda bothers me a bit :P (uhh.. just my opinion) Forget I said that :-[
-
Though that pipe's shiny line part kinda bothers me a bit :P (uhh.. just my opinion)
Well... it's a big brushed steel tubing directly located under a light source, so it will have a "shiny line". The variable that may affect that shiny line is mostly the "Roughness", which is relevant to the roughness of the brushed steel finish. A coarse roughness will give you a wider "shiny line" and more blurry reflections, a small roughness will give you a much more narrow shiny line and sharper (mirror-like) reflections. When I started texturing, I actually put less roughness, but I changed this variable after comparing it with the original picture (see page 3 of this thread), where a relatively coarse roughness is shown. We could have a long debate about choosing small roughness over big roughness (or the opposite), I prefer to submit a picture to figure which one is preferred by the majority. I may revise my textures later accordingly (it's rather a quick fix, it doesn't bother me).
Below is a picture with the non-painted part of the large horizontal tube set to a much lower roughness (a quarter of the roughness of the previous picture).
(http://i.imgur.com/0sj6mSV.png)
Let me know what you guys think.
-
I think it looks amazing, the piping looks almost photo realistic. My only suggestion would be to make the tarnish on the smaller pipe visibly consistent across the entire area of tubing. It seems a little too pristine for a musty dirty place like the slums
Or perhaps the store owner polishes his makeshift store counter, whichever way you want to rationalize it
-
I think you definitely had it right the first time. The shine had a settled look to it with the more coarse roughness that would fit a store in the slums. I think the second one loses it's realism with how it's reflecting the light.
-
(http://img.photobucket.com/albums/v287/Akabi/Itemshop.png)(http://i.imgur.com/Raz1EYD.png)
I would like a tone down on the highlights overall, right now it has more the feeling of polished chrome, instead of oxidated steal.
not only on the main horizontal pipe, but also the cylinder ar the back, and the reddish pipes on the right top.
Also the white of the poster feels too white, one note in CG is that there aro no absolute colors (rgb 255,255,255) best is to always keep well below that.
make it dirty, make it dirty....that's the challenge :-P
but that's just my humble opinion :)
8)
-
Forget I said that :-[
Don't be embarrassed about your opinion. Feedback is useful.
Or perhaps the store owner polishes his makeshift store counter, whichever way you want to rationalize it
That was was pretty much my take on it.
I think you definitely had it right the first time. The shine had a settled look to it with the more coarse roughness that would fit a store in the slums. I think the second one loses it's realism with how it's reflecting the light.
I am not sure there's a "more" or "less realistic" case. Today I have seen a steel pipe of mirror-like finish on a construction site, which itself was pretty dirty. The roughness is primarily a consequence of the metal finish itself, much less that of dirt and grime (which however affect the intensity of the reflections).
I would like a tone down on the highlights overall, right now it has more the feeling of polished chrome, instead of oxidated steal.
not only on the main horizontal pipe, but also the cylinder ar the back, and the reddish pipes on the right top.
I've done quite a lot of observation of metal ducts IRL, so I can try to give you an educated answer :P
I don't think the metal you see in my picture is too polished for being some kind of brushed stainless steel. The brushed stainless steel I showcased has significant roughness and anisotropy, I believe the scale at which we view the image (the main duct is about 80 cm in diameter) may give a false impression. Chrome usually has pretty much no roughness nor anisotropy, with very sharp mirror-like reflections (usually it's more with a smooth, electroplated finish, and not brushed with grit belt like stainless steel).
The other aspect is: stainless steel does not stain or rust under regular conditions. If it is corroded, you would see an oxide with a rather light purple color. It behaves a lot differently than cast iron or low grade steel (which would rust with the classic orange iron oxide). So in a sense, if it is not subject to harsh environments (such as acid rains, saltwater), stainless steel can only get dirty with whatever is smeared on it, not so much because of the degradation of the material itself. But that is just one side of the discussion.
When it comes to these cases of texturing, I believe the main problem is simply a matter of interpretation. What do we see in the original image? Basically we see brighter sections of grey metallic-like pipes and dirty-orange sections of pipe. There are actually various types of materials which could give that impression at low resolution, but once you get at much higher resolution and you see much details, you can end up with very different results.
1- My own interpretation was that the pipe was made of brushed stainless steel, which was painted in the more orange sections. I ruled out the assumption that the orange color came from rust, because of the area of the pipe which is painted with the "Welcome" kanjis. The main horizontal pipe is like a makeshift counter, which the shopkeeper would clean more regularly than the rest of the pipes. The pipes are only dirty because of grime, not because of rust.
2- One could interpret this as as assembly of aluminum (where the pipe is shiny metal) and rusted iron. The "Welcome" kanji painting could have had some corrosion-resistant coating, who knows.
Aluminum is fairly shiny, but generally its diffuse/glossy ratio is more than that of stainless steel, so it could have given a different result on the grey sections. If the orange is rusted cast iron, the resulting high res texture would be very different from what I've shown.
3- The pipe could be made with galvanized steel/orange-painted galvanized steel. Galvanized steel is also very corrosion resistant, but its diffuse/glossy ratio is quite higher than that of stainless steel, and it would certainly tone down the highlights on the metal. This interpretation could give a result much closer to what you suggest. Yet again, galvanized steel has a very particular grain to it.
Personally, I have to understand how the material is made of to try to replicate how it would look like. Based on your feedback, I can have a go on interpretation 3- and see what I'd end up with. I don't think interpretation 2- is valid (again, based on the painted kanjis).
Also the white of the poster feels too white, one note in CG is that there aro no absolute colors (rgb 255,255,255) best is to always keep well below that.
You're right, I hadn't taken care of toning this one down. I'll make the revision.
-
Hi folks,
After your feedback, I decided to have a try at re-texturing the pipes based on the idea that they were made of galvanized steel and painted galvanized steel.
Additionally:
- I put a bit of grime on the poster and toned it down
- I changed the texture of the shelf
- I put more reflectivity in the tea cups
- [EDIT: Not shown on the picture below] Also revised a glitching texture on the piping with the two steel holders
Below is the result:
(http://i.imgur.com/l5K7KT9.png)
I'd appreciate your feedback on this.
-
...galvanized steel and painted galvanized steel (https://www.youtube.com/watch?feature=player_detailpage&v=Xu3FTEmN-eg#t=106).
;)
-
I thought the first one looked great, the new one is just awesome. Lighting and texture on the pipe looks really good.
-
Getting started on 5min1_1....
(http://i.imgur.com/baplqqg.png)
-
Hi folks,
I wanted to start doing a bit of texturing on 5min1_1, but I could seriously use some advice/feedback.
Here's the original image...
(http://i.imgur.com/dtRQ6gG.png)
I'm actually done with modelling most of the pipes. To be honest, it was a fest of non-Euclidian geometry and escher-style perspective, more painful than enjoyable thing to model.
Now, I would like to start texturing the floor and the walls to get a better sense of the feel of the place. However, based on the original image, I have a hard time figuring out what the walls and floors are actually made of, and what kind of textures I should assign to it.
Here are some basic tests I made for floors:
(http://i.imgur.com/i1ZlOED.png)
1- very old and worn tatami-like flooring fabric
2- fabric flooring
3- fine concrete floor
4- coarse concrete floor
It's not so much about making a definite choice for one of these four options, but I'm rather looking for pointers. I tried to model the whole floor with texture 4 but it looks off somehow. If you have some ideas, I would really appreciate. :)
-
I like the first texture. I'm not sure if it's 100% match but I like it. I thought of that kinda floor when I look at the original picture.
-
(http://i.imgur.com/l5K7KT9.png)
I'd appreciate your feedback on this.
Astonishing job! :o Congratulations!
About the choice for the floor texture. The one I feel like excluding is number 3. Fabric looks good and while number 1 looks extra nice I feel like number 2 might give the best result.
-
Hi folks,
I wanted to start doing a bit of texturing on 5min1_1, but I could seriously use some advice/feedback.
I'm actually done with modelling most of the pipes. To be honest, it was a fest of non-Euclidian geometry and escher-style perspective, more painful than enjoyable thing to model.
Now, I would like to start texturing the floor and the walls to get a better sense of the feel of the place. However, based on the original image, I have a hard time figuring out what the walls and floors are actually made of, and what kind of textures I should assign to it.
1- very old and worn tatami-like flooring fabric
2- fabric flooring
3- fine concrete floor
4- coarse concrete floor
It's not so much about making a definite choice for one of these four options, but I'm rather looking for pointers. I tried to model the whole floor with texture 4 but it looks off somehow. If you have some ideas, I would really appreciate. :)
For me, it feels like none of the above.
Tatami is not used here, (too expensive for the slums).
fabric, not likely, too much moist.
Concrete also looks to be out of the question.
(http://www.cgtextures.com/thumbnails/textures/wood/Plywood/Old/PlywoodOld0025_1_thumblarge.jpg)
(http://www.cgtextures.com/thumbnails/textures/wood/Plywood/Painted/PlywoodPainted0039_1_thumblarge.jpg)
(http://www.cgtextures.com/thumbnails/textures/wood/Plywood/Painted/PlywoodPainted0041_1_thumblarge.jpg)
Personally I would look more into the direction of worn-painted-plywood.
In some places it looks like the floor is curled up at the end of the plates, might be due to moist.
Also the paint is worn more at the ends of some plates.
that's my two cents. :)
8)
-
I think the flooring kind of resembles wrestling mat material ???
-
Looks like concrete wood to me...
Here are some examples:
(http://upload.wikimedia.org/wikipedia/commons/8/89/Wood_grain_German_Occupation_bunker_concrete_1.jpg)
(http://www.textureimages.net/uploads/6/1/2/6/6126732/3814500_orig.jpg?896)
(http://upload.wikimedia.org/wikipedia/commons/9/92/Wood_grain_German_Occupation_bunker_concrete_2.jpg)
(http://[url=http://images.kaneva.com/filestore7/2673721/4708004/Concrete.CastUinUPlace.Formwork.Wood.Boards.bump.jpg]http://images.kaneva.com/filestore7/2673721/4708004/Concrete.CastUinUPlace.Formwork.Wood.Boards.bump.jpg[/url])
(http://0.design-milk.com/images/2010/06/joey-roth-ff-concrete-wood.jpg)
(http://caccstudiov.files.wordpress.com/2010/09/wood_grain_german_occupation_bunker_concrete_311.jpg)
(http://www.panespol.com/blog/wp-content/uploads/08-panespol-beton-tabla-640x355.jpg)
-
So... I guess I'll just go and try different textures on a larger scale and see what I end up with.
- I had several (out of the forum) people saying right away it's old worn tatami fabric. It could be very well some "upcycled" tatami that a wealthy resident from above the plate dumped before. It also concurs with the "wrestling mat" material idea.
- I wonder if the classic idea of slum houses having mold and humidity problems holds in Midgar. The Plate is above, and should have all sorts of drainage system, making the area under the plate never receive any rain. I would believe things get grimy because of dust and dirt, primarily.
- I'm adding plywood to the list of possibilities.
- As pointed above, I don't think I can rule out concrete because a/ the patterns and ridges are rather concrete-like b/ concrete is a building material that is very commonly used in slums (I've seen real life examples in Mexico City).
I'll just run along with these ideas. It's going to be time consuming, but there's no helping it, is there.
-
(http://caccstudiov.files.wordpress.com/2010/09/wood_grain_german_occupation_bunker_concrete_311.jpg)
I would be satisfied with this one though :)
-
Ok folks, I did some more tests...
Concrete test
(http://i.imgur.com/wkLAjGX.png)
Plywood test
(http://i.imgur.com/nhnzZbh.png)
Old tatami test
(http://i.imgur.com/4hyddov.png)
Well, it may not be very obvious on these pictures because of some sources of noise (having all walls and ceiling generating indirect light creates quite a bit of noise), but old tatami is actually the winner. That my final word :P
Coincidentally, it turns out the room is almost exactly 8 x 4 tatamis, based on traditional Japanese length units (http://en.wikipedia.org/wiki/Tatami). Now I just have to work on making an auspicious tatami arrangement (http://webhome.cs.uvic.ca/~ruskey/Publications/Tatami/TatamiCOCOONsubmitted.pdf). :D
-
If you could please make your tatami a little bit worn out :P :P
-
Update!
(http://i.imgur.com/obMlFUK.jpg)
The flooring has old tatamis, pieces of plywood and dirty concrete. All should be happy :)
I think I may add a bit more specularity/glossy to the square pipes. By the way, the TV (I've yet to model it properly) has an 80" screen :o Must have "fallen from the truck". I'm assuming it's an old fashion "giant flat(-ish) screen" (the largest CRT TVs were about 40", and needed 4 people to be carried around).
I hope you enjoy!
-
If you compare it to the original the tatami blocks look bigger (not like small squares), more grayish and not in a tiled sequence, but anyway you prefer it, should be fine. If you can change this btw it would be more awesome. Thanks!
-
Well, what you mentioned could open some discussion regarding how to model scenes...
When I started to model for TA, the goal was pretty much to make the scenes HD, and try to exactly replicate how it used to be in the original. My point of view has significantly changed since then. It turns out that if you try to simply upscale a scene as you remodel it, it just doesn't work fine. The main reason is a problem of scale. Many items in the original images were scaled in such a way to make them recognizable at the very poor resolution we had on PS1 era games. But if you keep things at the same proportions as the original and make it HD, it just looks dumb. For instance, the original scale of the tea cup in the item shop scene is 38 cm in diameter and 40 cm tall. When something looks absurd, there's no saving it on the reason that "it's set in a fantasy world anyway". For having believable environment, it's important for us to correct all these clunky details as we remodel things.
Additionally, as we remodel things in HD, we have the ability to bring in lots more details which also contribute to make the world more believable. I disagree strongly with the simplistic idea of enhanced graphics being detrimental to story-telling: if anything, enhanced visuals can tell more without a word. For example, in The Last of Us, just when entering random houses, you could start imagining what kind of people were living there, and recreate their lifestyle, just with the amount of details given to the environments.
All in all, when we remodel a scene, we have to interpret the original, and build something coherent from there. In the particular example of the 5min1_1 scene, after I decided to go for the old tatami floorings, I sized tatami according to the traditional Japanese standards. I laid out the tatamis according to the rules for auspicious arrangements, I don't think Japanese people would have done otherwise. Then, to keep all the tatami together in the area filled with clunky pipes, I either cut out some of the tatamis to have enough gaps for these pipes, or I inserted pieces of plywood like a slum resident would for an easy and cheap fix. If I had tried to exactly replicate the original, I would have ended up with oversized, awkwardly shaped tatamis, and I don't think that would have worked better. Don't get me wrong, I want to be respectful and faithful to the original design, but I want to work on ways to improve it, when it is possible. It is also the main advantage of our HD remake of field scenes over the "enlarge and photoshop filter" method you have in Omzy's graphical overhaul.
-
You have put lots of serious thought into it Mayo and that's what makes your modifications special and recognizable by the community! :) About the tatami I thought it was made like this on purpose in the original because of the industrial effect in this area. In such an imaginary reality big tatami blocks with no certain order with some concrete grayish moldy scent in it, would make more sense to me in that type of room... Well this is just a thought that you don't need or have to take seriously into consideration, especially if it also delays the progress of other scenes.
-
I read that there wasn't any 80 inch CRT tvs.. and I was like really?! lol I never realized my huge crts were never larger than like 43 inches....
I was thinking maybe they have hung a 80 rear projection television on the wall :P
though maybe not
http://reviews.cnet.com/8301-33199_7-57557233-221/rip-rear-projection-tv/ (http://reviews.cnet.com/8301-33199_7-57557233-221/rip-rear-projection-tv/) here is one of the last rptvs. i guess it was 92 inch.
http://reviews.cnet.com/projection-tvs/mitsubishi-wd-65737/4505-6484_7-33588548.html (http://reviews.cnet.com/projection-tvs/mitsubishi-wd-65737/4505-6484_7-33588548.html) review of last rptv as shown above... some side profile images and such. :P
-
Update... More details, namely finishing the TV, adding a VHS player and some VHS tapes... and adding all the cables with it :P
(http://i.imgur.com/DafFKjH.jpg)
-
Please add an old PSX console. It would be a lovley allusion.
-
Please add an old PSX console. It would be a lovley allusion.
Working on it ;)
(http://i.imgur.com/HeJYovr.png)
-
Missed the old design! :)
-
Great! :-D
But don't forget Shin-Ra is the only company in this world. So it's a Shin-Ra PlayStation.
Missed the old design! :)
Me, too... :'(
-
Will the TV be a light source that affects other parts of the scene when this is done?
-
Ok, the PSX model is done and UV unwrapped. I'll do the texturing tomorrow.
(http://i.imgur.com/3DWtSpa.png)
Great! :-D
But don't forget Shin-Ra is the only company in this world. So it's a Shin-Ra PlayStation.
I had that in mind as well. I think I'll have fun with the logo.
Will the TV be a light source that affects other parts of the scene when this is done?
Actually, this is already the case. It isn't really noticeable because the room is already brightly lit, so the little extra light from the TV doesn't make much of a difference. If I ramp up the TV light, then whatever is on screen will be saturated with brightness. For reference, I made a quick render at small scale to give you an idea of the lighting from the TV alone. I think it's fairly realistic.
(http://i.imgur.com/y65C1Qr.png)
-
... and DONE :D
(http://i.imgur.com/7xDKps7.png)
(http://i.imgur.com/FEP737h.png)
That was fun to make. I might have gotten a little carried away regarding the details :P
-
Name it MakoStation!!! :P :P
-
Wow I just nostalgia"d pretty hard, if you plan on making a controller to go with it remember that analog sticks weren't a thing until after ff7 came out I believe
-
Wow I just nostalgia"d pretty hard, if you plan on making a controller to go with it remember that analog sticks weren't a thing until after ff7 came out I believe
I do remember that (well, that was the setup I was playing FF7 with in the first place). Besides, not having to do the controller sticks will make my modelling job easier :P
-
Ah, neat console (and backgrounds too!) Looks like you guys are making good progress on the early Midgar area.
-
Mini-update with the PS1 added to the scene with controller, memory card and cables.
(http://i.imgur.com/QUb9TSf.png)
-
You can place a FFVII case somewhere there as well ;)
-
You can place a FFVII case somewhere there as well ;)
No, this would be strange somehow, but Xenongears is mentioned in the game by Cloud.
-
Here's my personal take on it...
(http://i.imgur.com/jKwbLob.png)
-
Blehhhh the EU version ;)
Lookin good
-
You could place an imaginary title though that would have to do with something we are anticipating from Square in the future!
-
A cookie for the person who'll get the reference from the game box ;)
Anyway, here's a quick update (50% res).
(http://i.imgur.com/iTOU9ch.png)
-
Looking good as always ;)
A cookie for the person who'll get the reference from the game box ;)
You mean this? (Log Horizon much?)
(http://i.imgur.com/JVkTBZPl.png)
-
Yup! Here you go!
(http://1.bp.blogspot.com/-94XK-zMZLFI/Tk5ekLtwDgI/AAAAAAAADs4/dEjtNbQsvXc/s400/Cookie.gif)
-
Very quick update... (20% res with not many samples)
(http://i.imgur.com/sXcvzfe.png)
Actually, the scene is nearly done. I need to model and texture a magazine, do the textures for dishes, tea kettle, candle, lighter and a couple of other things. I think it should be finished by Easter.
-
Hi folks,
Can anybody help me out there? It's probably about something simple, but I can't figure it out. After my last change of hard drive, I needed to re-install FF7 (I got the 2012 PC downloadable version) and then re-convert it to the 1998 version using Kranmer's game converted, before I could test my mods. However, after conversion, I can't launch the game (Please Insert disc message) despite I had a USB drive renamed for that purpose (it was working before, why doesn't it now?). If I could have some help to fix that, it would be much appreciated.
-
make sure the datadrive registry entry is set to the same drive letter that your USB drive is appearing as.
-
Thanks! I passed the CD check after changing the registry (BTW did I also need to change the registry for movies?). However, it still fails to launch, now for another unknown reason :(
EDIT - I kind of fixed the mess. I was trying to use the v 0.8 of Aali's driver, it didn't work. I managed to get things to work after switching back to v 0.7.11b. Hopefully I'll get back to testing my mods soon.
-
Update: Getting started on mds5_m (materia shop):
(http://i.imgur.com/IgoXHQp.png)
Two things regarding 5min1_1:
- I am missing two textures to complete the scene (a celebration card and a poster hanging on the wall), I hope they'd come my way sometime this week.
- I had to make some revisions because the original scene had a problem. Because the game is set as 3d characters walking on a 3d mesh arranged with 2d layers, the perspective of the 2d layers has to match the whereabouts of the 3d characters. In this particular scene, one problem was a table which would reach above the character's knees when walking on one side of the table, but the same table would be waist-high when walking along another side. I revised the table dimensions (making it more narrow, actually) so that the height would look consistent (by matching the perspective). These sorts of problem are only apparent when testing the fields in-game. Some revisions of this type may be needed in a good number of scenes.
-
GR8!
-
Hi folks,
I'm having more delays for the remaining textures I need for 5min_1. Nothing really urgent though, it will all come together eventually.
In the meantime, I carried on my work on mds5_m. Here's today's update:
(http://i.imgur.com/R0a6TlZ.png)
For some reason it's going to be much less noisy than many other scenes, which is good. I may have to add a bit more dirt on the floor planks. Also, the lighting setup was tricky, because the illumination of from TV in the original scene is completely off. I had to recreate something that would give a good match on the overall lighting, but using extra light sources (the effect of the tv lighting itself is not very obvious, beyond the shelf where it sits).
I hope you like it.
-
Update on the materia shop...
(http://i.imgur.com/0DbZEui.png)
A couple of comments: It may not seem like there's still a lot to do, but the toughest part of that scene has yet to be done: bonsai trees! Otherwise: please don't mind the region of the picture with lots of noise, it's the combo glass object near glossy objects which causes that - it'll be remedied in the final renders. As usual, I scaled the objects in the scene properly. As a result, it may not seem like there's as much trash in that scene as there is in the original. With that in mind, there are two possible options: 1- I could increase the number of trash items laying around, to enhance the feel of a dirty place 2- I could leave it as it is, because after all the shop clerk (a heavy smoker, no doubt) should be a minimum concerned about housekeeping.
On another note: the plastic strips by the back exit may look odd: this is because nothing (there is only a strip of grey untextured floor) is yet modeled on the outside of the shop. I'll use some objects from the exterior scene, as well as a special enviro light to address that problem.
Lastly, I was wondering about filling one of the glass jars with a bunch of materias - this being a materia shop after all. It's not in the original scene, but I think it would be relevant AND look cool. The only unknown is about the amount of noise generated by the corresponding material (I'm guessing the material will be an association of glass plus volumetric emission).
Comments, feedback and random blurbs are welcome, it's getting lonely here.
-
I gotta say, that scene is looking great. I'm really digging the wood textures for the floor.
My main criticism at this point is the lighting. I'm only amateur at best when it comes to modeling, so I don't know how I would go about it exactly, but the lighting around the tv looks off compared to the original picture. For instance, in the original, there's a shadow caused by the tv on top of the fridge and in front, but you have a shadow behind it.
I like what you've been doing in your scenes with regard to scaling. This is a case where I find myself not entirely on board. The pop cans seem extremely small with reference to the fridge in the corner. If you can make a scene have more parity with the 3D models, I'm all for it. However I think parity within the scene itself should take precedent.
All in all, pretty f'ing awesome.
-
Thanks!
The wood floorboards were really a case where I wanted to convey the original feel, but making something a bit different. Specularity and bumps seem a bit exaggerated in the original scene.
Regarding the TV lighting: it's a point I mentioned in a previous post. The original lighting actually fakes what would have come from the TV, and it's really off. In particular, the lights and shadows atop the fridge are actually inconsistent in the original picture. Why? Because you can see a bright lighting and a sharp shadow on the fridge as if a light was coming at a 90 degrees angle from above, while the TV is actually located right above the fridge and should be projecting light straight ahead of itself. With a bit of experience in modelling scene, sometimes you realize that the limitations they had in the past made them do a lot of compromise regarding modeling and lighting. It's up to us to see when it makes sense to replicate the original, and when it doesn't.
Regarding rescaling: the final "test" when it comes to rescaling is when you play the scene in the game. Now that some modders have made a great job at making character models with much more balanced proportions, rescaled objects appear a lot better not only when compared to other features of a scene (such as a door frame), but when compared to the characters themselves. For instance, if I play with Kaldarasha's models in the original scene, I'm pretty sure the soda can would be nearly the size of a character's head. I'll double check when I test the scene in game, but a good point here is about making the characters evolve in a believable environment. By the way, the dimensions of the fridge are not absurdly big compared with the soda cans (I think the fridge is about 1.70 m in my scene). Also, I'd have to ask you exactly what you mean by "parity" - I'm a little confused. Do you mean matching the original picture, or to make objects at the same scale across different scenes, or something else entirely?
-
I'm just in awe of your work... seriously, it looks so good!
Forgive my ignorance but how did you become so proficient in 3D tools? Was it all self-taught?
-
Thanks!
Yes, my abilities in Blender 3d are essentially self-taught, I started 3d modeling in August 2012. However, as I explained in this post (http://http://forums.qhimm.com/index.php?topic=13786.msg195178#msg195178), I used a good wikibook and some good tutorials for getting the hang on this. I still regularly check out tutorials online, mostly when I need to use a specific feature (namely physics, particle systems) where a lot of settings are not immediately obvious. I think you can get through my progress as a modeler just by browsing this thread, although my earliest work was here (http://http://forums.qhimm.com/index.php?topic=13786.msg195178#msg195178). I believe the big steps I made are really related to texturing - in the beginning I had no idea about how to do it. In that respect, switching to Cycles render was also a huge step - although I had to figure out how to do proper UV-unwrapping too.
It would be great if my work would motivate some noobs into picking up Blender and do stuff! There's a long way before we get anywhere near the 650 scenes of this game :|
-
OK - I'm impressed. Particularly by the PSX model.
Really great to see such attention to detail and faithfulness to the original's aesthetic.
-
Thanks! I hope the bonsai trees I'm doing for this scene will live up to that.
-
Update: bonsai!
(http://i.imgur.com/Z5eiYsh.png)
This is actually a 300% sized render, compared with how it would look like in the actual scene. I used the best method I could for having something that looked ok without spending an ungodly amount of time on it - in the final image it will fit within a 85 x 85 pixels square. All the branches were made manually (which was very tedious) and I added the leaves using a distributed particle system. It's a chestnut tree, by the way.
On another note, some other scenes in game feature bonsai trees, I'll be happy to provide this one as a reusable resource. The mesh is clean (all the junctions between branches are cleaned), so I think it can be exported fairly well, too.
-
Bonsai part 2!
(http://i.imgur.com/sivs8yv.png)
This time, instead of making the bonsai manually, I used Blender's plugin called "Sapling". The plugin generates parametric trees. I think the results are pretty awesome, it's very simple and quick to use, and one can achieve a large variety of results. I don't think I'll make trees manually branch by branch, ever again (it took me 4-5 days to make the previous bonsai, and barely more than an hour fiddling with the settings to make the second one).
Well... the scene is almost done, actually. Yay :D
-
Yes, my abilities in Blender 3d are essentially self-taught, I started 3d modeling in August 2012. However, as I explained in this post (http://http://forums.qhimm.com/index.php?topic=13786.msg195178#msg195178), I used a good wikibook and some good tutorials for getting the hang on this.
(snip)
Thank you - I'm reading through that wiki book now. If I can get half as good as what you make here I'll be a happy man. :)
Good job on the bonsai too!
-
I wish you the best in your 3d modelling adventures! When starting, I think patience is the main virtue, taking the times to get through the baby steps.
If at some point later you feel like trying out a field screen, let me know!
-
Update: I incorporated the indoor stuff of the materia shop into the main outdoor scene of the Sector 5 slums.
(http://i.imgur.com/1AceoEr.png)
There were some inconsistency issues to solve, since the house as seen from outside is actually much smaller (about half the size) of how it is laid out in the indoor scene. This is the best I could manage for the materia shop, I think it's fairly decent.
Now, I have another problem. I decided to go back to the weapon store scene, which is supposed to be located inside a "refurbished" bus. The whole scale of the scene is complete nonsense, let alone trying to match how it looks like from outside with how it's supposed to be inside. I realized, after making some preliminary tests in game (i.e. comparing Cloud's height to objects in the scene), that the height and width of the bus would be about 5 metres :-o (basically, it's assuming that the top of the crate where the kid sits is about 2 metres high). So I'm really wondering about how to turn that environment into something believable. While the height could correspond to a double-decker bus (which is absolutely not how it looks like from outside), the width is way to big for any regular road vehicle (it's about twice the size of a regular bus). So... suggestions are welcome, because I don't quite know how to get started. I could try to keep to the original idea, but some elements will be ridiculously sized. I could try to transform the scene completely and have a matching exterior, but it's pretty much like creating something from scratch. All in all, it looks like this mds5_w scene is going to be a huge pain >:(
-
Have you ever tried my community flevel project? I had rescaled the models to fit better to the backgrounds.
For mds5_1 the model size is 576 (previous 512).
And mds5_w has a model size of 640 (previous 512).
The only solution for the inconsistent of the scenes I could imagine, is to make the models smaller for mds5_1. But I fear the doors will look then a bit to huge. I have no idea how much it is possible for Makou Reactor to alter the walkmesh for it. It has the ability to do it, but the question is if it works.
Anyway all what I wanted to say is that we are not (entirely) stick to the size of the models.
-
Have you ever tried my community flevel project? I had rescaled the models to fit better to the backgrounds.
For mds5_1 the model size is 576 (previous 512).
And mds5_w has a model size of 640 (previous 512).
The only solution for the inconsistent of the scenes I could imagine, is to make the models smaller for mds5_1. But I fear the doors will look then a bit to huge. I have no idea how much it is possible for Makou Reactor to alter the walkmesh for it. It has the ability to do it, but the question is if it works.
Anyway all what I wanted to say is that we are not (entirely) stick to the size of the models.
Well, no, I've never tried. And thank you for mentioning, because that's going to be a life saver! Honestly, if I can have characters resized for mds5_w, then most of my troubles are over. If all characters were bigger, then it would mean that, in comparison, I could get the interior of the bus rescaled to reasonable proportions. We'd have to go by trial and error, because if I just get all character scaled by 1.5, I'm not sure how everything is going to look like. Still, it is by far the most accommodating option :)
I would not resize the character much for mds5_1. The walkmesh is what it is, and doors are already much larger than need be.
-
Update!
So... I've mostly worked on trying to make a compatibility of environments between mds5_w and mds5_1, and it's pretty much figured out. I initially tried to piece the walls of the bus together using what I originally made for mds5_w, but I wasn't happy of the result (among other things, the topology was really messy). So I decided to take the basic dimensions of the bus and remake it from scratch - which took a bit of time. I'm just done with the body of the bus (not textured), and I wanted to share that with you.
(http://i.imgur.com/DlEUAzM.png)
-
It's all coming together now 8)
Oh man am I excited to play this scene
-
Bonsai part 2!
<IMAGE REMOVED TO not be redundant>
This time, instead of making the bonsai manually, I used Blender's plugin called "Sapling". The plugin generates parametric trees. I think the results are pretty awesome, it's very simple and quick to use, and one can achieve a large variety of results. I don't think I'll make trees manually branch by branch, ever again (it took me 4-5 days to make the previous bonsai, and barely more than an hour fiddling with the settings to make the second one).
Well... the scene is almost done, actually. Yay :D
Bit late but LOL that is funny. Now you might be wondering why I would say that. When I was actively using POV (back 10 years ago) I did the exact same thing. I first made a pine tree by "hand" actually I used a random perturbation calculation along a spiral for each "branch" and scaled the branches (you may now wake up). However this was tedious (even too me) so I found a fractal tree maker that did the same thing in POV by a function call.
The nice thing about parametric tree creation is you can add a bit of randomness for each tree you create so they all look just a bit different.
The down side is the object data explodes. However you are rendering to 2d so the amount of data is all the same (LOL).
Congrats you aren't the only one who has done this (IE don't feel bad).
Update!
So... I've mostly worked on trying to make a compatibility of environments between mds5_w and mds5_1, and it's pretty much figured out. I initially tried to piece the walls of the bus together using what I originally made for mds5_w, but I wasn't happy of the result (among other things, the topology was really messy). So I decided to take the basic dimensions of the bus and remake it from scratch - which took a bit of time. I'm just done with the body of the bus (not textured), and I wanted to share that with you.
<IMAGE removed for redundancy sake>
Oi looks good.
begin brain ramble
With all these details it would be interesting to see it in a preview like one would in Oblivion/Fallout3/NewVegas/Skyrim. Granted that is only partially possible (I did say partially). As you can only insert data into those systems programatically that preexists in the data base for the game. You would also have issues with Vista information for the area (what does the surounding area of midgaard look like relative to sector 5 slums giant pizza in the sky?). Lots of technical difficulties with doing such things.
end brain ramble
I've never been able to "get into" using blender sadly because I learned to make models programmatically (IE snythesize everything instead of using and making chunks of image data to paste on things). It only has certain advantages (for example the models in FF7 look pretty good irregardless the resolution they are displayed at for a similar reason), in that it is far more scalable. It is quite a bit more work for simple results however (LOL). Hence likely it was not a popular method in the end to do things.
Back to the image at hand, you know I never really paid attention to what the bus looked like in game. I've played through it a few times too. Hmmm my guess is the details weren't enough for me to "care". I tend to focus on something that has any detail in games when viewing.
Cyb
-
So I'm really wondering about how to turn that environment into something believable.
Switch off your brain. Making sense out of FFVII's physically impossible BGs is a waste of time. The stairwell in upper Shinra Building, for example, is impossible. The way the stairwell goes up would cause the each subsequent floor to have an overbite. I know because I tried to model it as a continuous model for the Mako Dawn mod. You don't notice because you never compare the BGs side by side. No one besides developers will notice — you're just hyper concious of it because you're making it.
Unless you're planning on solving all continuity issues, my suggestion would be not to worry about it and focus on making it look nice, not logical. Unfortunately, that'll occasionally mean recreating tweaking / things even when they're supposed to be part of the same object.
-
Do we have the ability to make/modify walkmeshes? That's obviously taking greater liberties with the game than this project was ever meant to, but it's pretty much necessary if you want the world to actually fit together properly.
-
Makou Reactor has some possibilities to edit the walkmesh, but it seems to be a bit buggy and uncomfortable. But the option is there...
-
Ah! Scaling and consistency is almost an endless topic. So...
Switch off your brain. Making sense out of FFVII's physically impossible BGs is a waste of time. The stairwell in upper Shinra Building, for example, is impossible. The way the stairwell goes up would cause the each subsequent floor to have an overbite. I know because I tried to model it as a continuous model for the Mako Dawn mod. You don't notice because you never compare the BGs side by side. No one besides developers will notice — you're just hyper concious of it because you're making it.
Unless you're planning on solving all continuity issues, my suggestion would be not to worry about it and focus on making it look nice, not logical. Unfortunately, that'll occasionally mean recreating tweaking / things even when they're supposed to be part of the same object.
I don't think that making sense of the backgrounds is a waste of time. I agree that making a perfect matching of all the connected scenes with extremely realistic proportions is not possible, however. Nevertheless, I do believe in the importance of ensuring consistency in these environments, because having characters in a believable surrounding helps having the player immersed in the world (suspension of disbelief, if you would call it). With that in mind, I'll always do my best to make the environments look logical, or at least "as logical as they can possibly be".
Now, with these criteria, there will always be some compromise to be made (as you mentioned by "tweaking things"), and I have already made a few. For instance, the houses of the outdoor scene of the Sector 5 slums are coarsely undersized (particularly when compared to the indoor counterparts of the shops). Hence, when trying to do the outside/inside matching, it can't work perfectly on a 1:1 scale. The compromise I made in that occasion was to just cut out pieces of the shop that wouldn't fit in the outside scene (it visually works because, from the outside, you only see a small fraction of the indoor environment). For the bus scene, there are some details I am going to change (namely, the driver's seat location). And there will be some other design changes along the way. I also believe it makes sense to take some liberties from the original design, since scaling everything up on the same proportions as the original would create a bunch of absurdly sized objects (mostly smaller objects, like a tin can, a soda bottle, a tea cup, the steps of a staircase, etc).
All in all, I think it is worth to put a lot of thought into scaling, redesigning, and what I would call "reconciliation of discrepancies between environments".
One other thing: thanks to Kaldarasha, I am now aware of what can be done about the character sizes using Makou Reactor, and that will be a huge help for solving most scale inconsistencies.
Do we have the ability to make/modify walkmeshes? That's obviously taking greater liberties with the game than this project was ever meant to, but it's pretty much necessary if you want the world to actually fit together properly.
Makou Reactor has some possibilities to edit the walkmesh, but it seems to be a bit buggy and uncomfortable. But the option is there...
Given the extremely remote possibilities of walkmesh modifications, I have based all my designs on the constraint that the walkmesh cannot be changed. I will keep on working that way until the day someone ends up with a bug-free in-game test. Even if the walkmesh can be altered and help to solve consistency issues, I would rather go for small edits, not something that would drastically change the whole scene (or at least, not in the near future). That being said, I'll try to keep an eye on the progress made in that department (or, if the person willing to test walkmesh alteration would keep me informed, it would be greatly appreciated :)).
-
Update!
This past week, I have kept on working on the outdoor scene of the Sector 5 slums, mds5_1. I did not finish the bus yet (mostly the texturing, adding wear, rust and dust). I spent some time modeling the broken asphalt in the scene, until I was happy with the result. It was quite tricky - the objects are very large, so I mostly relied on a combination of procedurals to make all the cracks.
Below is a render with only 50 samples (hence it's a pretty noisy), but 100% resolution (click on image to see the high res):
(http://i.imgur.com/MwJgB0J.jpg)
So... there are about 4 things to do to complete this scene:
- revise the packed dirt texture for the ground where the asphalt doesn't show
- Make a transition in the house on the right to lead to min5_1 indoor scene (I don't think I'll make a straight import, the scale inconsistencies are too great).
- finish the bus, and the hanging canvas (cloth simulation)
- tune the lighting (it's not bad, currently, but needs a few tweaks)
I also thought about adding particle systems to litter the scene with tiny gravels, debris, trash etc. But I think I'm not too far from the limits of this computer. The scene is very, very heavy at this point (about 200 Mb, and roughly 4000 objects, without particles). I'll have to see what I can do.
-
- tune the lighting (it's not bad, currently, but needs a few tweaks)
Does this involve adding a whole lot more lighting in general, or at least brightening what's there? The lighting in any given part of this image is good relative to the rest of the image, but on the whole it's a lot darker than the original.
-
That's true, the brightness level is currently quite lower on my scene than it is on the original image. Lighting is actually rather complex a topic, and we could have an open debate about it.
One thing I started to observe as I was modelling various scenes was the following: generally, the amount of lighting in the slums scenes is fairly high in the original images, and if I try to replicate the same amount of brightness in my remodelling, it's not "working" perfectly. Mostly, it feels the scenes are overexposed, and brightly lit scenes do not convey the "slums feel" very well. Often it seems too saturated. Lighting in original images is also inconsistent at times: very often you would find a light source illuminating an object with a yellow color, while the same light source would have a blue glare associated to it.
My current take on lighting is: use original brightness level as a rough guideline, rather than something to be replicated or perfectly matched. What mostly matters is to convey the "feel" of the place. If the lighting of the scene is significantly modified with respect to the original, the lighting of the character models can always be tuned using Makou reactor. That being said, I am open to suggestions.
On that particular scene: I would like to increase the overall lighting level a little, because the overall brightness is a bit too low when compared with the original. However, I won't try to replicate it. I also made some basic tests already where I just increased the intensity of my enviro lights, and the result wasn't really good. I think that what I should aim for is to have a relatively low enviro (or ambient) lighting and multiple smaller brighter lights, which would be more appropriate for describing a place where there's little difference between daytime and nighttime, and most of the lighting is provided by man-made light sources (as opposed to sunlight). More specifically, I would like the rubble in the top-right corner to remain rather dark - it's meant to be a messy rubble which shouldn't draw much attention to the viewer (even if I spent a lot of time working on this area, it's absolutely not the highlight of the scene). Then, in the original image there are many other "cone-lights" which only have a volumetric effect (i.e. a visible beam of light), I thought of having extra lighting from these sources. There are also some lights in the "tunnel" (for lack of a better word) below the massive tv screen) which are currently missing in my scene, they'll be added. I'll also probably ramp up the intensity of some of the light sources that I already have. The lighting intensity from the tv screen and the four blue lamps on the left hand side is already fairly high, I don't think I'll change these ones. Overall, I also think that if I can have many objects casting multiple shadows, it would help giving the right impression. That's pretty much what I'll be working towards - if you have some suggestions please let me know.
-
I mainly brought it up because at least on my work computer (2012 27" iMac, default color settings) it was by and large too dark to see most of it. Virtually all the detail behind the bus was completely lost.
-
Ok, that's good to know. On my monitor (27" ACER) I can make out details and nuances on low RGB values (even if it looks very dark) - oddly enough, the brightness is at the setting of 15 on a scale of 100 (on max setting, I could probably illuminate my entire apartment with it :P ).
By the way, if you have that sort of brightness issue, can you let me know if the brightness is fine on your end for the colne_b1 scene (http://forums.qhimm.com/index.php?topic=14810.msg207821#msg207821)?
-
I quite like the lighting, perhaps a little brighter in the right side, but not much. I am a fan of the slums being dark and dreary... they are slums after all.
-
The biggest problem I see are the buildings in the center. They look very small (somehow unimportant) compared the things around of them. I currently counter check it to the original and will make some notes. The darker lighting looks correct for me, the image is usually darker on PSX. Speaking of lighting Aali's driver can't replace the light layer(s) correctly. You can replace the light, but the transparency can't be changed dynamically by the game.
Here are some suggestions of me:
(http://s1.directupload.net/images/140603/jmd2dcpn.jpg)
I would try overall to merge the buildings with the tower and put some pipes through the roofs.
-
Interestingly, it looks a lot lighter on my laptop... which highlights the points Kaldashara has made. However they were not so noticeable on my rig and moniter where the buildings seemed more dominant due to the lighting. Don't know what to add to that, but there are my observations for you! :)
-
By the way, if you have that sort of brightness issue, can you let me know if the brightness is fine on your end for the colne_b1 scene (http://forums.qhimm.com/index.php?topic=14810.msg207821#msg207821)?
I can't get the preview to show up (dammit Imageshack).
-
That bus looks a bit shinny for that area...
-
My goodness! So much feedback all of a sudden! I almost feel overwhelmed!
So... first:
I can't get the preview to show up (dammit Imageshack).
Try this (http://i.imgur.com/vMANlUE.png), freshly uploaded to imgur. If you feel like you cannot make out much of the details in the pile in the bottom of the image, it's pretty much normal - I don't think it's meant to be a brightly lit scene. However, if you cannot see the main features, then I'll be concerned. By the way, I also advise you try to look at these pictures when surrounded by a black background on the screen (as it is in game), instead of a white background (as it is on this forum). It makes a real difference in how your eye can perceive it. If you're curious, colne_be3 has a brighter lighting. If you can test these fields in game and see for yourself, that would be the best kind of test (both scenes are layered and everything).
I quite like the lighting, perhaps a little brighter in the right side, but not much. I am a fan of the slums being dark and dreary... they are slums after all.
I agree. I may brighten the scene a little, but it won't be much, especially in that top-right corner behind the bus (it's not meant to draw attention to the viewer, besides).
That bus looks a bit shinny for that area...
You can only guess how much texturing work I'll have to do :o
The biggest problem I see are the buildings in the center. They look very small (somehow unimportant) compared the things around of them. I currently counter check it to the original and will make some notes. The darker lighting looks correct for me, the image is usually darker on PSX. Speaking of lighting Aali's driver can't replace the light layer(s) correctly. You can replace the light, but the transparency can't be changed dynamically by the game.
Here are some suggestions of me:
I would try overall to merge the buildings with the tower and put some pipes through the roofs.
Well... you're asking for a huge revision here. You should have given me this feedback when I started, 6 months ago ;) As far as scaling is concerned: the buildings in the centre are matching the original very closely. I agree that they look small, for that matter they are much smaller than their "indoor" counterparts. Beyond the amount of time I would need to spend on these kinds of revisions, I am not sure I could even make some of them them because of layering constraints (I don't have control on how Palmer cuts some of the layers). Walkmesh constraint may be a hindrance if trying to make larger buildings, too. I would have to do some testing on these constraints before trying to change these dimensions.
Overall about scaling: most of your suggestions would be good in principle, but their implementation is not straightforward. They would also be a significant deviation from the original image - I would probably be ok with that, but I don't know how it would be perceived in general. Also, to be honest I wanted to be able to deliver that scene around end-of-june/early-july, and that won't happen if I try to implement these revisions! I would probably have more enthusiasm about more work on that scene if there were some 20 people actively working on the project.
About the other details: often it's a matter of preference. I think the chimney cap on the foreground is fine as it is, and the design of the straight "Cargo" rooftop was actually deliberate (I wanted it to be an "upcycled" piece of a container for boat transportation, and dimensions had to make sense). As for the bent plates of the rooftops, I could do that. I agree it would look better, and I perfectly know how to do that (make a bunch of extra loop cuts and use lattice modifiers). I am not sure how much I will do though, again because of how time consuming and how tedious that is.
Honestly, I will probably put most of that on the back burner. I would prefer to be able to wrap up a 1.0 version of that scene and move on (fair enough, there's a certain amount of quality required of version 1.0, and I have to deliver that). Still, I'll do some testing.
-
Yeah, it's too dark on this screen. I can't tell what the water is, among other things; it looks like there could well be another floor under there, or a void, or anything.
I tried running both fields through photoshop. The sector 5 image was still darker than the original, but bright enough to see. The sewers actually had a brighter platform above, but the water underneath was just barely visible.
As an aside, they both look pretty good on my home computer monitor. I don't know how well calibrated an iMac is out of the box, but I'd imagine pretty decent.
-
The sewers on both scenes have very murky water. I could probably tuned that in a better way now, because 6 months ago Blender didn't have volumetric absorption shaders (so I had to fake the murkiness by blending a glass-like shader - for water - and a diffuse shader for the mud-like appearance). That being said, The bricks underneath are barely visible - for the fluid simulation to run properly, the water was needing a non-negligible height to just flow, so there's between 15 and 20 cm of sewage there, which would realistically yield a fairly opaque result.
By the way, re-rendering these sewer scenes would take a while, since I would need to render 8 animation frames for each scene. I believe I still have a backup of my fluidsim files, though.
Then, as a whole, what would be your recommendation? Do you think increasing the brightness on my renders is necessary, or do you think it's rather up to the players to adjust the brightness of their monitors? To be sincere, I think I have a fairly well-tuned monitor on my end, so I am unsure if it's up to me to make the adjustments. I could try a better match with the original images on the overall brightness, but as I said, I tend to find corresponding renders a bit over-saturated and over-exposed (especially for slums scenes).
-
Well, they're fine on my screens at home, which I know are properly calibrated because I did it myself. My big concern is that the average user will be running every piece of software and equipment at default settings; if a game does not have any gamma/brightness correction features, you cannot assume they will be able to see at all in particularly dark or light areas.
I think it might be easier but just as effective to brighten them up a tad in Photoshop, particularly with those time consuming sewer scenes. The results aren't quite as perfect as a fresh render, but when I tried it they were still pretty good quality (http://i.imgur.com/HmMJThg.png) as long as you didn't go overboard. And as for Sector 5, it can be brightened significantly while remaining darker than the original, a good balance between the two. I really do think you're right that they're better off darker than the original, but it's still better to be safe and put a lower limit on how dark you're willing to go.
-
I checked out your brightness correction by Photoshop on colne_be1, it looks pretty good indeed. You may also want to double-check colne_be3 (http://i.imgur.com/6BCp9yY.png). Overall colne_be3 is brighter, so a brightness correction may not be necessary, but I prefer to let you judge of this. There's actually a technical reason behind the lower brightness of colne_be1: I had to tune some render settings (namely, add a "cap" on specular values on the specular light passes) to avoid some persistent fireflies (for that matter, the reflections on the water of colne_be3 are a bit too saturated, because there was no such cap on the specular light passes). If that's fine with you, I can send you all the layers of colne_be1 (and colne_be3 if the situation warrants) for you to tune the brightness. Then they'd just need to be combined in Palmer.
I'll be coming to you for feedback on the Sector 5 main slum scene when I'm working on tuning the lights, then.
-
colne_be3 could use some tweaking on the ladder and railing; additional contrast wouldn't be accurate from a realistic lighting perspective, but it could definitely do to stand out more. The overall light level isn't going to be an issue on this one though.
I did save my settings for colne_be1 to a PSD just in case, so I'd be glad to apply the same settings to all of the layers. It's a little more complicated than a simple brightness tweak; I actually did one uniform minor adjustment with the levels tool, and then a stronger adjustment just to the water and a few elements that I felt needed to stand out better.
I've gotta say though, aside from the issues with the brigtness (which even then are minor and don't apply to properly configured monitors) this is all some real good work here. Certainly better than I could do (which, therefore, is why I'm not really involved with TA).
-
Regarding colne_be3: I'm not sure how I could make the railing and the ladder stand out more without adding an extra light source... for the railing, I can't see of a "relevant place" where I'd put one. I don't really think it is necessary, since the layering will make the player notice it's there. For the ladder, I could imagine another light source from the top, since the top of the manhole leads to the train graveyard. Still, I would keep these modifications for later (when I have the time to revisit this scene, which is not a priority), unless you can phototshop such a light source in the scene.
How would you like me to send you the layers of colne_be1? You could PM an email where I'd send them, I could upload them to the file repository, upload a .7z archive somewhere (or upload the bunch of them to imgur)... just let me know. Even if you're not really involved with TA, I appreciate the help :) Feedback is a good form of contribution.
By the way, after Kaldarasha's list of suggestions, I took a look at how Palmer is cutting out layers for this scene. Actually, there is quite a bit of room for modifications. I'll leave that for later, but I won't discard the possibility. For now, I'll try to bring the viewer's attention to these buildings in the centre with how I arrange my lighting setup.
-
It's easy with Photoshop (http://i.imgur.com/E63Upgc.png). My chief concern is that it does start to sacrifice realism when you make tweaks like this without actually modifying the lighting. I dunno if it's enough anybody would notice something's wrong, but it's worth noting.
My email is covarr, gmail. Send me stuff however's easiest for you.
As an aside, I checked out this mac's calibration. I must've calibrated it before at some point; the gamma setting was already at 1.8, lighter than Mac's default 2.2. If I was getting nearly-crushed blacks and not enough contrast, I'm even more sure than before that other people are bound to have the same problem.
-
Regarding the retouches with Photoshop: you don't have to be that concerned. As a matter of fact, in the world of CG, the use of post-processing techniques (such as Photoshop) is not only common practice, it is actually recommended. I have limited tools and skills when it comes to photo manipulation, so I personally tend to keep post-processing to a minimum. However, post-processing is very useful, and is used in a wide variety of situations. Some examples:
- Get a better balance between your light passes (for example, if you want to emphasize a specular pass over a diffuse one, or vice-versa).
- Color grading
- Mimic a variety of camera lens artifacts (sun glare, chromatic aberration, lens flare, depth of field, vignette, etc).
- Denoising
In Blender, you can actually apply a bunch of post-processing method in what is called the "Compositor", but if you make similar effects in Photoshop, it doesn't make the end result "less valid". As long as the post-processing effects still look realistic and are somewhat subtle, there's nothing wrong with that. For that matter, the glares and soft blurs around the lamps in both sewer scenes are made through the compositor, you cannot have this effect from a raw render. I agree you have to be careful when applying post-processing techniques, but given what you showed me previously, I'm confident in the end result.
So... well, I just sent you all the layers, by the way. I would even have a request: if you could please see about softening the harsh specular reflections on the water in the colne_be3 layers, that would be a welcome improvement. I'll be looking forward to seeing your results.
As for the lighting, then we're on the same page :)
-
I dunno what I can do about those specular reflections. They're simply too harsh; there's no recoverable color data there, and I don't have the talent to completely redraw that. But other than that, I've got all the layers done. Sending 'em back right now.
-
Thanks for the files. I just recombined them through Palmer and tested them in game:
- For colne_be1, everything is fine
- For colne_be3, there's actually a small problem. After double-checking, it seems like there's some sort of weird aliasing, which is most likely due to the fact that the image is not displayed at full-scale (HD is actually 4.5 times the original, while my renders are 5 times the original image). The problem exists for both non-retouched renders and your retouched image. In this situation, I think it is best to use the set of images where there's non contrast enhancement on railing and ladder, because it makes the aliasing very subtle. On your set of retouched images, the aliasing "pops" to the eye, and it's not looking very nice. One possible cause could also be a problem of alpha values management in Palmer when combining the layers, but I don't know enough about Palmer to confirm this.
I heard there was an issue if the image resolution was a non-integer multiple of the original, which is why I was always choosing "5 times the original resolution" instead of 4.5 times (for each dimension, of course). I've never tested the problem, and I don't know where the issue is from. If somebody could fill me in, that'd be much appreciated.
-
The biggest problem I see are the buildings in the center. They look very small (somehow unimportant) compared the things around of them. I currently counter check it to the original and will make some notes. The darker lighting looks correct for me, the image is usually darker on PSX. Speaking of lighting Aali's driver can't replace the light layer(s) correctly. You can replace the light, but the transparency can't be changed dynamically by the game.
I would try overall to merge the buildings with the tower and put some pipes through the roofs.
Let's have a closer look at the buildings.
(https://farm6.staticflickr.com/5479/14427664431_3812b31fde_o.png) (https://flic.kr/p/nYVxZ2)
Looking at the lines of the walkmesh compared to the background, it looks pretty close.
The one small difference I notice is the little shed behind the circular well (?), it is a bit taller on the right side, but should not really matter.
With regards to the lighting
(https://farm4.staticflickr.com/3916/14451274183_35713b5b34_o.png) (https://flic.kr/p/o21ykH)
As I see it, there are still some light sources missing, that would light up the whole scene a bit more here and there.
Specularity...well like you said, still a lot of textures to go ;)
Hope this helps a bit, overall I like it growing as it is.
8)
-
Thanks for your feedback, SpooX :)
Regarding the mds5_1 scene, since the last update I made the texture of the dirt ground, which was fairly complicated to pull off. Now I would like to move on to the details of the bus, but in order to do that, I'll pretty much have to work specifically on mds5_w scene (yes, there still will be the concept of "seeing the inside of the bus from the outside scene).
In the meantime, anaho latest update on his scene inspired me to do something very useful... as some may know, Blender just included volumetric scattering shaders in their latest release. Thus, I've made an update on mds5_dk (aka "The man in pipe", aka "This guy are sick"):
(http://i.imgur.com/AS7iyE5.png)
The great thing is the new ability to convey the feel of a dusty place where the air seems "thick", which is perfect for all slums scenes. Not only that, but it is exactly what was needed for increasing the overall brightness of these scenes without increasing light intensity, and also lower the saturation in the process. Thus, I think I'll go ahead and apply this method for all slums scenes. The downside is obviously longer render times :P
-
Update! I included a touch of volumetric scattering for the mds5_i scene (item shop). The number of samples isn't too high here, meaning it can be increased to minimize the remaining noise.
I wanted the "dust" effect to be very subtle, a "foggy" shop wouldn't make sense in my opinion.
(http://i.imgur.com/GkqsEBP.png)
-
Looks amazing! :)
-
Looks wonderful!
-
Micro-update.... details on mds5_1
(http://i.imgur.com/jddfaJJ.png)
Apologies to all the followers for the very low productivity of the moment... mostly because of the world cup :lol: (that's about 4 hours a day I am not spending on Blender :P ). Hopefully my modelling should get back on track in a couple of weeks.
-
Apologies to all the followers for the very low productivity of the moment... mostly because of the world cup :lol: (that's about 4 hours a day I am not spending on Blender :P ). Hopefully my modelling should get back on track in a couple of weeks.
lol, no worries, there is some breathing space untill friday...
8-)
-
Hi folks,
Quick update on the main scene (mds5_1) where I am tuning my lighting setup, and have included a volume scattering shader to create a dusty feel. (this is a 50% size render)
(http://i.imgur.com/2RQqD5l.png)
I think I'm getting close to my final lighting setup, however there's something which is bugging me. The original scene has a few light beams (namely, blue light beams in the top right corner) which are not on the "lights layer", they're baked into the environment if I may say so. I tried a few things to replicate these light beams (volume scatter cones, lights in post-processing like paintshop) but so far results have been poor (mostly, it looks blatantly unrealistic). There's one last thing I would like to try, which is to overlay some "cone lights" with the halo from the Blender Internal engine on a pre-rendered scene made with Cycles, although preliminary results have not been all that promising.
Would anyone have any suggestions regarding this detail?
At this point, if I cannot get a satisfying result when trying to recreate these light beams, I'm thinking of ditching them altogether (I think the scene can still look good without these).
-
So just for fun I had a go at adding the light beams in in photoshop. 3D modeling may still be beyond me, but I like to think I'm alright in 2D. IMHO I think the top right light beams came out alright, but the cone on the right and the blue area on the left definitely aren't great. Obviously being a test render I didn't want to waste a whole lot of time on it, but I thought I'd give it a shot. Yay? Nay?
(http://i.imgur.com/4kibURV.jpg)
-
Most of the original backgrounds had been edited in an image editor. My problem is that the design of ealin(?) is entirely against Aerith statement that the church is the only place where flowers grow in Midgar. Also from where is the lighting coming?
-
For reference
(http://img3.wikia.nocookie.net/__cb20080913145853/finalfantasy/images/c/cb/Sector5.jpg)
I think what you have looks pretty good, the rays of light must be shining through cracks in the upper plate. The light rays also show how saturated with dense particles, the air is down there.
The only criticism that I have, which is really splitting hairs at this point because that render looks awesome, is that the 'light bulbs'
in the original give off a radiation sort of feeling or like there is fog around the light sources.
-
The only criticism that I have, which is really splitting hairs at this point because that render looks awesome, is that the 'light bulbs'
in the original give off a radiation sort of feeling or like there is fog around the light sources.
Correct me if I'm wrong, but I believe that the glow around the bulbs and the rays of light are essentially the same idea, which is volumetric lighting. It's something that's only recently been introduced to Blender/Cycles in the form of Volumetric shaders. I don't think it gives true volumetric lighting, but there are ways of faking it I think, which is what I believe Mayo is struggling with.
Again my knowledge of 3D is spotty, but I think that's the gist of it. Once I get a job and can afford to upgrade my rig, then I'm gonna invest some time in actually learning this stuff. :wink:
-
Cool!
-
Thanks for the feedback, people!
So just for fun I had a go at adding the light beams in in photoshop. 3D modeling may still be beyond me, but I like to think I'm alright in 2D. IMHO I think the top right light beams came out alright, but the cone on the right and the blue area on the left definitely aren't great. Obviously being a test render I didn't want to waste a whole lot of time on it, but I thought I'd give it a shot. Yay? Nay?
So... I think "Yay", and thanks for giving it a shot :) Indeed the top right beams are alright. If I cannot get a decent solution within Blender, I think I'll be able to safely hand my render to you for post-processing, which isn't really my forte.
Just letting you know: these light beams in the top right corner are meant to be on a proper "background layer", whereas the other beams of light (such as the yellow-greenish on the right and the purple one on the left) are on a separate "light" layers which would have to be made by post-processing (compositing within Blender, or Photoshop on your end). You can learn more about how the fields are layered if you open a field file using Palmer. In any case, I can guide you through whatever post-processing which would need to be done.
Most of the original backgrounds had been edited in an image editor.
That's true, and there's more. Pretty much any advanced CG art has some post-processing to it. It's actually recommended practice.
Correct me if I'm wrong, but I believe that the glow around the bulbs and the rays of light are essentially the same idea, which is volumetric lighting. It's something that's only recently been introduced to Blender/Cycles in the form of Volumetric shaders. I don't think it gives true volumetric lighting, but there are ways of faking it I think, which is what I believe Mayo is struggling with.
Again my knowledge of 3D is spotty, but I think that's the gist of it. Once I get a job and can afford to upgrade my rig, then I'm gonna invest some time in actually learning this stuff. :wink:
As far as I'm aware, Blender doesn't really have a volumetric emission of light, and I'm not sure how to mimic this. Blender has a very nice volume scattering shader, but I was unable to use this shader in order to create the intended "light beam" effect, mostly for 2 reasons:
1- If you define a volume with many light scatterers in it, the density of scatterers is isotropic in the volume. Thus, it is difficult to have a lateral "fading" of the beam of light, and the result really feels unnatural.
2- The volume of scatterers will scatter all the light sources. Given the multiple light sources in my scene, this makes a weird result.
Maybe there's a solution out there, but I haven't found it.
Now, the light around the bulbs or just surrounding lamps are (in my opinion) a different effect - although the cause (scattering) may be similar. Often, we're talking about glow, or glare effects. These are meant to be done by post-processing, so it will be best if I leave these to you, as well. I'm not sure Blender users have tried to replicate glare effects by means of shaders, but I believe it would be overly complicated while it can be made easily via post-processing.
All in all: I'll keep investigating for solutions on my end, but I'll be sure to contact you when it comes to do the post-processing.
My problem is that the design of ealin(?) is entirely against Aerith statement that the church is the only place where flowers grow in Midgar. Also from where is the lighting coming?
I'm assuming you're referring to the field file describing the outdoors of Aerith's (and Elmyra's) house. As for "where is the lighting coming", I've always assumed it was daylight, which could be explained by a combination of
a/the house being located at the periphery of the city
b/ the Plate construction being incomplete in some parts
As for the Flowers question: the presence of flowers can be explained from two angles, in my opinion:
- One angle is rather "symbolic": Aerith is a special character, her Cetra lineage grants her particular powers, such as "interacting with the Planet" (sure, it's a pretty vague term). The presence of flowers where Aerith lives symbolizes this interaction, in my view.
- One angle is more pragmatic: Aerith loves flowers, which only grew originally at the Church. She could have perfectly transplanted some in the backyard of her home (you'll notice there's plenty of water around, besides).
Well, that's my take on it. At least it allows for reconciling what she says with how the environment is like.
-
Just letting you know: these light beams in the top right corner are meant to be on a proper "background layer", whereas the other beams of light (such as the yellow-greenish on the right and the purple one on the left) are on a separate "light" layers which would have to be made by post-processing (compositing within Blender, or Photoshop on your end).
Haha yeah, right after I added those other two spots I realized they would be on different layers, but that shouldn't be a problem. I'll google around for tutorials specifically for post process lighting, since I'm just going off my general photoshop knowledge atm, and worry more the layering when the time comes.
I'm glad I can finally contribute something :)
-
In the original image some things seem more "shiny" and the air seems more "foggy" and "bright" if this makes sense?
Will this be applied to your render via the post processing? At the moment it looks a little off compared with the original (also in comparison with your other works which are true and better than the originals :)).
-
In the original image some things seem more "shiny" and the air seems more "foggy" and "bright" if this makes sense?
I understand the comment. I will try to tune a bit more the lighting to match this. Mostly, It will involve increasing the density of the volume scattering shader and also increasing the intensity of some lights, all within Blender. Then, the "light beams" and other glare effects will be made by post-processing (most likely, I'll give the renders to Dubular who'll do a Photoshop job).
One thing to consider: the lighting may not end up matching the original very well, because the original lighting is not realistic (think about this, in the original image: the environment has red-orange tones, while it's mostly concrete/metal, and the lights have blue glows).
One thing I also realize is that my scene seems a bit more saturated than the original. Having more density in the main volume scattering shader will help that to some extent. However, desaturation can also be handled via post-processing too.
So... it goes to show I still have quite a bit of work to do on that scene. I'll also need to take care of the bus/weapon shop.
-
Ok, here's a small update after some more tuning... mostly: increased the volume scattering to have a bit more "dust mist" effect, increased the intensity of some of the lights. I also gave my best shot at creating volumetric lights in the top right corner (plugging an emission shader into the volume output of the object). The big problem is that there are very very few tools which enables to control the density of the emission over the volume, so that may be the best that can be done with the current nodes available. (again, this is a 50% res render)
(http://i.imgur.com/wveOj2n.png)
In all fairness, I think that Dubular photoshop job will look better. I don't think this render isn't useful, since the volumetric lights I made can serve as a very useful guide as to where the light beams are located with respect to the metallic beams of the big pylon.
Otherwise, I think the rest of the scene is pretty good - I don't think I'll want to increase the main volume scattering (i.e. dust mist effect) more than that. The overall level of brightness seems satisfying on my end (Covarr's feedback would be useful in that respect, given our earlier discussion), although it is a bit darker than the original.
Let me know if you have suggestions regarding the lighting or other stuff... next I'll probably move on to mds5_w (bus/weapon shop).
-
The overall level of brightness seems satisfying on my end (Covarr's feedback would be useful in that respect, given our earlier discussion), although it is a bit darker than the original.
It's not so dark as to cause loss of detail or visibility issues. It's right on the threshold; any lower would be too dark, but as it is, it's fine.
-
Update on mds5_1... (click for 100% res)
(http://i.imgur.com/gaG85kC.png)
I hope you like it!
Well... this means I don't have much left to do for mds5_1 except perhaps finalizing my lighting setup and importing the contents of mds5_w (bus/weapon shop). I'll be moving on to that scene (as mentioned in the past 4 posts :P ) at last.
-
Yes! That bus was looking way too clean ;D. Looks great.
-
The bus needs scratches specially the green paint looks too clean at the moment. The reflection is too strong, there should more dust on it which would prevent most of the reflection of the car paint. The last thing is nitpicking, but the bus body looks very thin at the door.
Considering that this is a weapon shop I would add some bullet holes on it, maybe a target, too.
But it is overall great work and I can't wait to see it in game.
-
The bus needs scratches specially the green paint looks too clean at the moment. The reflection is too strong, there should more dust on it which would prevent most of the reflection of the car paint.
Based on your feedback, I made some adjustments on the texture. It's rather discreet. Namely, I added a map to fade the green paint over the grey paint a little, I enhanced the dirtmaps a bit, added a scratch map, and toned down the reflections. Below is the result.
(http://i.imgur.com/H7LbrhA.png)
At this point, I don't want to lower the glossy shader further, since I think it is necessary to have a minimum amount of glossy to convey the feeling of painted metal. I hope that's a satisfying update.
The last thing is nitpicking, but the bus body looks very thin at the door.
It's 2 cm thick. I based my design on the style of old buses, and I think it's consistent. I don't think the good oldie VW Samba bus has much thicker doors than that.
Considering that this is a weapon shop I would add some bullet holes on it, maybe a target, too.
I don't think that is a good idea. If you're the owner of the weapon shop, and you want your customers to have some practice shooting, I don't think you want them to aim at the bus or even risk an indirect bullet impact. If I had to place targets and a shooting range, maybe I'd try to set it up inside the tents, parallel to the bus (and they wouldn't be visible anyway).
But it is overall great work and I can't wait to see it in game.
My personal goal was to finish the outdoor slum 5 marketplace and all related indoor scenes, plus outside and inside Aerith's house scene before the end of this year. I don't think I'm ahead of schedule :-\
-
More realistic... you die! :D
-
My personal goal was to finish the outdoor slum 5 marketplace and all related indoor scenes, plus outside and inside Aerith's house scene before the end of this year.
That's very optimistic planed. :P
But the bus is (nearly) perfect now.
|
L > It's simply that my brain tells me that the body needs to be thicker at the door.
http://commons.wikimedia.org/wiki/File:1961_T-2_Double_Door_Camper_VW_Bus_-_Flickr_-_starlord.jpg
I don't think that is a good idea. If you're the owner of the weapon shop, and you want your customers to have some practice shooting, I don't think you want them to aim at the bus or even risk an indirect bullet impact. If I had to place targets and a shooting range, maybe I'd try to set it up inside the tents, parallel to the bus (and they wouldn't be visible anyway).
OK I see your point here. But I think there is another point of view and that's the question what the scene should represent. I think that Shinra has forged a careless society and that this should be represented in many actions of the people who live in Midgar and the surrounding where they live in. It's a recommendation by me, but it's nothing I would stay with, if you think it doesn't fit.
-
Update on mds5_1... (click for 100% res)
I hope you like it!
Well... this means I don't have much left to do for mds5_1 except perhaps finalizing my lighting setup and importing the contents of mds5_w (bus/weapon shop). I'll be moving on to that scene (as mentioned in the past 4 posts :P ) at last.
I suggest that you try rendering at 200% then downscaling to normal size, if you want to remove pixelation (natural grain will remain intact)--pixelation is common to Blender's default anti-aliasing mode and even in full anti-aliasing some features still seem to cause pixelation at least from my experience.
-
That's very optimistic planed. :P
Well... it's not unrealistic if you think about it. Here's the tally:
mds5_1: almost completed
mds5_i: maybe missing one texture
mds5_m: missing one texture
mds5_w: very little is done (I'd say 5% completion)
mds5_dk: completed.
5min1_1: missing one texture
5min1_2: needs to be completely retextured
eals_1: needs to be completely retextured
ealin_1 and ealin_12: needs to be completely retextured
ealin_2: needs to be made from scratch
When I've done retexturing jobs on mds5_i and mds5_dk, each took about 2-3 weeks, so I'm expecting the same amount of time on 5min1_2, ealin1/ealin_12, and eals_1. ealin_2 may be the scene which will take me the longest time. So... it's still feasible, albeit on a tight schedule.
Anyway, I think it is necessary to set ambitious yet realistic objectives in such big projects which you undertake as a volunteer. There is a big problem in projects you work on "when you have time": if you don't set deadlines and objectives, generally nothing ever gets done. Another big problem is to establish when you're content with the job done, and move to the next task. Most of these scenes could be individually improved on and on for years, but if you spend one year on a single scene, the project will be neverending. I think you can relate to your character modelling project: you could probably spend a crazy amount of time sculpting each wrinkle of a single character, but that would be at the expense of the 50+ other character models you would need to work on (and as far as I'm concerned, one strength of your project is the relatively consistent and even quality of your models all across the board).
I suggest that you try rendering at 200% then downscaling to normal size, if you want to remove pixelation (natural grain will remain intact)--pixelation is common to Blender's default anti-aliasing mode and even in full anti-aliasing some features still seem to cause pixelation at least from my experience.
It's a good suggestion, I've considered it myself before. The only issue is that the resolution we're aiming at is already fairly high (mds5_1 is 2560 x 2560, for instance). Rendering it takes already a long time. So... rendering at 200% would multiply these render times by 4. I'm guessing the question is: what image has the better quality between a render at 200%, scaled down by half, or a render at 100% with 4 times the number of Samples. To be honest I don't know the answer to that (it may even be dependent on the scene in question).
-
My suggestion would be: test it in game first before you go overboard with 4k resolutions...
before you know it we might need to do twin cam 3d glasses compatible versions... :o
8-)
-
Well... I wasn't planning on testing 4K in game. I merely thought of rendering at 4K and scale down to HD in order to remove noise (maybe not the most efficient way to do so), and then build the field files.
Although... testing 4K in game may be an interesting experiment :P
-
Update... Coming back at last to mds5_w, mostly working on textures.
(http://i.imgur.com/u0tcel6.png)
There are a few things which will be changed from the original scene, which was riddled with design issues. The most drastic change is that the bus driver seat will not be visible: to be consistent with the proportions of the bus in mds5_1, the driver's seat is position a bit more than 1 metre in front of where the visibility of mds5_w ends. I'll replace it with other stuff. The dimensions of the other seats have followed some minor alteration (the ones I textured), so that they would have similar dimensions as a proper bus seat (http://www.mslindia.net/pages/seats-school-bus.html). I'll probably use models from Blend Swap to get the weapons done and speed up the whole job on that scene.
I have also successfully imported the environment of mds5_1 into that scene, so that it can be seen through the windows and windshield of the bus (you may see some canvas through the windows, and the rubble of the broken construction through the windshield. I'll have to work on the enviro light, though.
[EDIT] After double-checking, my seat dimensions do not correspond to those of reclining seats, which a/match the original design b/correspond more to the concept of a touring bus. So... I guess I'll redesign them. Oh well. At least I got the materials done.
-
Wow! I'm blinded by immersion!
-
You've come a long way Mayo, it's great to see your still contributing to the 'cause' :P.
Are you using generic units 0_o?
-
One of the first things I do when working on a scene is to establish a proper scale. Usually I take the reference of an object whose dimensions are standard (example: door frames have standard height of 2.1 m) and normalize the whole scene based on that reference. For mds5_w, it was bouncing back and forth with how large the bus is in mds5_1, and match it with plausible bus dimensions. Once a scene is properly scaled, it's very easy to put units to all the dimensions in your scene.
Scaling a scene is important since many objects at original resolution have absurd sizes (namely, small objects were considerably enlarged to make them be recognizable at the very low res of "standard tv", rather than be small pixel blotches), which needs to be corrected in our HD renders. Another aspect is that some physics simulation in Blender (namely fluid sims) need to be scaled correctly, as well.
-
Update! Basically I revised and textured all the objects that were made previously.
(http://i.imgur.com/JXoLspo.png)
I hope you like it so far.
-
That's coming along nicely.. Can't wait to see the machine gun :mrgreen:
-
Update (click for 100% res)...
(http://i.imgur.com/7GYxFER.png)
So... a few comments:
- Given how the bus was redesigned, the driver's seat was relocated much more in front, in a place not rendered in the indoor view. So, for that indoor scene, I needed to model something to fill the same place. I decided to opt for a luggage rack. I hope you find this suitable.
- Since the last update, in addition to the luggage rack, I added a bunch of cigarette butts, soda cans, posters/papers, clay pots, and a whisky bottle (which I will probably copy paste in several other places - I still need to work on the labels). I feel like the scene could use for "trash" objects (like cigarettes and soda cans), given that the rescaling process made them significantly smaller than in the original.
- I have a chainsaw model taken from BlendSwap (I can use it as long as I give credits, and it's for non-profit stuff). I'll pretty much render it for final renders (not WIP renders) so that I don't have to give acknowledgements 10 times in this thread. I had to retexture it, though (it was not made for Cycles). So, you'll see it eventually. The rescaling process made it look a bit small, although I would say the original chainsaw was humongous. I'll most likely do the same for all the other guns in the scene. I may want to model the katana myself, though.
I guess that's about it for now.
-
I dunno if it's just me or what, but all that littery stuff seems really small relative to the scene. Either that, or this bus is enormous. It might be worth getting the scale technically wrong for the sake of feeling right, if that makes any sense; the wonky design in this game isn't always gonna be compatible with realistically-proportioned objects.
-
That is a stunning bus you've got there mayonnaise, everything looks great but I agree that maybe the soda cans and cigs could be blown up 1.5x and made a tad fatter. Look at the hand-fists on the field models, to scale, next to the cans. Other than that, killer job so! I'm excited to test this area out
-
Beautiful.
-
Scaling was complicated because I wanted a match between the indoor and the outdoor scene, and this is the best compromise I could make (I'm not sure there's much room in that department). Yes, the bus is pretty big (slightly more than 2 metres wide, which is more than common buses - I think the length was around 9 metres and the height 3.5 metres). Still, the original was actually a lot bigger, and sizes were absurd (5 metres wide and tall, soda cans 50 cm tall, etc). I'm staking reference with Kaldarasha's models, and I made some pre-testings in game (that was a while ago, though, when testing ovrall dimensions of the bus) using myst6re Makou Reactor field editor (it works pretty well when models are about twice the size of the originals, i.e. around 1000).
Regarding the trash objects, since they look relatively small, I'll make them more numerous. I wouldn't want to modify the scaling I've come up with.
-
Regarding the trash objects, since they look relatively small, I'll make them more numerous. I wouldn't want to modify the scaling I've come up with.
How does the scale look with chibies running around..?
just wondering
8)
-
How does the scale look with chibies running around..?
just wondering
8)
To be honest: not so good. But there's the thing: either we want the scene realistically made for characters with realistic proportions, or we keep it deformed for chibis characters. I don't think we can make a rescaling process which would be a suitable compromise for both kinds of characters. My take was always to make a design for realistically proportioned characters (in my opinion, the game gains a lot with new high poly models with more realistic proportions, while chibis look a bit off). I have to admit I have never been fond of chibis, even back in 1997. My own tastes aside, creating realistically proportioned environments (for realistically proportioned characters) makes them more believable (giving more immersion). It also helps the modelling of objects by taking real life examples and take their dimensions as guidelines. If an environment is scaled for chibi models, it will look more like a dollhouse than a real place. That pretty much sums up the motivation behind this design choice. This design choice will have to be consistent across 3D modelling artists, so I hope you're following me on this one.
-
This design choice will have to be consistent across 3D modelling artists, so I hope you're following me on this one.
Errr...not to be contrary here but...isn't using chibis one of TA's big things? I thought that was all part of keeping with the style of FF7 but just increasing the resolution.
In any case the bus is looking awesome!
-
Errr...not to be contrary here but...isn't using chibis one of TA's big things? I thought that was all part of keeping with the style of FF7 but just increasing the resolution.
In any case the bus is looking awesome!
As far as I'm aware, the chibis were made as a first project, at the same time as the GUI. I believe at the time that HD field files were not even a priority. All this was made before I got on board - maybe SpooX or sl1982 would have a better idea of the history. Basically, you may say that the chibis were made a a time where:
- we hadn't even tested HD fields in game, and the "scaling" problem was not even outlined.
- Kaldarasha wasn't around, so we didn't have many good character models to play with (the shortcoming of many "character upgrade mod" is the consistency of the quality between one single very nicely detailed model and all the rest made of blocky chibis).
As we went through more and more modelling of the field scenes, the "scaling issue" became more and more apparent, and it looked like something we had to change (such as my favorite example of the tea cup measuring 40 cm in diameter). In a sense, we adapted the work after learning new elements.
On another note, some details just for fun...
(http://i.imgur.com/5x9PrQO.png)
-
I would totally buy whiskey that said "Mayo Master" on the bottle. Or Mayonnaise Master.
-
If I'm drinking mayonnaise whisky, it had better have more than 40% alc
-
I would totally buy whiskey that said "Mayo Master" on the bottle. Or Mayonnaise Master.
Sadly I'm not into making whisky, so that's not going to happen. I did however make a batch of mayonnaise with 12 years old whisky in it, it was pretty darn good :D
If I'm drinking mayonnaise whisky, it had better have more than 40% alc
If I were to make whisky, it would be for the flavor, not the burn. But I could give you cask strength if you want more than 40% :P
-
As far as I'm aware, the chibis were made as a first project, at the same time as the GUI. I believe at the time that HD field files were not even a priority. All this was made before I got on board - maybe SpooX or sl1982 would have a better idea of the history.
errr...
Mayo, don't know how to put it, but did you ever read the TA faq?
Q: Things look great but we want full sized models instead of chibi. Can you make some?
A: Nope, look elsewhere.
Of course in the end evrything might be modular, i.e. from the package you don't install everything, but that is basically the reason that some objects are in fact in 'Chibi-scale'. The annoying part of making everything in exact scale is that those little can as example will be lost in the scene. And furthermore, any field models will look out of place.
just my two cents to keep it in mind, in the end it will be our artistic freedom to do whatever we want... :evil:
Oh and if the taste is good, I'm in for a 25+ year old wiskey, now that would be 37 years by now... :o
I like the details...
8)
-
errr...
Mayo, don't know how to put it, but did you ever read the TA faq?
I did. Back when sl1982 and Timu started this endeavor, the original scope was to replicate the fields as simple magnifications of their original counterparts. This is probably more in line with the idea of chibis they had back then. If we wanted to strictly stick to this idea, I do not think the project of Team Avalanche would have significantly added value over what Omzy and then yarLson have accomplished. Should the scope of the project drafted by sl1982 and Timu more than 4 years ago be a rigid framework? I do not think so.
Of course in the end evrything might be modular, i.e. from the package you don't install everything, but that is basically the reason that some objects are in fact in 'Chibi-scale'. The annoying part of making everything in exact scale is that those little can as example will be lost in the scene. And furthermore, any field models will look out of place.
just my two cents to keep it in mind, in the end it will be our artistic freedom to do whatever we want... :evil:
Well, among field modellers we have to agree on something, for consistency's sake. I've already explained at length why I was opting for "realistically scale" modelling. I also think realistic-scale modelling would also facilitate consistency (since each modeller can refer to real life standards to design things). If modelling is made for chibi proportions, to be sincere it is very difficult to decide what items to deform and what items should be kept "realistic-ish".
I don't know if you'd want to make a poll to see who wants fields tailored for chibis vs. who wants fields for characters with realistic proportions. Or if you'd rather debate the direction of the project before we write a sticky about possible changes of scope.
-
While in the end it is entirely up to the modelers I would like to apply my two cents. First of all, by changing the scale of the game your implying that realistic texture models will then have to be created on top of the fact that animations would have to be reworked entirely, probably from the ground up. In fact you might have to facilitate some sort of creation of a new model format entirely so that we can actually animate the models using modern standards rather than cutting them up into pieces.
Secondly, this isn't just a minor deviation from a rigid framework but an entire reimagining of the projects goals. From the the start TA has been about accurately recreated the graphics based on their original artwork and proportions. To go from a faithful recreation to scale realism is at odds with the very core philosophy of the project.
Now to be clear, I would thoroughly enjoy playing either project in the end and I am not saying any of this to pick sides. More to illustrate that if you want to continue in this direction you might want to consider changing the name of this project to reflect the change in philosophy and leadership. Furthermore the implication of work involved compounds exponentially if your goal is to actually bring the project to absolute visual consistency (because of model animations, probably walk meshes and any other unforeseen issue that arises in changing scale). Not to mention the fact that many field scenes have already been worked in TA style and would have to be redone again to fit the new scale.
Lastly, just in my opinion, if I were to choose I would prefer an artistic recreation if your going to stick with the core game-play mechanics such as prerendered backgrounds, text interactions and everything that implies. However I would prefer an accurate scale recreation if we could somehow break from prerendered backgrounds in the future and add in voices and more complex world dynamics like a controllable camera and a new model format.
Anyway that's just how I feel about it.
-
Ok... just to make sure, I decided to double-check a few things in game regarding scaling and the use of chibis. After further examination, here are my observations:
- I used chibis only, tested original fields scaling vs. revised realistic scaling. It turns out that realistic scaling isn't really detrimental to the use of chibi models (my earlier comments (http://forums.qhimm.com/index.php?topic=13786.msg217016#msg217016) were actually flawed, since not based on actual testing). I tested it on my finished sewer scenes (colne_be1 and colne_be3), as well as my WIPs on the Sector 5 slums. In fact the scaling from original fields doesn't do a better favor to chibi models (for instance, soda cans in the original fields are still 3/4 the size of the chibis head). In ealin1/ealin12 original fields, objects in general are way too big with respect to the chibi models on default settings (ie. without using Makou Reactor). It used to be not-so-noticeable when playing on low resolution, but in 1920 x 1080 it's a bit of an eyesore.
- In fact, one more important parameter regarding scaling and how it fits character models is about how the character size is adjusted for each field, which can be edited using Makou Reactor tool. For instance, if you wish to have realistic proportions of field and characters, you need to edit that infamous bus/weapon shop scene (mds5_w). I haven't tested what kind of scaling the chibi character would need in Makou Reactor in order to get good results for that scene.
All in all, these tests have comforted me in the realistic rescaling direction. The bottom line is: realistically scaled scenes will look better than original scale scenes for both realistic and chibi models, and for achieving the best results one will need to use Makou Reactor in either case.
Do I find this situation awfully convenient? Well, yes, and I am shameless about it 8)
Ok, a new comment popped up while I was typing this...
While in the end it is entirely up to the modelers I would like to apply my two cents. First of all, by changing the scale of the game your implying that realistic texture models will then have to be created on top of the fact that animations would have to be reworked entirely, probably from the ground up. In fact you might have to facilitate some sort of creation of a new model format entirely so that we can actually animate the models using modern standards rather than cutting them up into pieces.
Secondly, this isn't just a minor deviation from a rigid framework but an entire reimagining of the projects goals. From the the start TA has been about accurately recreated the graphics based on their original artwork and proportions. To go from a faithful recreation to scale realism is at odds with the very core philosophy of the project.
Now to be clear, I would thoroughly enjoy playing either project in the end and I am not saying any of this to pick sides. More to illustrate that if you want to continue in this direction you might want to consider changing the name of this project to reflect the change in philosophy and leadership. Furthermore the implication of work involved compounds exponentially if your goal is to actually bring the project to absolute visual consistency (because of model animations, probably walk meshes and any other unforeseen issue that arises in changing scale). Not to mention the fact that many field scenes have already been worked in TA style and would have to be redone again to fit the new scale.
Lastly, just in my opinion, if I were to choose I would prefer an artistic recreation if your going to stick with the core game-play mechanics such as prerendered backgrounds, text interactions and everything that implies. However I would prefer an accurate scale recreation if we could somehow break from prerendered backgrounds in the future and add in voices and more complex world dynamics like a controllable camera and a new model format.
Anyway that's just how I feel about it.
The character modelling side of Team Avalanche has been down for a long while now, and as far as I am aware there is no active contributor at all. As far as I am concerned, TA's work should now be solely focused on 3d modelling of fields for HD rendering, and that's it (this scope alone is insanely huge, probably more than I can chew). Practically, that's actually what's been going here for more than a year anyway. My aim would be to rather team up with other people or teams working in other areas (I was thinking of Kladarasha's characters, Bloodshot's HD battle fields, etc.), trying to pick the best of what's out there and work in correspondence with them. If people feel like we should make some official notification, I don't mind.
With that in mind, it's mostly up to the character modellers to work on the character's animation, not something we have to address in Team Avalanche. Then, when you mention that the work involved increases because I have to check visual consistency, walkmesh and layering, well I'm already doing all that. One last thing is: there are actually very few scenes which have been actually completed and released. So the amount of work which would need to be done to "fit the new scale" would be rather minimal (I think it would only involve rescaling a few items in anaho's finished scenes of Mideel), all the other completed scenes (by yours truly) either have already undergone the process, or are going to be reworked anyway (for instance, I need to rework all the texturing of eals1 and ealin1/ealin12 to be able to render them with Cycles, revisiting scaling won't be a lot of work in comparison).
Among the possible change of scope, what I would like to achieve is a relatively high visual consistency between fields, combat, and FMVs. That is not like it was made in the original game, but I do believe the visual discrepancies from the original were mostly due to technical limitations rather than artistic choice. For that matter, Square tried to address this issue in their following iterations of the franchise (fields and battle models were the same as early as FF VIII - and gradually they tended to try to match field with FMVs).
One other thing: our work (at least mine) on pre-rendered background won't be compatible with the concept of a dynamic camera, for 2 reasons: a/ so far I have been working with no restrain on poly counts whatsoever. The modelling logic behind making objects which need to be displayed dynamically would be totally different (and probably imply more work). b/ Things which are off camera on the fields need only a very coarse modelling/texturing (they only interact on the basis of their reflection and shadows, so no need to spend much time on that). If you want to model entire districts like in current open world games, it's a completely different ball of game.
As one last remark, I would like to reiterate that if TA's job is to make strict HD replicates of original fields, your own project would make TA's endeavor rather pointless (not blaming you there, your work is pretty awesome).
So I'll put it plainly: if you feel like my vision is a departure from the original scope, then I would very much like to change the scope. Now the questions would be: Do I need someone's green light to proceed (most likely SpooX's, but who else?)? If so, is it necessary to create a thread to discuss about it?
-
Have any in-game screenshots to compare?
-
I haven't found a way to take in-game screen captures while running on full screen (if anyone knows how to, I'd appreciate some tips). The best I would recommend so far would be for you to check out some completed works I made (http://forums.qhimm.com/index.php?topic=14810.msg207821#msg207821), and do the test yourself.
-
Use a borderless window tool and run the game in window mode.
http://forums.qhimm.com/index.php?topic=15214.msg212938#msg212938
-
Fraps takes screenshots. If you are on Windows, and you must have a beefy gfx card, you can set hotkeys for screenshots within amd and nvidia control panels... I think.
-
Okies, so here are screenshots...
Below is one original field with scaling uses (character size has not been edited).
(http://i.imgur.com/GL697hq.png)
Following the original scale, even with chibis, many elements of this scene have an absurd size. The TV and the flowers are the most blatant example. A general character upscale (using Makou Reactor) may partly solve it, however some elements are not properly scaled when compared to each other (for instance, the lilies in the background are huge with respect to the table and the chairs).
Now, for proper comparison of various cases, I chose the colne_be3 scene. I picked this scene because I've completed my work on it, and I've done the "rescaling process" in doing so. This scene does not present many instances of the "scaling problem", though. I placed Cloud next to the boot because this is one particular object which went through rescaling revision.
First, let's have a look at the original field with original chibis:
(http://i.imgur.com/e1Gmt5C.png)
As you can see, the size of the boot is a bit too big in comparison to Cloud's size.
Next, let's see my moded field, after fixing scaling issues, using the chibi model:
(http://i.imgur.com/n1ig6mG.png)
I believe the scale of the boot in my moded version is much more suitable than how it is in the original field, even when using chibi models.
Finally, let's see my moded field with Kaldarasha's (old) model of Cloud, which has more realistic proportions.
(http://i.imgur.com/SJ3Pjue.png)
Obviously the scale is suitable, because I had taken this model to test the scale of the objects in my scene. In my opinion this is how the game is meant to look like.
All in all, I think realistic scale is more suitable than original scale, even with chibi models.
That's all folks :P
-
I admire this scene everytime I look at it man! Looks so well designed and detailed/lifelike!
-
Chibis4life. I don't know what it is about the realistic models but they always look so unnatural and out of place. I guess the chibis look fine on those backgrounds regardless.
-
For nostalgia yes, but i personally don't believe I can stand those blocky Lego like/models with today's graphic potential...
-
Chibis4life. I don't know what it is about the realistic models but they always look so unnatural and out of place. I guess the chibis look fine on those backgrounds regardless.
If everything would use chibi size like FFIX (or Crono Cross) then it would be OK. But the game use the chibi design only because of the inexperience of the former developers and that they thought they could realize emotions better this way. However they had canceled the idea with the face expression (I think Square has forced the dev team to bring the project to an end), the :o faces of the models in the PC version are a proof of that. These models are an earlier built which weren't cleaned. FF7 is in it's model design pretty inconsistence so making the models to a more realistic shape is the only option we currently have. Because we can't change the battle models so much without getting serious problems with the animations. There is also a higher demand for a non chibi design.
Anyway I still would love to see a review of the games most memorized scenes with the chibi characters, when the staff roll is seen.
Besides my next update will cover some resizes for the field models for the vanilla game (or the Reunion). So the chairs, which currently looks to huge even for Barret, will look to small for him.
-
I would just add one thing: it's always been in the plans of TA to recreate environments in order to re-render the FMVs in HD (although I admit it's rather for the far future), and one thing we need to work on is visual consistency between field environments and FMV environments. I would believe FMV environments were meant to fit FMV models, so it would make sense the field environments suit these models as well.
-
Hi folks,
Here's an update on mds5_w (ie bus/weapon shop). Since last time, I've mostly added details. I hope you'll enjoy.
(http://i.imgur.com/h2H85OR.png)
So... most of the stuff is done, I mostly need to import weapon models from Blend swap, plus add a few details on the rear parcel shelf (I need to model a toolbox, for instance). Final touches are about importing an enviro light based on a render from mds5_1 (the outdoor scene), and add a bit of volumetric scattering to give the place a bit of a dusty feel.
However, I'm off tomorrow for some (much needed!) vacations - I'll be a bit off the grid for a week.
8-)
-
I really love it, Mayo Master!
Great stuff!
-
Hi folks,
So... I've made some progress on the weapon shop scene - I imported a few weapon models and modelled a few other things, it's almost complete. Therefore, I made the import from that scene back into the main outdoor scene of the Sector 5 slums, mds5_1. Below is the result:
(http://i.imgur.com/bBrZBjx.png)
[EDIT]: I forgot to update a texture on that render - please do not mind the yellow color of the door, it will be replaced.
Anyway, this means that mds5_1 is almost complete as well. I only need to make final adjustments to the lighting and volume scattering. However, I don't think I'll be able to create particle systems to litter the place with "trash objects", since it seems like I'm finally reaching the limits of what 16 Gb of RAM can handle. I'll try to upload a render of that scene soon.
-
your work is truly amazing! I hope you keep the scenes in case Q-Gears really kicks off!
-
Well... the scope of this work doesn't really depend on the status of Q-Gear. The fields I make already work on the current engine.
To be honest I am not sure what the possible benefits of Q-Gears would be like, as far as the fields are concerned. I am curious - It would be nice if somebody could fill me in.
-
Anyway, this means that mds5_1 is almost complete as well. I only need to make final adjustments to the lighting and volume scattering. However, I don't think I'll be able to create particle systems to litter the place with "trash objects", since it seems like I'm finally reaching the limits of what 16 Gb of RAM can handle. I'll try to upload a render of that scene soon.
Time to upgrade to haswell-e with 128Gb of DDR4?
-
Time to upgrade to haswell-e with 128Gb of DDR4?
DDR4 is a consumer thing now? Holy crap I haven't been paying attention. Might as well rent time on a render farm for whatever that would run you.
That is the most beautifully war-torn looking bus that I have ever seen, IRL or otherwise. Perhaps a little too polished/reflective for how rusty it is, but by the beard, that is beautiful.
-
Update!
Below is the latest 100% render I made for mds5_1. It took a whopping 38 hours to be completed (with 2000 samples) - I'm stuck with CPU rendering (still, I have a good CPU). Click on the image to get the full resolution...
(http://i.imgur.com/Pp6ccSx.jpg)
So... regarding this scene, this is as far as the modelling will go. I could see ways of improving it, such as lattice deformations on the rooftops and particle systems for debris and trash, but I am not going to do that for "mark I" of this scene, for two main reasons:
a/ I think the modelling has reached a satisfying level, and I have already spent enough time on this scene (which I started to model back in November 2013). Time for me to move on (the time I would be spending for more details here would be less time spent on making new scenes).
b/ I am reaching my computer's limits (7 cores with 16 Gb of RAM), and I am not planning an upgrade anytime soon (bought this PC last year).
I am however still considering some improvements concerning the texturing and the lighting. Namely, I forgot to apply a concrete texture on a pillar, and I think I should tone down the bumpmapping on the wood beams. The lighting is pretty good in my opinion, although:
- I am open to suggestions regarding the lighting setup, which includes the volume scattering setting (I used volume scattering to create a dusty atmosphere, which is why the image is conveniently desaturated).
- Eventually I'll need to add post-processing effects (glare, light beams).
I hope you like it. Feedback and comments are welcome - if you have revisions to suggest, bear in mind that I will like to wrap this one up quickly.
-
looking at each section in detail up close, I really can't fault anything. All that it needs is the final post effects for glare and lighting to make it really fit and it'll be 100%. Awesome, awesome work.
-
If you need rendering I can probably do it. My pc is watercooled and can run at 4.7ghz 24/7. I leave it on all the time anyways
edit: can blender not do gpu rendering?
-
looking at each section in detail up close, I really can't fault anything. All that it needs is the final post effects for glare and lighting to make it really fit and it'll be 100%. Awesome, awesome work.
Thanks you! I already made some preliminary tests in game (background layer using this render, other layers from original fields), it's really promising.
If you need rendering I can probably do it. My pc is watercooled and can run at 4.7ghz 24/7. I leave it on all the time anyways
edit: can blender not do gpu rendering?
Thanks for the offer - I'll see if I become desperate about rendering. Blender can do GPU rendering, but not necessarily for everything (for instance I am not sure if strand rendering - for furs and carpets - is GPU supported). How much RAM do you have? Basically, for Blender, RAM will condition whether something can or cannot be rendered. Then CPU/GPU power defines how fast the render is going to be. So I'd say RAM is even more crucial than CPU/GPU when it comes to feasibility.
Otherwise, I thought I'd make this update regarding some new details for the bus/weapon shop scene (this is a 1000% render, the final result will be much smaller)
(http://i.imgur.com/st42o9x.png)
I hope you like it.
-
Haha! Great! :mrgreen:
-
Thanks you! I already made some preliminary tests in game (background layer using this render, other layers from original fields), it's really promising.Thanks for the offer - I'll see if I become desperate about rendering. Blender can do GPU rendering, but not necessarily for everything (for instance I am not sure if strand rendering - for furs and carpets - is GPU supported). How much RAM do you have? Basically, for Blender, RAM will condition whether something can or cannot be rendered. Then CPU/GPU power defines how fast the render is going to be. So I'd say RAM is even more crucial than CPU/GPU when it comes to feasibility.
3770k @ 4.5GHz
16GB Ram
R9 290x with 4GB graphics memory.
I could possibly upgrade to 32 GB if needed.
edit: if you throw a scene up on the repo i will see what I can do with the hardware i have
-
Hi folks,
Just letting you know that I've started to work on 5min1_2. anaho sent me the .blend file he previously made using Blender Internal rendering engine. My aim is to retexture it for Cycles (with lots of UV unwrapping), but there are a few objects to be remeshed and a couple of others whose modelled from scratch. Below is what I have done so far.
(http://i.imgur.com/fzxBqLs.png)
-
Well... the scope of this work doesn't really depend on the status of Q-Gear. The fields I make already work on the current engine.
To be honest I am not sure what the possible benefits of Q-Gears would be like, as far as the fields are concerned. I am curious - It would be nice if somebody could fill me in.
No limits on animations etc, and the image file could be used as-is instead of converted to that strange "block" format. Also you could use 3D field files rather than static renderings. This is probably possible with the current engine too using the custom hook/driver/rendering engine as it could simply render the mesh itself and not draw the 2D field images.
Edit: Actually I have been wondering, wouldn't it be cool to add 2D lighting to the current engine like in the "the swapper" game?
See here: https://www.youtube.com/watch?v=I7DxYZN6fu0 this is all 2D
-
That would be really awesome if pulled off the right way. Have you made much progress on field maps/wm? High framerate field with a much smoother 'panning' of the bg when running around would make a nice touch :-D
I know I've reposted this before, but it is awesome. I believe this is a QGears build from a few years ago. My computer wouldn't stand a chance with these new slums scenes.
(http://pics.livejournal.com/crazy_otaku/pic/00017pzq)
-
Yeah in QGears it should be easy enough to do. On the wm side of things the wm0 can be rendered - I need some spare time to carry on working on it. A decompiler to decompile the scripts to QGears WM lua is the next step.
-
Love them all!
-
No limits on animations etc, and the image file could be used as-is instead of converted to that strange "block" format. Also you could use 3D field files rather than static renderings. This is probably possible with the current engine too using the custom hook/driver/rendering engine as it could simply render the mesh itself and not draw the 2D field images.
Edit: Actually I have been wondering, wouldn't it be cool to add 2D lighting to the current engine like in the "the swapper" game?
See here: https://www.youtube.com/watch?v=I7DxYZN6fu0 this is all 2D
Having freedom for decomposing the fields along layers, as well as field animation (such as fluid flow) would be a great bonus - they would allow more freedom to rework scenes. As for 3d fields, I am not sure. The framework we're currently working in does not set a limit on poly count, and objects are not worked so that they could be rendered at 30 or 60 pfs - there's no such optimization (and to be frank, I'm not sure how to do it). In addition, texture/shading is specific to a rendering engine, so all the corresponding work I am doing for Cycles render would have to be redone for whatever 3d rendering system used in-game.
That would be really awesome if pulled off the right way. Have you made much progress on field maps/wm? High framerate field with a much smoother 'panning' of the bg when running around would make a nice touch :-D
Do you mean to make a rigged camera movement like in the urban environments in Resonance of Fate? https://www.youtube.com/watch?v=ltrPi4yt5Ic (https://www.youtube.com/watch?v=ltrPi4yt5Ic)
-
Having freedom for decomposing the fields along layers, as well as field animation (such as fluid flow) would be a great bonus - they would allow more freedom to rework scenes. As for 3d fields, I am not sure. The framework we're currently working in does not set a limit on poly count, and objects are not worked so that they could be rendered at 30 or 60 pfs - there's no such optimization (and to be frank, I'm not sure how to do it). In addition, texture/shading is specific to a rendering engine, so all the corresponding work I am doing for Cycles render would have to be redone for whatever 3d rendering system used in-game.Do you mean to make a rigged camera movement like in the urban environments in Resonance of Fate? https://www.youtube.com/watch?v=ltrPi4yt5Ic (https://www.youtube.com/watch?v=ltrPi4yt5Ic)
I think you would have to bake everything so they just textures/bump/spec maps or whatever - if you've seen Andrew Prices latest cycles videos you'll see what I mean. I think pretty much any real time rendering engine can load that data in some way.
What are the current poly counts? To be honest for real time rendering I think counts into the millions would be fine as by the time all scenes are completed it would probably be easy for entry level hardware to handle it. For example PS1 could only render the FF7 WM at 30fps and only limited sections of it. Today we can render the whole thing un-optimized at 200fps.
-
I would imagine the 3rd person top view from GTA/GOW is the only type that wouldn't confuse the hell out of everyone. It certainly would make for an interesting tech demo
-
I would like to see it even with a static camera as it just looks more natural and scales with future higher resolutions more easily. Also the lighting (even if baked) can be casted in real time on the chars making it look a lot better.
-
I would like to see it even with a static camera as it just looks more natural and scales with future higher resolutions more easily. Also the lighting (even if baked) can be casted in real time on the chars making it look a lot better.
I would need to work on a "test scene" and figure what would be needed to convert a scene into an environment you would render with QGears. However, I don't consider this a top priority. That being said, having the characters being exposed to the same lighting as the environment would be a huge bonus.
-
Update!
I've finished most of the texturing in 5min1_2, below is where I'm at.
(http://i.imgur.com/knPMPbd.png)
So... there are a couple of things I may want to revise: mostly I may need to redo the sculpting of the folds on the pillows (I've done that quickly yesterday, let's just say there's room for improvement), I'm also wondering about the red lamps next to the bed (anaho's original take was that they were simple colored glass globes). The top of the "cube" doesn't look so great, but it will be eventually covered by a cushion, so I don't think it's a concern. Otherwise I'm pretty happy with how the rest turned out.
There are a few things left to do: the aforementioned cushion, a crate with a couple of objects on it, as well as a number of trash objects (soda cans) littering the place, which I could import from other scenes. I don't think it should take long to wrap this one up.
-
In game test
http://youtu.be/rohAoIngmOA (http://youtu.be/rohAoIngmOA)
-
raptr, raptr, raptr! :D
-
I already had it on the computer and it made uploading it to youtube fairly simple.
-
Thanks for that video :)
Would it be possible for you to make one at 1080p? After all, 1080x1920 res. was the objective, might as well showcase it.
-
Yeah I dont know why it did that. I was playing in 1080p. Must be a raptr thing.
Sent from my SM-N900W8 using Tapatalk
-
I already had it on the computer and it made uploading it to youtube fairly simple.
Interesting how the guraud shading and wot not, still looks good after ... 17 years!
heh.
@Mayo Master nice stuff.
Tis been a while, doing too much work stuff sadly.
Cyb
-
Amazing - too bad the char models now look really really really really bad instead of really bad with the new awesome back ground images.
-
Amazing - too bad the char models now look really really really really bad instead of really bad with the new awesome back ground images.
That's precisely why I am testing and playing the game with Kaldarasha's character models (http://forums.qhimm.com/index.php?topic=13960.msg195913#msg195913).
By the way, sl1982, have you tried this combination of character/fields? As I mentioned before, I think that's how the game is meant to look like, and the best way to showcase our mods.
-
I havnt looked at it. I am a little behind on what mods exist these days. I have now added in team avalanches chibi cloud though.
-
Hi folks!
Looks like I'm pretty much done with 5min1_2. I'll just have to import and enviro light for the final render and that will be it - unless someone has useful advice about how to improve the scene some. So, here it is:
(http://i.imgur.com/f7VDIwE.png)
Who is going to spot the Scottish Easter egg?
-
Amazing work as always
-
Hey guys,
Need some n00b help, please. I just wanted to test something in game, and for a reason that I cannot figure, my game does not launch anymore. I have no idea why (basically, nothing seems to have happened between when it used to work fine - like yesterday - and now), I'm at a bit of a loss.
Basically, upon launching the game (using Run as admin), the window of FF7 opens and then closes as soon as it's opened. For info: I'm on Windows 8, Aali's driver version 0.7.11b.
Any pointers would be welcome - if you feel like I'd be better off uninstall/reinstall just let me know.
-
Had the same thing happen. Tried replacing the exe with a fresh 1.02. Still nothing.
Had to do a full reinstall. Added all my modded files back and it worked fine. I think something happened to the registry somehow.
-
Had the same thing happen. Tried replacing the exe with a fresh 1.02. Still nothing.
Had to do a full reinstall. Added all my modded files back and it worked fine. I think something happened to the registry somehow.
Thanks for the tip :) I was wondering if a Windows updated had messed up the registry, too. I'll do a re-install then.
-
What does the app log say? Sometimes it is sound related. Cant initialize whatever sound device is set up in the config
-
What does the app log say? Sometimes it is sound related. Cant initialize whatever sound device is set up in the config
You may be onto something. Here's what the App log says:
INFO: FF7/FF8 OpenGL driver version 0.7.11b
INFO: Auto-detected version: FF7 1.02 US English
INFO: NVIDIA Corporation GeForce GT 640/PCIe/SSE2 4.3.0
INFO: Found swap_control extension
INFO: Max texture size: 16384x16384
INFO: Number of texture units: 4
INFO: Original resolution 640x480, window size 1920x1080, output resolution 1440x1080, internal resolution 1920x1440
INFO: FBO extension detected, using fast scaling/postprocessing path
INFO: FFMpeg movie player plugin loaded
INFO: FFMpeg version SVN-r25886, Copyright (c) 2000-2010 Fabrice Bellard, et al.
INFO: No shaders, codecs with YUV output will be slow.
LOCK UNLOCK TEST
MATRIX INITIALIZE
INITIALIZE DD/D3D END
initializing sound...
How is this remedied (short of a reinstall, that is - I tried to change the Sound setting in FF7Config, it did nothing)? Thanks.
-
When i had that problem i changed the sound driver in the config file. If that doesnt work make sure that an audio device isnt disabled in windows (Right click the speaker icon in the system tray and go to playback devices)
-
Hi folks!
Looks like I'm pretty much done with 5min1_2. I'll just have to import and enviro light for the final render and that will be it - unless someone has useful advice about how to improve the scene some. So, here it is:
(http://i.imgur.com/f7VDIwE.png)
Who is going to spot the Scottish Easter egg?
Being a British person myself how could I not notice the fact all them sexy bottles are Irn Bru!!!! Bloody love that stuff and another great reason why I am happy the No vote won! Plus my Scottish uncle just brought down square sausages for us, seriously why do you have square sausages, they look burgers?
P.S. Looks amazing!
-
When i had that problem i changed the sound driver in the config file. If that doesnt work make sure that an audio device isnt disabled in windows (Right click the speaker icon in the system tray and go to playback devices)
Thanks for the input - I'll try that tomorrow.
Being a British person myself how could I not notice the fact all them sexy bottles are Irn Bru!!!!
Thanks for having spotted that! Although, they're not quite Irn-Bru :P
(http://i.imgur.com/1Mns1hl.png)
-
Fe instead of Iron, eh? I like the cut of several of your jibs.
-
Fe instead of Iron, eh?
Professional deformation ;)
When i had that problem i changed the sound driver in the config file. If that doesnt work make sure that an audio device isnt disabled in windows (Right click the speaker icon in the system tray and go to playback devices)
Well, amazingly enough, I managed to fix everything by... plugging speakers ???
-
Professional deformation
Don't tell me you're into chemistry.... ::)
8-)
-
Don't tell me you're into chemistry.... ::)
8-)
It's probably even worse than you think :P
-
Update!
I've started to revise my works on eals_1 (outdoors of Elmyra & Aerith's house) to render it with Cycles. I redesigned the cliffs, and here's what I have so far:
(http://i.imgur.com/yDd9F9E.jpg)
The mesh quality is much better now, which should allow me to have a lot more control on the particle systems that I'll use for the grass and the flowers. I also intend to do some work on the water material - it seems very tricky to give a realistic water rendering where you can have the opalescent blue color you find in glacial lakes - which is pretty much what the water on the original image looks like (although the reason may be mako-infused water).
-
Wow! Such detail! Great!!!
-
Can i buy one of the empty plots of land please to build a house?
-
Sorry I will be the highest bidder! (raises hand) :D
-
There is more than enough land for both of us :P
-
Hi folks,
I'm pretty much done with eals_1, I'm quite happy about how it turned out. Here is the scene:
(http://i.imgur.com/q0hBwkv.jpg)
I'll have a crack at the waterfalls later, which I intend to do by smoke simulation (a relatively thin waterfall this high would just become a big spray), I hope it will be well-integrated to the rest.
However, I think I'll move on to the indoor scenes now (the top floor of the house has to be made from scratch).
-
Looks great. If you want to send it to me for render and splitting ill give it a go
Sent from my SM-N900W8 using Tapatalk
-
Blender 2.72 has some improvements for you.I 'm sure you want to re-render all the .blends.
-
Looks great. If you want to send it to me for render and splitting ill give it a go
Sent from my SM-N900W8 using Tapatalk
This render was very manageable, I don't think I'll need to send it over (the target image size is "small", and there aren't too many layers). But I have enough scenes on my desk to give you some work later ;)
By the way, I wanted to discuss your earlier renders and layering of mds5_i. Namely, I wanted to know whether you separated the layers with Photoshop, or if you used the layering capabilities of Blender. My question is related to the fact that, in your test of the scene, it looks like there is a bit of aliasing along the objects which are placed on a separate layer (like the big transverse pipe). I personally have used layering within Blender before, and I don't think I had these kind of aliasing problems on my end (since Blender renders are with antialiasing). I'll need to do further testing to verify that, though.
Blender 2.72 has some improvements for you.I 'm sure you want to re-render all the .blends.
I've just checked 2.72 release, it looks interesting and I'll update shortly. However, I don't think there are that many scenes which will need a revamp. I think the main thing that 2.72 offers for my past scenes is the Sun Beam effect in the compositor. It will certainly come in handy when I model the indoor of Aerith's house.
-
I basically used blender to render out specific objects and used them as a mask in blender. If your way is better I will need you to show me how to do it.
Sent from my SM-N900W8 using Tapatalk
-
Could you please upload and link the renders you made (could be uploaded in the repo, if you prefer)? By this, I mean the layer images you imported into Palmer to recreate the "tiled fields". I wonder if you might have used masks on instances when it's not necessary. Namely, I found out that you shouldn't mask anything (ie render everything) on the base layer. Perhaps part of the aliasing issues comes from this.
-
They are up now.
-
I can smell those
daisies lilies now!
-
Thanks for the upload. I just had a check at your pictures: it does look like there is some weird aliasing problem on the foreground layers, the base layer is fine.
I may have a go at it, but not right now. It would be difficult to give you a very straightforward procedure for this, however I can give you some tips from my experience at working on that process on previous scenes (the sewer scenes, namely):
- it may not be mandatory to follow all the cuts of the original layer arrangements. I would advise not to duplicate objects from the base layer on a foreground layer when it isn't necessary.
- for Blender: I'd recommend you put foreground objects on, say, Layer #13 (3rd layer of the right block). For instance, you isolate the main pipe, the table, the computer and the cables on that layer. Then, on the Scene tab of the Properties panel, on the "Scene" side (left), you select all the layers which are going to be involved in the render (ie layers 2-10, 11 and 13) but on the "Layer" side (right) you only select Layer #13. If you render that, it should only give you the pipe, the table, computer and cables with minimal aliasing. Then, you can place these items back on one of the background layer, and select another group of objects to be rendered on Layer #13 (for instance, the chest of drawers, with the medkit and the brown flask), and render these ones, and so forth. Once you have good antialiased renders of these items, you can recombine them with Photoshop later, as needed (you'll need to recut layers via photoshop anyway, for instance when you need to put a chunk of the pipe on the third layer). Note: I don't know if you went through this process precisely already - it's a bit of a guesswork on my part. In any case, I did manage to have neatly antialiased renders on my end when it came to render distinct separate objects.
Now, with all these considerations: it turns out that the fields still produce some aliasing in game. However, I am not sure whether it is actual aliasing, or an artifact due to the fact that there are problems to display images which are at a resolution which is not an integer multiplier of the original (since HD is actually 4.5 times the original - my renders being 5 times).
I hope that helps. I don't know if you'd like the layers I made for the sewer scenes on the repo, to serve as a reference.
I can smell those daisies now!
I thought I had based my flowers on lillies :-[
-
Fixed ;)
-
Yeah throw them on the repo and ill take a look
Sent from my SM-N900W8 using Tapatalk
-
I just uploaded the layer files of colne_b3 on the repo. You'll notice how smooth the antialiasing is, in comparison.
-
Yeah it does look alot better. Ill see what I can do to fix the other one up
Sent from my SM-N900W8 using Tapatalk
-
Hi folks,
I've been working on the indoor scenes of Elmyra & Aerith's house (ealin_1, ealin_2, ealin_12). I've done quite a bit of ground work in order to have some correspondence (albeit not a perfect match) between the indoor and the outdoor scene, as well as a matching between the ground floor and the second floor. Now that's pretty much done, I was able to get going on the retexturing of the ground floor scenes (ealin_1 and ealin_12 are set in the same environment, but with different cameras and different layers visible). Below is an update, taken from the ealin_12 camera.
(http://i.imgur.com/dg5VRM3.png)
One thing to note: after studying the scene from a scaling perspective, it looks like the characters sizes are way too small with respect to the items and objects in the room. Makou reactor will have to be used in order to remedy that. I would expect a x1.5 rescaling of the characters to be necessary.
-
Update! Making some progress on the texturing...
(http://i.imgur.com/y6edZOH.png)
Since the kitchen closet glass panels are a bit noisy, I wanted to confirm that everything in the closet is textured and done. I may want to revise a couple of elements (scratches on the shelf on the left of the image, lack of contrast between the TV and its "off" screen) and I'll carry on with more furniture (kitchen appliances are visible on ealin_1, but not ealin_12). I'll finish with the flowers and the tea set.
If you have some feedback/comments let me know.
-
You never fail to amaze me Mayo. Keep up the good work!
-
Really pretty!
Great job there!
-
Update! Got some furniture done in the kitchen area, and made a render from the ealin_1 camera.
(http://i.imgur.com/V4hqCCC.png)
I have also made a bit of work to make a "fall-themed" still life painting, it will be in the next update. Now, I have to work on the tea set and the flowers - it shouldn't take all that long since most of it is already UV-unwrapped.
-
It is crazy how good you are getting at this, looks immaculate ;D
-
Lighting and textures are amazing! keep it up!
-
Update!
Well, ealin_1 and ealin_12 are essentially done. I may need to tweak a couple of things (such as the volumetric absorption of my tea material, which is currently too high), but it would be the matter of minutes.
I decided to change the lighting coming from the left window, as I felt the sun beams were coming at too horizontal an angle. Besides, this new lighting is more in accordance to that of the outside scene (eals_1).
Anyway, here's the almost-final version (from ealin_1 camera):
(http://i.imgur.com/4BDYYYV.png)
I hope you like it. Now, I'll have to start to work on ealin_2 (top floor) which is the last scene I set out to do for the year.
-
I like it a lot.
My only humble suggestion is - and it's a general gripe I have with many games - to lower the refraction for mirror-like surface. The way it is now it looks too clean and to me, it feels like the pavement extends to the wardrobe's doors.
-
I like it a lot.
My only humble suggestion is - and it's a general gripe I have with many games - to lower the refraction for mirror-like surface. The way it is now it looks too clean and to me, it feels like the pavement extends to the wardrobe's doors.
I appreciate you feedback, although I am actually not sure I would change anything to the current glass material, and here is why:
First, I would say that glass materials in games or in CG renders used to be handled in different ways, because a proper glass shader takes quite a bit of processing power. In video games, up to the previous generation of consoles, glass shading was roughly approximated, because it was too demanding to process (and keep a steady framerate), and it was often treated as a reflective (specular) shader. Proper glass shading in video games started (I think) with the last games of the PS3/Xbox360 generation (and earlier on PC of course).
For a proper glass shader, you have to decompose into the refracted/reflected rays, which depends on the angle at which your seeing the glass material (Fresnel laws), and as a rule of thumb the more you're going to look at the material from a normal angle, the less reflection you're going to have (conversely, the larger the viewing angle from the normal, the more reflection you have). Additionally, whether you see "through" or a reflection will depend on how bright the environment is before and after your glass panel. Assuming a proper glass shader, given that this scene and most video games look at the environment from "the top", if you look at vertical glass panels, you'll have a great angle from the normals of the panels, so you'll see mostly reflection. But that's physically accurate! Generally, the effect is enhanced if the light comes from the direction of your view into the glass material. That's why, in general, when you're outdoors by daylight you see only reflections on the windows of building because the light outside is much greater than the light inside the buildings.
I can try to tweak things a little and see if I can increase the refraction/reflection ratio, although I am not expecting drastic changes (and as I mentioned, it will be less physically accurate). For one thing I could make the glass not cast a shadow onto the objects inside the closet, which should make these objects more illuminated. We'll see.
-
Thank you very much for your detailed explanation. I appreciate it. Technical but on layman's terms for my benefit. I guess that a way for it to be better (in my opinion) would be... how to say... make the contrast between the surface and the reflected surface a bit more prominent somehow? Perhaps it will be less accurate but at the same time I won't be thinking that the pavement extended into the wardrobe... ;)
-
The only other suggestion would be to have a non glass material. A material that is almost transparent but doesnt have the refraction/reflection of glass. I would be interested in seeing this. Remember we arent looking for photo quality renders and not everything needs to make logical sense as in real life if it looks good.
-
The only other suggestion would be to have a non glass material. A material that is almost transparent but doesnt have the refraction/reflection of glass. I would be interested in seeing this. Remember we arent looking for photo quality renders and not everything needs to make logical sense as in real life if it looks good.
FF7 should be as much realistic as possible, at last that was the plan of the dev team. However, my problem with this very strong looking reflection is that the character model won't be mirrored on it, which will be more noticeable with this. So a weaker reflection would be better indeed.
-
Realistic is fine as long as it looks good. If it doesnt then what is the point? This is a fantasy game after all
Sent from my SM-N900W8 using Tapatalk
-
My approach to modelling is to be able to have something looking realistic AND good - or at least tend to the best compromise I can manage. I also think we can't really be bothered by seeing environment reflections and no character reflections, because the first thing is that character models do not even cast shadows. I'll see what I can do for the kitchen closet of that scene.
Otherwise, I wanted to present some early works on ealin_2 (top floor), and I have decided on several drastic changes with respect to the original scene (a bit like my works on mds5_w). My motivations for these changes were:
- Make the top floor consistent with the fact that it's located under the slanted roof. Thus, the walls are slanted in the top part.
- Make the light sources consistent with the windows layout from the outside scene picture. At this point, the sun beam source hasn't been made consistent with the one of ealin_1 and ealin_12. Additionally, I believe that extra light sources such as those from light bulbs will be needed.
- Ensure scaling consistency. This means that the scaling should be consistent with the bottom floor of the house. Actually, they even share some of the mesh, the fact that they're using the same stairs objects is the most blatant in this case. However, I haven't shown the stairs on the render below. This scaling process revealed a very big problem: the size of the beds had poor scaling, or at least one could say they were scaled for small chibis. Meaning: the beds were about 80 cm wide (which can be plausible), but with the original aspect ratio the same beds would be about 1.20m long! Thus I went for a drastic redesign of the beds, with an aspect ratio more in the line of 1.9 m x 0.8 m. I made some in-game testing to see what parts of the walkmesh I could use for making the beds longer. Because of the longer bed in Aerith's chamber, then I do not have the room for the chest of drawers, which I decided to replace with a shelf (I am thinking of stacking several of these shelves, actually).
- By the way, my works on scaling gave me an order of magnitude of the characters' sizes for ealin_1, ealin_2 and ealin_12 (with Makou reactor tool), using Kaldarasha's material. I obtained good settings with 1024 size for ealin_12 (yeah, that much - but it does look much better), and about 768 for ealin_1 and ealin_2. I suppose these sizes could be further optimized, but they're a good starting point.
- One thing that is a bit annoying: if the scale of the character models is adjusted for ealin_2, then there is a problem with the animation of Cloud getting up from bed. More particularly, there's a discontinuity between the end of the "getting up" animation and where the character is placed afterwards, as he's walking out of the room. It'd be great if someone could look into it.
Anyway, so here's the picture:
(http://i.imgur.com/5GUV5OS.png)
Still quite a bit of work to be done - unfortunately work has been piling of late, hence less time (and much less energy) to do much Blender on my evenings.
-
Realistic is fine as long as it looks good. If it doesnt then what is the point? This is a fantasy game after all
True. That's why I said as realistic as possible. It needs to be immersive so that the player won't draw out of the fantasy and realize that he is playing a(n old) game.
My approach to modelling is to be able to have something looking realistic AND good - or at least tend to the best compromise I can manage. I also think we can't really be bothered by seeing environment reflections and no character reflections, because the first thing is that character models do not even cast shadows.
I think adding drop shadows isn't to hard to make. All what is needed is to add a shadow model to the current position of the model, but I'm not sure if it is really that easy. ::)
- One thing that is a bit annoying: if the scale of the character models is adjusted for ealin_2, then there is a problem with the animation of Cloud getting up from bed. More particularly, there's a discontinuity between the end of the "getting up" animation and where the character is placed afterwards, as he's walking out of the room. It'd be great if someone could look into it.
If I remember rigth there are used two models of Cloud for this. I think after the 'Out-of-the-bed'-animation a store the current position of the model can solve the problem, if the playable Cloud model will be placed at the saved position.
Keep in mind to change the field scale, too. Or else the movement animations will have a slight sliding look.
-
Progress update (I haven't revised the lighting setup though)...
(http://i.imgur.com/jN9NK2Y.png)
-
Really really nice! I must congratulate you, Mayo Master. I hope to see more from you.
-
Lovely lighting!
-
Really really nice! I must congratulate you, Mayo Master. I hope to see more from you.
Thanks! I hope I can keep things going, but mostly I'll need people to join me on this endeavor (600+scenes is way more than I can chew). I have some plans to showcase my works in collaboration with Kaldarasha, so that we could eventually see about attracting people to the project. We'll just have to see how this pans out - it could really go either way. My goal is to be able to put this short "demo" together by early 2015. Come to think of it, it is very similar to what was initially intended with the "Bombing mission" project.
Lovely lighting!
Well... I was meaning to change it quite a bit :P
Namely, my plans would be to kill the sun light and add indoor lights (as if from light bulbs). The way things are, I don't see it sensible to have Cloud sleeping in a room which gets a direct sun beam.
I'll try to get things going for that scene, despite the work load and the release of Dragon Age Inquisition :P
-
Dat finished wood look, looks like the real thing. Kind of crazy to think that they even had access to trees to make that banister inside Midgar in the first place.
I'm definitely getting these puppies into qgears when they are ready. Awesome awesome work
-
...my plans would be to kill the sun light and add indoor lights (as if from light bulbs). The way things are, I don't see it sensible to have Cloud sleeping in a room which gets a direct sun beam.
I'll try to get things going for that scene, despite the work load and the release of Dragon Age Inquisition :P
:o :o :o
...Kill the sun??? is that our solution against global warming??
Beware with doing that, no on a more serious note, if my mind remembers correctly, the field is faded to black when the team goes to sleep. This in a way can be considered as becomming night. But...
When they wake up, it fades back from black to the normal field. Basically I would consider this to be morning....
So why kill the sun?
8-)
-
Beware with doing that, no on a more serious note, if my mind remembers correctly, the field is faded to black when the team goes to sleep. This in a way can be considered as becomming night. But...
When they wake up, it fades back from black to the normal field. Basically I would consider this to be morning....
So why kill the sun?
8-)
Precisely. When I speak of "killing the sun light", I mean to remove the light causing a very bright sunbeam inside the place. If you want to show that Cloud goes to sleep in the evening, and wake up in the morning, the corresponding sunbeam would have to come from nearly opposite directions, which is not possible since the same field images are reloaded. By removing the sun beam, I would retain an illumination from the outside which would correspond to twilight (from the "enviro light", or "sky light"), which could work to describe the place both when Cloud goes to sleep (after sundown) and when he wakes up (before sunrise).
Actually, for some time on the back of my head I have been having the idea of a "day and night cycle" mod. It's obviously not a priority by any means, but I wonder how much interest that would get. I have a few ideas about how it could be implemented (conceptually speaking - I don't have enough insight to see how the code could be altered accordingly), and we'd have to see if/how it could affect gameplay (opening/closing of shops, monster encounter rates, etc.). Maybe I'll open a thread elsewhere to discuss that in details.
-
Actually, for some time on the back of my head I have been having the idea of a "day and night cycle" mod. It's obviously not a priority by any means, but I wonder how much interest that would get. I have a few ideas about how it could be implemented (conceptually speaking - I don't have enough insight to see how the code could be altered accordingly), and we'd have to see if/how it could affect gameplay (opening/closing of shops, monster encounter rates, etc.). Maybe I'll open a thread elsewhere to discuss that in details.
Technically speaking it would be interesting to have an engine which complies more with FFVII, where it is possible to overlay the walkmesh, and all other 3d objects with movies as background. This should already be a possibility with the current engine, as many of the current mvies already have this (think of the intro, or mkup).
An other possibility would be more in line with the limited way the backgrounds are made, but simply putting more layers in there, one for day and one for night. But then we need to be able to switch the layers. My guess would be that this needs to be done at the scripting level...
just my thoughts anyway. :)
But this is not impossible I would say.
8-)
-
This basically the primary reason why my output is so low.
You create very intricate renderings but they simply do not match with the rest of the game.
I would like to be able to cast shadows and relections like REMake and Zero does and create nice looking characters that blend with the HD backgrounds.
The way things are now it always seems as if you throw two different games in a blender.
I also want to annouce that I will not do much modeling in the near future and want to focus more on texturing, shading and lighting.
Reason for this is, that I am suffering from severe necrosis in my right hand and moving the mouse a lot, well, simply hurts my wrist more than the pen does.
-
My understanding is that, with the current game engine, the best that can be done for character models is baked bumpmaps and specularity into the diffuse shaders, and then have the characters with very basic drop shadows. If we want something more sophisticated than that, we'll have to look at another game engine (QGears?).
My approach is to do the best I can with the current tools and engines available, and keep an eye on what else is developed. If we happen to end up one day with an engine that allows for a real time 3D rendering of environments interacting with characters (shadows and reflections), the backgrounds I've made could potentially be integrated in such an environment (now that Cycles does have texture baking...). But that's not my priority - I am patient about seeing how this all pans out.
-
By the way, a quick "micro-update" on ealin_2: I finished modelling all the flowers (there are 6 bunches in total). It took quite a bit of time. However, it hasn't changed the overall look of the scene drastically, so I'll just wait for being done with a bit more texture work (vases, beds, etc.) before I show an updated picture.
-
Update!
Since the previous one, all the flowers and corresponding flower pots are done, lighting was revised, and some of the furniture was textured. Here's what I have so far:
(http://i.imgur.com/mT4cbUP.png)
However, I ended up with an odd "artifact" as I have the bottom right part of Aerith's bed frame which has a weird illumination (it's oddly more white than it should), I haven't really puzzled out why that happens. The artifact appeared after I brought in a point-type light source as an indoor source of light (light bulb), but the light emission isn't strong and rather have an orange tint, so I don't really get why it reflects so much white. I'm also not sure if it will matter though, after I add the mattress, bed sheets and covers. We'll see.
-
Simply lovely texture for the wood! Great job!
-
Looking great! Also just as a fyi, the scenes of yours i have tested look great in 4k as is. I doubt we need to render larger for it.
-
Thanks! Well... if your monitor isn't huge, I don't think you can tell much of a difference between HD and 4K for whatever is displayed. Also, my scenes are rendered at 5 times the original, while theoretically HD is 4.5 times the original.
Most of the stuff I have made should still be looking good at proper 4K renderings (9 times the original scale), because I took care of having fine enough textures (most of the time, at least). But for actually re-rendering the stuff, indeed that's not a priority.
-
Its a 28 inch. So not massive by any stretch. Its difficult to read text without scaling of some sort at my normal distance from the screen
-
Update! I finally got the bedding done - it took quite a while, I wanted to use cloth simulations and a lot of them were not producing the intended results (broken meshes, force field issues...).
Here's where I'm at:
(http://i.imgur.com/v7UAhVy.png)
So... I still have to make the pillows, the doors, and some random junk in the cardboard box in Elmyra's room, and that's about it.
-
That is stunning. Good work
-
Just been scrolling through this thread and I am incredibly impressed with the quality of your work.
Also I havent tested it, however isn't it possible to edit the walkmesh via Makou Reactor?
I know you can via Deling in FF8 and quite sure I remember seeing a similar feature in Makou Reactor.
In which event you could do architecturally sound field designs and if necessary adjust the walkmesh c:
-
Thanks!
Well... so far I am working with the assumption that the walkmesh is not editable. I would like to finish this set of field scenes under that assumption (if I keep reworking scenes forever, I'll be getting nowhere). As far as I've been told, there's no easy way to make edits, I don't think Makou reactor can provide that option - that being said I haven't taken the time to explore the possibility myself (and let's be honest, I don't have the time :/ ). With walkmesh editing, I think that scripts would also have to be modified, so I don't really see that being straightforward. However, so far Makou Reactor offers the option to scale characters easily, which helps a lot when solving scaling inconsistencies.
If in the future, if it turns out that we have a tool which could edit the walkmeshes somewhat easily, I'd be inclined to reconstruct some of the scenes. That's the sort of stuff I have on the back burner ;)
-
Haha, yeah I havent played with Makou Reactor much myself as with packages like Bootleg I didnt need to do much tedius work within the field LGP's,
however I am quite sure I remember seeing the walkmesh within Makou Reactor with a similar design to how the Walkmesh editor is within Deling for FF8 with extensive amounts of triangles that make up an invisible mesh on the field images
Yeah I can imagine how much work this already is for you especially with the quality standard you are displaying,
I've tried messing with Blender in the past and couldn't use it, the camera controls and general user-friendliness of the programe were a nightmare for me haha
So you have extra respect for getting that programe working with breeze
-
About Makou reactor, maybe you could ask on the dedicated thread on the Tools section. You're probably right, but it would take someone to dig into the stuff to explore how easily the walkmeshes could be revised.
As for blender, I picked it up 2 years and a half ago - if you've browsed this thread you probably saw some progression on the way. I was lucky that I tried to get started after the 2.5 version; the interface looked a lot less user-friendly before that point. That being said, the type of work I'm into does not involve much camera control - on the contrary we want it to be rigged (SpooX's Shinratool script providing all the data in most cases). I admit that getting familiar with shortcut keys was the first necessary step, but once you know these keys then I do find Blender rather easy to use.
Oh, and by the way, bonus gift as I needed to generate a texture for ealin_2 ;)
(http://i.imgur.com/7XvLbWD.png)
-
Oh man the nostalgia, I immediately pictured Aerith " don't step on the flowers!!!!"
Those flowers are so detailed O.O
It's really surprising to be honest that square doesn't just do this with their PC ports of classic FF games,
simply recreate the scenes, animations, character models, textures, UI and fans would actually be satisfied with the quality of the ports :roll:
It's really amazing the quality of content the modding community creates, it's just unfortunate that square leaves us feeling it is necessary to spend our own private time patching a game that we paid for
Still wonderful work as always really am loving the detail especially when compared to the original :DD
perhaps I will try to learn blender again then start overhauling the character models in game with more anatomically realistic ones
so we could have character models that fit in with excellent quality environments ;D
-
perhaps I will try to learn blender again then start overhauling the character models in game with more anatomically realistic ones
so we could have character models that fit in with excellent quality environments ;D
That won't be enough. I'm already on the limits of the games possibilities (there is only room for textures). The main problem is the skeleton and the animation files with there Z-position. If we could edit them a bit better and could use wight painted models with a unified skeleton, then things would look a lot better. But we need also tools for this. I guess mesh deformation is possible with the current engine, because it seems to be used in Safer Sephiroth special attack, but someone has to enable it for the models and expand kimera to display such models.
-
Oh man the nostalgia, I immediately pictured Aerith " don't step on the flowers!!!!"
Those flowers are so detailed O.O
It's really surprising to be honest that square doesn't just do this with their PC ports of classic FF games,
simply recreate the scenes, animations, character models, textures, UI and fans would actually be satisfied with the quality of the ports :roll:
It's really amazing the quality of content the modding community creates, it's just unfortunate that square leaves us feeling it is necessary to spend our own private time patching a game that we paid for
Still wonderful work as always really am loving the detail especially when compared to the original :DD
perhaps I will try to learn blender again then start overhauling the character models in game with more anatomically realistic ones
so we could have character models that fit in with excellent quality environments ;D
Actually, I find that remaking FFVII from Square Enix standpoint would not be straightforward at all. If they wanted to do a HD port, they would pretty much need to recreate all the assets (a/ apparently they've lost a lot of the assets of the original game b/ said assets are for the most part completely archaic now), and get it to work with a new engine... and at that point, they might as well go for a full-fledged remake. Not to mention that a bunch of elements of the game (storywise or gameplay) were widely acceptable when the game was released, but the current customer-base wouldn't be able to stand it. I also believe that FFVII is unfortunately plagued by the syndrome of the unpleasable fanbase (http://tvtropes.org/pmwiki/pmwiki.php/Main/UnpleasableFanbase), in that respect I agree with Square's decision to dedicate resources to entirely to new games.
What would be nice, however, would be a more supportive attitude towards the modding community, because whatever we do, we'll always be at the mercy of a Cease & Desist.
That won't be enough. I'm already on the limits of the games possibilities (there is only room for textures). The main problem is the skeleton and the animation files with there Z-position. If we could edit them a bit better and could use wight painted models with a unified skeleton, then things would look a lot better. But we need also tools for this. I guess mesh deformation is possible with the current engine, because it seems to be used in Safer Sephiroth special attack, but someone has to enable it for the models and expand kimera to display such models.
Indeed. And then there are severe limitations when it comes to shading (we can only do so much with only diffuse shaders)...
Oh, and by the way, it is DONE :)
(http://i.imgur.com/eSAxdim.png)
-
Hmmm! Flowers smell nice!
-
You've really captured the essence of the original which hit me right in the feels, just like when I played the psx game when I was 8. Do you think you'll release the set by the new year?
-
Even though everything is 99% done when it comes to modelling and texturing, I'm afraid I won't have it ready before the new year. Here is what still needs to be done:
- Make a few textures. There are a couple missing in mds5_m, mds5_w, 5min_1, 5min_2.
- For each scene, the final renders (with very high amount of samples) plus the re-renders each corresponding layers. Even though rendering time is much less than modelling time, it's still non-negligible.
- There is still a lot of layering, cutting out layers and testing them in-game. Sometimes the way the original layers are cut is not actually optimal. So, there's a loop render-assemble layers-test in game-back to re-render which needs to be done. Of course, I can quicken the process by making test renders with few samples, and only go for the final renders when I know exactly what to do.
However, I'm aiming to have that done by early January, and finishing these scenes is my priority (meaning: I'm not starting anything else) - except for enjoying Xmas holidays, that is ;)
-
I've been messing with 3D fields in QGears, here is what I could hack together in an hour using one of Mayo's scenes:
(http://s21.postimg.org/nztm67l6t/export.png)
(http://s24.postimg.org/esyesd7c5/export2.png)
(http://s8.postimg.org/ivkzifpj9/wire.png)
-
Thanks for sharing. It looks like what you got are the shaders from the Blender Internal engine, whereas the scene was made for Cycles (I have an older version of the same scene with Blender Internal shaders, but it does look very good in comparison). By the way, it is possible to do texture baking in Cycles (http://www.blenderguru.com/tutorials/introduction-baking-cycles/), so maybe you'd like to check that out. Unless you'd like me to upload that older version to the repo?
-
54 fps... not too shabby for just dumping the scene with no optimization. :mrgreen: Which blender plugin did you use to export the scene, i've found a few but all seem buggy. But I also have no idea how to use blender-soo many buttons
-
Thanks for sharing. It looks like what you got are the shaders from the Blender Internal engine, whereas the scene was made for Cycles (I have an older version of the same scene with Blender Internal shaders, but it does look very good in comparison). By the way, it is possible to do texture baking in Cycles (http://www.blenderguru.com/tutorials/introduction-baking-cycles/), so maybe you'd like to check that out. Unless you'd like me to upload that older version to the repo?
Ah so thats what its doing, if you have a blender internal version that'd be cool (to see how it looks :)). I don't have access to the repo though, I was just given this one file directly.
@Knife, I dunno what I'm using, I just used whatever the first google result was.
Edit: Oh btw about the 54fps, that was just a dip, it hits 120fps easily, and in a few years you'd prob be able to get 250fps easily too. If the project ever went down the real time 3d route then optimization probably wouldn't matter due to the hardware speed when its all finished being so much faster.
I was thinking that it would actually be technically possible to get FF7 itself to render 3d fields with hook that replaces the rendering functions.
-
Could I get access to a blender scene or two? I'm messing around with/testing the ogre3D side of qgears-there are some pretty interesting guides over at their wiki o.O
-
I was thinking that it would actually be technically possible to get FF7 itself to render 3d fields with hook that replaces the rendering functions.
Theoretically we could use 3d backgrounds as a character model. Problems are only to place them correctly on the field and the texture management of the game. It would easily blow up the size of the char.lgp.
-
I was thinking more external files, for example Aali's driver has that "direct" mode, well it could say if the file im loading is "reactor.field" (or whatever) then disable the 2d field rendering and load all of the meshes and things from some where else and render those itself. It would require hooking a lot more stuff in the game though.
-
Even if it could be done what would be the point? The renders that we do take days to complete and look vastly better then anything that could be done realtime? Its not like we could use the scenes to create fully rotatable 3d.
-
They would probably look better - in qgears at least with updated models too and real time shadows. And more future proof, one day even the best we can do now will look low res (say when 4k is the norm) at least polygons tend to scale better.
-
I don't see the harm in experimentation. :mrgreen: Real shadows from an actual light source would be really awesome. Having a higher range of motion running around the field would also add a touch of newness. Animations for standing still and a smooth transition from running to standing still or running to walking would also be a nice touch.
-
I was not suggesting not trying it out in qgears, I just dont think using aalis direct mode and the ff7 engine itself would be of much benefit.
-
Oh I'm sorry, I thought we were talking about qgears. Qgears is by far, a better time investment, it's kind of crazy to think what it is capable of. It just really needs the rest of those damn function reversed first
-
Ah so thats what its doing, if you have a blender internal version that'd be cool (to see how it looks :)). I don't have access to the repo though, I was just given this one file directly.
I can try to send you the older version with Blender internal shaders.
By the way, I have a question: do you know how the lighting and shading works? Mostly, I would like to know if there are ways to do better than what Blender internal engine can do - physical lighting (i.e. with light bouncing on objects to provide indirect light sources) and the kinds of shaders which can be used are important.
By the way, I'm into the layering process of ealin_2, and it's a pain in the butt. The way the layers are pre-cut is very annoying: the doors are glitchy, I even had to redesign them. For that matter, the original field has glitchy doors, if you pay attention (namely when the doors are open). I think I may have a workaround, but it's not straightforward (meaning, I have to deviate from how the scenes are originally pre cut for it to look decent).
-
I can try to send you the older version with Blender internal shaders.
By the way, I have a question: do you know how the lighting and shading works? Mostly, I would like to know if there are ways to do better than what Blender internal engine can do - physical lighting (i.e. with light bouncing on objects to provide indirect light sources) and the kinds of shaders which can be used are important.
By the way, I'm into the layering process of ealin_2, and it's a pain in the butt. The way the layers are pre-cut is very annoying: the doors are glitchy, I even had to redesign them. For that matter, the original field has glitchy doors, if you pay attention (namely when the doors are open). I think I may have a workaround, but it's not straightforward (meaning, I have to deviate from how the scenes are originally pre cut for it to look decent).
I'm not sure what you're asking? The lighting and shading in ogre/qgears? I think there is a material editor: http://www.ogre3d.org/forums/viewtopic.php?f=11&t=48774
-
According to ththe Complete Guide: Exporting Blender to Ogre from the ogre wiki, dotSceneLoader can deal with the lighting(?)
-
I think I (or someone) would need to focus a lot about how to export Blender cycles scenes into Ogre properly - it doesn't seem straightforward if you what to get the shaders right as well as radiosity (or physical lighting).
On a more important note: ealin_2 is done (ready to be played in game) and uploaded to the repo. I'll make a more open release later (probably making a notification on the Graphical mods forum, as well), and if you want to try the scene without access to the repo, you can PM me and I'll email it to you. Some comments:
- Unfortunately, the way the layers are pre-cut created some severe limitations about how to get the doors closed and open. The way it is set in the release is the best that can be achieved, in my opinion. In more details:
- the layers of the doors (closed and open) are not laid over layer #2. As a consequence, if you get the thickness of the doorframe as part of layer #2, it would be laid over the closed door, which is not a desirable result. For that matter, in the original field files, the open doors look glitchy because of the overlay of layer#2 over the open door layers.
- the layers of the open doors are pre-cut in such a way that it does not cover the entire area where the wall should be shaded by the open door. At first, I thought I could shade the wall appropriately depending on whether the door is open or closed, and after a lot of testing, it is doable for only one of the two doors. For the sake of consistency, I preferred not to make any shading of the walls by the doors.
- The lighting and the windows are set in such a way that it is consistent with the exterior look of the house. There is a mismatch with the original field, which features a window on the right hand side (with an overlay of crepuscular rays), because the original field is not consistent with how the windows are located on the exterior. I initially thought of adding crepuscular rays so that it would match the updated location of the windows, but this is not possible because of how the layer of crepuscular rays is pre-cut into the scene.
- The bottom line is: there is a number of small imperfection which cannot be corrected, and the release is the best which could be managed.
Lastly, below is a picture of the final render for the base layer. I hope you'll enjoy :)
(http://i.imgur.com/Hc85x8v.png)
PS: I personally recommend you play with Kaldarasha's models, and edit the characters size with Makou Rector. I personally use 768 for the character size (512 is the default size)
-
Alright, folks! ealin_1 is done and uploaded to the repo. Again, please PM me if you want the files while you don't have access to the repo - I'll make the big "Public release" once everything's done.
A few comments:
- I actually redesigned the kitchen area. The decision was made because of very poor perspective effects. Because 2D stuff is laid over the 3D models, if some 2D elements (like the image of the kitchen counter) is shown over the 3d model at an ill-defined location, the results are poor. In this particular instance, the kitchen counter was seemingly as high as Cloud's shoulder, which doesn't make sense. Matters were further complicated by the way the layers are pre-cut, and it took quite a bit of trial and error to figure a nice way to redesign and rearrange the kitchen furniture so that it gives a good impression of perspective, without any kind of glitchy appearance. In the end, I was able to get a satisfying result with no glitchiness, so I'm quite happy about it :)
- I changed the color of the crepuscular rays (more yellow) which makes more sense considering the lighting setup. It was rather easy, thanks to the new "SunBeam" compositing effect of Blender's newest release.
Below is the result:
(http://i.imgur.com/bfqx6dX.png)
Please enjoy! :)
Next moving on to ealin_12.
PS Again, using Kaldarasha's model, I recommend 768 for the character size using Makou Reactor for this scene.
-
And finishing the indoor scenes of Aerith's house with ealin_12. It's uploaded to the repo. Again, should you wish to test it and you don't have access to the repo, please PM me.
Here's the final result:
(http://i.imgur.com/U8nUyLR.png)
Using Kaldarasha's models, I find that a size of 1024 (thus, twice the original size!) gives good results.
I hope you'll enjoy!
-
Will notes about the sizes be in the final release? I'd like to make all the corrections you recommended, But i won't be playing until Beacause is fixed.
-
Blender 2.73 has been released and it looks like they fixed performance and lighting problems
-
Will notes about the sizes be in the final release? I'd like to make all the corrections you recommended, But i won't be playing until Beacause is fixed.
Of course :) My aim is to get the "best possible experience" with the tools available, and tuning the character size is an important part of it, so the notes on the character sizes will be mentioned (or written in a Readme.txt or something).
-
Hey guys,
I was working on finalizing eals_1, but it looks like there are some problems and I could certainly use some help - namely, if you have some insight about how field background animations are handled, it would be useful. The problem in this scene is about the animation of the waterfall, in the back.
So, here are the details:
- When eals_1 has the layers decomposed in Palmer, it looks like there are 3 images related to the waterfall animation (eals_1_0_00065664, eals_1_0_00065665, and eals_1_0_00065666). These images look about the same, I am guessing the waterfall animation should be made some sort of loop in alpha value between these layers (or something of the sort). Now, when importing the images in Palmer and asking Palmer to write the modpath textures, Palmer outputs 5 files (eals_1_00_00, eals_1_01_00, eals_1_02_00, eals_1_15_00, eals_1_16_00).
- However, when loading the scene, an error message is displayed: "Glitch: missed palette write to external texture field: eals_1/eals_1_15". As a consequence, the image of the waterfall is static, there is no animation.
- I wondered if the glitch was caused by an incorrect file name (because eals_1_15 seems "missing", while the image file is actually named eals_1_15_00). So I renamed the last two files as eals_1_15 and eals_16, respectively.
- When loading the scene, the error message is no longer there, and the waterfall is animated. BUT: it is actually the original waterfall animation images which are displayed, NOT the modified images from modpath.
So... is there a way to fix this issue? As things stand, either we get a crappy low-resolution animated waterfall, or a high res waterfall which would remain still (along with the Glitch message).
Any help would be greatly appreciated. If you think the question may be better asked elsewhere, please advise - Maybe Omzy or yarLson would know).
-
It is the problem with the transparent layers yarLson has mentioned on Aali's driver thread. Atm these layers can't be animated (problem with handling a dynamic alpha value I guess).
The last _00 is to mimic color variations which the textures have as extra information (mainly the magic textures). I wonder if the game use this or something similar to handle the dynamic alpha channel of a texture.
-
Oh. I'll do a bit more reading about this then, although you seem to say there's currently no workaround. I did manage to make the static transparent layers work, though. I have them in my release of ealin_1 and ealin_12. However, for it to work, the layer I have to import into Palmer is not processed with respect to its alpha value, the transparency is actually given by the level of black (in a way, I simply have to overlay my transparent layer over a black background to get it to work). But that doesn't seem to work for this case (most likely because this is a case of handling dynamic alpha values, not static).
-
The animation is done via field script, some set "parameter" states which seem to just swap tiles around, and some do it by changing the pallets of tiles at runtime. The waterfall appears to be done via the pallet method.
From Makuo reactor tool I think its this script:
Load the palette #8 in the position 0 (color count=256)
Label 1
Copy palette (sourcePal=0, targetPal=16, color count=256)
Multiply RGB(130, 130, 130) on the colors in a palette (sourcePal=0, targetPal=16, first color=Var[5][0], color count=10+1)
Multiply RGB(162, 162, 162) on the colors in a palette (sourcePal=0, targetPal=16, first color=Var[5][3], color count=10+1)
Load the position 16 in the palette #8 (color count=256)
Var[5][0] = Var[5][0] + 12 (8-bit)
Var[5][3] = Var[5][3] + 9 (8-bit)
If Var[5][0] > 100 (else go to label 2)
Var[5][0] = 0 (8-bit)
Label 2
If Var[5][3] > 138 (else go to label 3)
Var[5][3] = 0 (8-bit)
Label 3
Wait 1 frame
Go to label 1
Return
Seems like there is another script in that field related to this kind of animation too, I'd have a try with this tool? Perhaps you could add more layers/states and update the animation to use that instead?
-
Thanks paul. The problem is also discussed on the forum thread about Aali's driver. One of the problem is that these odd animations (now that I think about it, thank goodness the flow of the sewer scenes were not relying on that!) are handled by changing color values on an 8-bit color palette, and there's a problem with these palettes:
there's no hardware from this millenium that supports paletted textures
One solution may be to recreate such color palettes - although ultimately they'll be in 8-bit colors, so I'm not sure if such a solution might be decent-looking in the end. I tried to convert a test .png into such a color palette (my antique version of Paintshop can do that), which prompted a brand new error message when loading that scene in the game "Error: Wrong color". I may be onto something, I'll make a test where I set the colors of the png to match the rgb values listed in the script and see what happens.
In the meantime, I think the best for me is to release eals_1 without the waterfall at all (and even prevent it from loading the original palettes), I think the scene looks good as it is. Adding the waterfall would be extra bonus when we actually have the means to do so.
-
So there are only issues with the animations when a color cycle is being used? can't we change the script part as a workaround?
8)
-
So there are only issues with the animations when a color cycle is being used? can't we change the script part as a workaround?
8)
This would demand to create new files for the waterfall animation and these files must be in the flevel.lgp as well.
I wonder why Aali doesn't use the same trick as he use to mimic the color pallets for magic effects? Like **_15_00, **_15_01, **_15_02, etc. If this would work then Palmer may need an update too to extract the different palettes as textures.
-
This would demand to create new files for the waterfall animation and these files must be in the flevel.lgp as well.
I wonder why Aali doesn't use the same trick as he use to mimic the color pallets for magic effects? Like **_15_00, **_15_01, **_15_02, etc. If this would work then Palmer may need an update too to extract the different palettes as textures.
Going by the field script above, the issue is that the palette is being changed through code: it's not like the file contains 3 different palettes that get switched between (so you could just have 3 PNG files, one for each version). Instead, some script code changes part of each palette (and because it's script, in theory it could change it by a different amount based on variables, so there could be hundreds, or thousands of different possible palettes that get generated.)
In practice I guess that actually there are a limited number of variations of the palette that actually get set up by the script, but that's why there's no simple solution, if I understand it correctly.
(If I were coding a solution to this - which I'm not - then I'd probably set up the fragment shader to effectively emulate a paletted texture; have the actual data as an integer texture, and send a lookup table to the shader containing the palette data for it to translate. Would have reasonable performance on anything even semi-modern. Certainly a little complex though.)
-
Going by the field script above, the issue is that the palette is being changed through code: it's not like the file contains 3 different palettes that get switched between (so you could just have 3 PNG files, one for each version). Instead, some script code changes part of each palette (and because it's script, in theory it could change it by a different amount based on variables, so there could be hundreds, or thousands of different possible palettes that get generated.)
In practice I guess that actually there are a limited number of variations of the palette that actually get set up by the script, but that's why there's no simple solution, if I understand it correctly.
(If I were coding a solution to this - which I'm not - then I'd probably set up the fragment shader to effectively emulate a paletted texture; have the actual data as an integer texture, and send a lookup table to the shader containing the palette data for it to translate. Would have reasonable performance on anything even semi-modern. Certainly a little complex though.)
QGears will have the same issue, I was thinking about building some meta data table that stores which tiles/pallet's are changed. Then it should be possible to generate "normal" texture based animations. But it would also mean rewriting some of the script opcodes so that they simply change the current animation frame rather than changing pixel data.
-
Going by the field script above, the issue is that the palette is being changed through code: it's not like the file contains 3 different palettes that get switched between (so you could just have 3 PNG files, one for each version). Instead, some script code changes part of each palette (and because it's script, in theory it could change it by a different amount based on variables, so there could be hundreds, or thousands of different possible palettes that get generated.)
In practice I guess that actually there are a limited number of variations of the palette that actually get set up by the script, but that's why there's no simple solution, if I understand it correctly.
(If I were coding a solution to this - which I'm not - then I'd probably set up the fragment shader to effectively emulate a paletted texture; have the actual data as an integer texture, and send a lookup table to the shader containing the palette data for it to translate. Would have reasonable performance on anything even semi-modern. Certainly a little complex though.)
Aye, you are right. I've absolutely forget about that. Aali did said himself that he can only think of a shader based solution for it.
-
Hi folks,
I've released eals_1, the files are on the repo. I'll upload a picture tonight to show you how that looks. Again, if someone happens to be interested in trying out the field, while not having access to the repo, you can PM me so that I can email the files to you. 3 notes about that release:
- Given there's no simple way to get the waterfall to work properly, I've released this scene without the waterfall.
- I changed the color of the crepuscular rays (yellow to emulate a winter sun)
- The default character size (512 in Makou Reactor) works well.
That wraps up Elmya and Aerith's house. Now, onto the Sector 5 slum market.
-
Wonderful release man, I wish I had the artist talent to join you all but anytime I have tried I have trouble making a cube haha!
Look forward to seeing all of sector 5 completed!
-
What a gorgeous scene once more! Aeris's house revamped! Astonishing work! Looking forward for the update!
-
Ok, here's a rendition of the eals_1 scene, I hope you'll enjoy.
(http://i.imgur.com/1Jla1rO.jpg)
-
Hi folks,
I've released eals_1, the files are on the repo. I'll upload a picture tonight to show you how that looks. Again, if someone happens to be interested in trying out the field, while not having access to the repo, you can PM me so that I can email the files to you. 3 notes about that release:
- Given there's no simple way to get the waterfall to work properly, I've released this scene without the waterfall.
- I changed the color of the crepuscular rays (yellow to emulate a winter sun)
- The default character size (512 in Makou Reactor) works well.
That wraps up Elmya and Aerith's house. Now, onto the Sector 5 slum market.
This means wall market is next?
-
This means wall market is next?
No, it's the big plaza of the Sector 5 slums with a cluster of shops in the middle, and the rusting bus turned into a weapon shop. And all related indoor scenes. In other words, mds5_1, 5min1, 5min_2, mds5_i, mds5_dk, mds5_i, and mds5_m.
I do not intend to model the Wall Market afterwards.
-
So nicely dirty! It's the slams anyway!!! :)
-
No, it's the big plaza of the Sector 5 slums with a cluster of shops in the middle, and the rusting bus turned into a weapon shop. And all related indoor scenes. In other words, mds5_1, 5min1, 5min_2, mds5_i, mds5_dk, mds5_i, and mds5_m.
I do not intend to model the Wall Market afterwards.
From your last progress on that scene I thought it looked completed :)
-
You might say it's 95% completed - the modelling and texturing is essentially done, but I need to make them playable. The remaining work to do is to make all final renders (currently rendering mds5_1, estimated total rendering time: 6 days! and I'll have to re-render the layers...), and do the job of layering which sometimes can be a pain (it took me a week to get ealin_2 correctly).
-
Quick update:
I'm into making the renders of mds5_1, which is taking a long time since it's the largest and most complicated scene by a long shot. The biggest chunk is done, though - I'd say that I'm through 70% of the rendering job, but I'll still have to cut the layers and arrange them.
In the meantime: I actually wasn't too happy with the way I made the crepuscular rays for eals_1, so I remade that layer. The files in the repo have been updated. Below is the result:
(http://i.imgur.com/KLFRP7R.jpg)
I hope that looks better to you.
-
Just letting you know: some unexpected delays on the mds5_1 layers. Basically, the renders I made for the extra layers were not showing the volumetric scattering effect, so I have to re-render all these things. I should have tested that earlier. Pretty much 5 days of rendering down the drain :(
EDIT: Found a workaround, so it wasn't all for nothing. Currently into adding post-processing effects (beams of light, light glows/glares, etc.). Hopefully I can have it finished by the end of the week.
-
Update: And yet more delays for mds5_1 >:(
After a lot of work, upon testing the scene in game, it turned out that the automated layer cut is causing an awful lot of problems: mostly there are many places where the background layer is not showing, leaving the area black. So... I even have many areas to remodel in order to "hide" the areas where the background is cut, by adding "stuff" on foreground layers. And then I'll have to re-render and re-cut all that.
Cheap %#*^......
What he said.
-
Hi folks,
mds5_1 is finally done and uploaded to the repo. Below is the final render with light layers:
(http://i.imgur.com/4N0s6yV.jpg)
Again, anyone who wishes to try it without having access to the repo, please PM me. I hope you like it.
As a side note: finalizing this scene has been tedious and painful. As a lesson for future scenes: when modelling a new scene, it is key to refer to the background layer as given by Palmer. Referring to this layer in particular is helpful in preventing the following issue: Considering the background layer given by Palmer such as the one below
(http://i.imgur.com/Sr4l1C2.png)
It is essential that the tiles in white shown on that layer be covered by objects located on a foreground layer. If not, it will instead reveal an area not covered by foreground objects, where the background layer will not be displayed either, resulting in black pixels. I had to go through an awful lot of troubleshooting to get that fixed, because I had been basing my modelling on the complete scene (where all layers are displayed together). The tiling process of the various layers is very limiting (not to mention the problem of alpha values).
Anyway, I need something more straightforward to do next - I've lots a few neurons and some hair completing that one.
-
Hi folks,
mds5_1 is finally done and uploaded to the repo. Below is the final render with light layers:
Again, anyone who wishes to try it without having access to the repo, please PM me. I hope you like it.
As a side note: finalizing this scene has been tedious and painful. As a lesson for future scenes: when modelling a new scene, it is key to refer to the background layer as given by Palmer. Referring to this layer in particular is helpful in preventing the following issue: Considering the background layer given by Palmer such as the one below
It is essential that the tiles in white shown on that layer be covered by objects located on a foreground layer. If not, it will instead reveal an area not covered by foreground objects, where the background layer will not be displayed either, resulting in black pixels. I had to go through an awful lot of troubleshooting to get that fixed, because I had been basing my modelling on the complete scene (where all layers are displayed together). The tiling process of the various layers is very limiting (not to mention the problem of alpha values).
Anyway, I need something more straightforward to do next - I've lots a few neurons and some hair completing that one.
Amazing work. You are a beast.
I believe the note about the white tiles also applies to FF8. Alpha channel shenanigans.
-
That is fugging awesome!
You really nailed the lighting on that scene
-
Excellent job!
Thanks for sharing!
-
So how much of the backgrounds are already finished? I would really like to see all those beautiful renders in the game, even if it's just the first few screens of the game. :)
-
The number scenes which have been finished, and finalized to be playable, has reached the amazing amount of... ehm, 7.
I have 6 other scenes which are almost done, where I have to finalize the layering and post-processing effects, and that's currently underway. I'd like to believe I can do these before the end of the month, but recent experience has shown that unforeseen problems may pop up at that stage.
Once I have the package of my Sector 5 scenes done (market, related shops, and Aerith's house), I'll make a "public" release.
anaho made a few scenes for Mideel, I think a couple of them are almost complete and playable.
And then there's Ulpians's scene where the layering and post-processing wasn't done.
So, yeah, our progress is about 9/650, soon to be around 15/650 :-o
-
Update!
The item shop scene mds5_i was updated (a couple of tweaks here and there), and uploaded to the repo (the tiled field files were replaced). Below is how the scene looks like:
(http://i.imgur.com/04nBJVS.png)
-
Update!
The mds5_dk scene, aka "Man in pipe" aka "This guy are sick" scene, has been finalized and uploaded to the repo. Below is the final render with the lighting layer:
(http://i.imgur.com/J8P7j6y.png)
I hope you enjoy.
-
Excellent work! I check here often hoping that there are new screens :-P
-
Amazing work Mayo! I can't help but notice you're using a Chinese font for 歓迎 and 漢 in the shop, I know it doesn't really matter given that it's a fantasy world but they feel a little out of place due to the inconsistency (or did the original scene use Chinese fonts for those too?).
-
Amazing work Mayo! I can't help but notice you're using a Chinese font for 歓迎 and 漢 in the shop, I know it doesn't really matter given that it's a fantasy world but they feel a little out of place due to the inconsistency (or did the original scene use Chinese fonts for those too?).
Actually, original scenes have a lot of kanji writings. For texturing, it's best to figure out what they were meant to be, so we can get the high-res version of them, so we do make some search. SpooX has been recently struggling with the original writings on the train car, for instance. For quite a lot of material, I was able to find information here:
http://www.glitterberri.com/final-fantasy-vii/map-analysis/map-analysis-part-5/ (http://www.glitterberri.com/final-fantasy-vii/map-analysis/map-analysis-part-5/)
and I've been trusting this source so far (sometimes double-checking with my girlfriend who has rudimentary knowledge of Japanese).
All in all, I've made an effort in order to get some Japanese writings which would make sense to a Japanese reader - which is very tricky since I've pretty much no knowledge of the language myself. If you have some advice in that area, please let me know. Or did you mean that the font was specifically Chinese and would be unused for Japanese writings?
-
Yeah sorry, the font in particular is Chinese, there are subtle differences in how the CJK characters are rendered depending on the font and that scene has a mix of both Japanese and Chinese fonts.
-
To be honest I thought that kanjis which were common to Chinese and Japanese could be written with whatever font. I have other fonts for kanjis which generally look more "modern" to my Westerner eyes, but in that particular case I wanted a more "rustic" kind of font for the kanjis you mentioned. If you could redirect me to someplace where I could find a variety of Japanese (rather than Chinese) True Type fonts, I'd appreciate.
If there is a place where Chinese fonts were possibly used in the original game, that would be in the Wall Market / Don Corneo's mansion. If you could please check out the corresponding scenes
http://www.glitterberri.com/final-fantasy-vii/map-analysis/map-analysis-part-2/ (http://www.glitterberri.com/final-fantasy-vii/map-analysis/map-analysis-part-2/)
and let me know? I don't have a discriminating eye for that. If there are Chinese fonts used in any of these, then maybe I could get away with a Chinese/Japanese font mix in mds5_i. If not, well then I guess I'll have to re-texture these.
Anyway, that's an area where I could definitely use some help :)
-
This clip from the unicode charts should give you a better idea of what I'm talking about
(http://i.imgur.com/lPTTrwy.png)
Take a look at 8FCE
To the untrained eye the characters are all pretty similar, but if you can read Chinese or Japanese the differences kind of stand out.
Don Corneo's might make sense in a Chinese font, although the feeling I always got from his den is that it's going out of it's way to be stereotypically Chinese, sort of how a lot of Chinese restaurants are decked out in Japan, in which case there is a very specific clerical script font that's always cheesily used. I'll have a look at all of the backgrounds and get back to you, perhaps we can find the original fonts used if they weren't hand drawn.
Ok so for the item shop I'd suggest a semi-cursive gyosho font for the welcome sign on the counter and the medicine jar at the back, something like Ricoh's HGGyoshotai. Here (http://www.hakusyu.com/download/kyokan.html) are some free fonts, the third and sixth are gyosho although these fonts only have the basic 1,000 or so educational kanji, so only 漢 is available. It should give you an idea for what to look out for though.
Incidentally those boxes at the back of the shop in the original field look like they say *配 to my eyes, can't make out the first character but it's a delivery company.
Corneo's place has a bit of a mix of stuff, but I think most of it would look good with a reisho font, the 5th font on the same page above is reisho. Then again if you want to stick as close to possible to the original design then perhaps only the golden lettering above the doors in the main hall should be reisho, the rest uses a gyosho font again, except for the gong outside which uses seal script (second to last font in the same list).
-
Wow, that's a lot of useful information! Thanks :)
You've given a good example, I see what the differences can be. I would guess a Westerner equivalent would be modified letters such as è, é, ç, þ, ţ, ž you can find in various European languages.
Thanks for the fonts, they're going to be useful. I'll see about re-texturing these elements of the item shop then.
If possible, I would very much like to have your help for "Quality Control" purposes - I have to rely on knowledgeable people if I want to keep Japanese writings in my scenes and have them make sense.
By the way, if you could double-check what I had already considered "final", that'd be much appreciated: in the indoor scenes of Aerith's house (a couple of pages earlier), I have textured many books and folders where I tried to write a title, I'd like to make sure it's consistent (and that I didn't write gibberish). I have also a number of scenes "in progress" where I'd like your review before I launch the final renders.
-
I would also recommend to use a handwriting japanese font instead. I did a quick google search, you can download some here: http://www.freejapanesefont.com/category/handwriting/ (http://www.freejapanesefont.com/category/handwriting/)
Keep up the good work :wink:
-
If possible, I would very much like to have your help for "Quality Control" purposes - I have to rely on knowledgeable people if I want to keep Japanese writings in my scenes and have them make sense.
By the way, if you could double-check what I had already considered "final", that'd be much appreciated: in the indoor scenes of Aerith's house (a couple of pages earlier), I have textured many books and folders where I tried to write a title, I'd like to make sure it's consistent (and that I didn't write gibberish). I have also a number of scenes "in progress" where I'd like your review before I launch the final renders.
Sure thing:
料理書 should just be 料理, although like a book in English called "Cuisine" it feels like it needs a little more, so perhaps 基本料理 - Basic cooking, or 家庭料理 - Home cooking (or if you're feeling cheeky チョコボ料理!)
For the photo albums (unless they're books on photography), Japanese photo albums ironically tend to have gibberish English (http://www.amazon.co.jp/SEKISEI-%E3%83%8F%E3%83%BC%E3%83%91%E3%83%BC%E3%83%8F%E3%82%A6%E3%82%B9-%E3%83%95%E3%83%AC%E3%83%BC%E3%83%A0%E3%82%A2%E3%83%AB%E3%83%90%E3%83%A0-151~200%E6%9E%9A-XP-3257/dp/B00CCXXZ6Y/ref=sr_1_5?ie=UTF8&qid=1423693102&sr=8-5&keywords=%E5%86%99%E7%9C%9F%E3%82%A2%E3%83%AB%E3%83%90%E3%83%A0) (or often French) on them so these would look fine with an English title, like PHOTOS (or PHOTO for that authentic, grammatically incorrect feel).
作り方 is a little odd because it doesn't say what it's about. A gardening book might fit so perhaps 園芸, or 趣味の園芸 if you want to plagiarize a popular nhk series of magazines.
I can't quite make out the large green book and smaller blue book in the back, can you post the text?
I would also recommend to use a handwriting japanese font instead. I did a quick google search, you can download some here: http://www.freejapanesefont.com/category/handwriting/ (http://www.freejapanesefont.com/category/handwriting/)
Keep up the good work :wink:
Personally I'm not too sure about those, they're pen and pencil fonts, the cursive fonts I've seen so far are all brush fonts.
-
Thank you so much :D
I'll definitely take your advice for the books. Below is an image where I pasted all the writings of the textures I had made for the writings on the book covers. I only have the writings as images, not as fonts, but I'm sure you could help me out on that one. Sadly enough, I actually can't recall exactly what each title was intended to be (I have done this texture job almost two years ago!). The titles which are horizontally written are on the front of the cover, the ones written vertically are for the cover sides. On the image below, there may be some redundancy between titles because I've just quickly combined all the titles of all book together.
(http://i.imgur.com/kDwA0en.png)
What I can recall, however, is that I wanted some books about cooking indeed (I think something like "Home cooking" would be the best), photo albums (wouldn't have guessed about the Engrish on that one), gardening and floral arrangements. I even think that for one of them, I copy/pasted the title of the oldest book in Japan which happened to be about gardening
I had set 作り方 for some books as I was trying to make "how-to" books such as "1001 tips for your home" including a variety of odd jobs, the kind which would cover "DIY table", "applying wallpapers" or "fix your plumbing".
I'm very grateful for your help, although I can't help feeling a little down because I have to rework all these textures, and re-render the images :'(
-
Well to be honest this is all really nitpicky stuff so please don't feel like you need to go out of your way to fix things.
In any case:
ラーメン: Ramen, on the front of one of the cooking books. This is a bit strange because ramen isn't something you usually make at home in Japan, I mean it's certainly possible but uncommon (unless we're talking cup ramen but those don't have cookbooks :-P). おかず is better imho, these are your typical small home cooking dishes that accompany rice.
作り方: DIY is just DIY in Japan, although it's not nearly as common a practice as western countries to fix up your own stuff, perhaps another gardening book like I mentioned above, so 園芸?
ユリ: Lily, but if you're going for an "encyclopaedia" style books series then perhaps this should be ユリ科, which would be the lily family (I just find it a bit odd that there's a single huge volume on just the lily).
料理書: As above probably better as 家庭料理, home cooking
裁縫: Needlework, fine
生け花: Ikebana, fine
花の百科事典: Flower encyclopaedia, fine
-
Thanks! That's very insightful.
Well, all in all I think I can fix stuff without too much of a problem. I think that, if I have the means to be thorough and correct, I might as well do it right. Besides, your comments will be very useful if I wish to "recycle" some of these books on other scenes later.
And at least I'm happy to have gotten a couple of titles correctly :P
-
Personally I'm not too sure about those, they're pen and pencil fonts, the cursive fonts I've seen so far are all brush fonts.
Yeah I agree, in the original renders it is more of a brush type of font. I just wanted to note that most kanjis in the original game are handwritten. Here there are some free brush fonts (sorry, sites are in japanese):
http://free-font.biz/page1.html (http://free-font.biz/page1.html)
http://matome.naver.jp/odai/2136161327768901401 (http://matome.naver.jp/odai/2136161327768901401)
http://sitebk.com/summary/brush-font/ (http://sitebk.com/summary/brush-font/)
-
Now this is an interesting topic, could have used some of these insights while making the signs for midgar, those were a nightmare to decypher... Don't remember quite what I've put on them anymore. :-\
oh well, wonder if anyone will notice those tiny bits.
8)
-
Quick update:
mds5_i and ealin_12 were patched, following luksy's remarks, with corrections on the fonts used for the writings (Japanese instead of Chinese), as well as corrections on the book titles on ealin_12. New versions have been uploaded on the repo.
Otherwise, I've been working on the layering of mds5_w, which was another pain in the butt. I have all the alpha-masks ready for the layers. All I need is to finish a couple of textures (notes and posters stuck to the shelf and behind the seat on the left), make the final render and then it should be straightforward. Hopefully the remaining scenes (5min_1, 5min_2, and mds5_m) should be much less complicated to layer and cut out - I'm just missing a couple of textures in each of these scenes.
Since my adventures with the rendering and layering of mds5_1, my workflow has changed a little: I make a rough alpha-test render (with small amount of samples), which I use to prepare the layers. Depending on the results, the modelling may have to change (objects added, or moved), especially if we're in the case of "space not displayed by background layer, not covered by foreground objects". I go through the iterative process Model > Rough render > Cut and layer > test in game until the layering in the game fits the desired result. All that needs to be done afterwards is to make one final render (with lots of samples for high quality), then use the layer cuts from the rough renders as alpha masks on the final render to generate the layers in high quality quickly. Add post-processing effects (light glows, god rays, etc.) and there you have it.
-
Quick update:
All the alpha masks for the layers of 5min1_1 and 5min1_2 are done. I'm currently making the final render for 5min1_1, it should be finished Friday evening (UTC-7). A list of what remains to be done:
- Get the alpha masks for the layers of mds5_m.
- Finish 1 texture for 5min1_2, revise a couple of meshes, and make the final render for that scene.
- Finish 3 textures for mds5_w, make the final render for that scene.
- Finish 1 texture for mds5_m, add a couple of very simple objects, make the final render for that scene.
And that should wrap things up. Optional would be to patch mds5_1, because I had to change a bit the layout of mds5_w, which affects what can be viewed from the outside of the weapon shop.
-
Hey how do we get access to the repository to use and test these scenes? I've been trying to get registered on the forums for a while just now finally succeeded instead of getting an error. I'm using the set of sewer fields you released and it looks amazing in game!
-
The file repository is on restricted access, so you can't have these fields right now. However, once I finish all the scenes I've been working on (hopefully by the end of the month), I'll make a public release where these scenes will be just as available as the ones you already got. I also have plans to get that included in the next pack of the 7th Heaven mod.
Just a little patience ;)
-
Hi folks,
mds5_1 is finally done and uploaded to the repo. Below is the final render with light layers:
Again, anyone who wishes to try it without having access to the repo, please PM me. I hope you like it.
Ahh, very well. I only requested because of your post there.
-
Oh, well actually you're the first to make a personal request. If you wish to test a scene as preview, I can email you the corresponding files. Just for a quick reminder, at the moment I have the ground floor and top floor of Aerith's house (ealin_1/ealin_12, and ealin_2, respectively), the outdoor of Aerith's house (eals_1), the main sector 5 market scene (mds5_1), the item shop (mds5_i) and the "main in pipe" scene (mds5_dk). Just let me know the ones you'd like to try.
That being said, I have the whole package coming up soon enough (about a week), which will have 5min1_1, 5min1_2, mds5_w, and mds5_m in addition - I'd also like to make a small patch on mds5_1 for that release.
-
Quick update:
- mds5_w is complete, ready to be in game.
- mds5_1 has been patched to reflect the changes made to mds5_w. It's ready to be in game.
- 5min1_1 and 5min1_2 are almost done. Final renders are done, I only need to add post-processing effects (ie. glows and volumetric lighting effects) for both scenes.
- mds5_m is currently in the final render. I still need to cut the layers for that scene (although it shouldn't be too complicated) and add post-processing effects.
None of this is on the repo yet.
So... the end is within sight. Any recommendation about where to upload the files for the public release?
-
Technically they're images right? So you could use something as simple and reliable as imgur. I think your best bet would be zip files on Google drive or Dropbox, however.
I'll download them early if you'd like but I don't mind waiting. Didn't realize you were so close to release.
-
If you don't mind waiting, then it's more convenient for me that you'd hold on until the release :)
As for hosting, I was planning of putting all directories in a 7z archive, and upload that somewhere everyone can access to. For the user, all that would be needed would be to extract the archive into their modpath directory. Imgur wouldn't be convenient because the user would need to download each and sort them out.
-
Imgur would be inconvenient because each background is made of several separate files
I have a soft spot for Google drive, I don't believe they cap the dl speed and works in most countries
-
Google drive tends to work pretty well. Otherwise the others I use are Mediafire and Mega. Mega works best in Asia and that whole region, but can be slower in the UK and the US.
-
You can release from the repo. Just need to change the share settings. I can do it when you have it up.
-
Thanks for the feedback, guys :)
The easiest will be to upload a .7z to the repo and then change the share settings indeed. I'll do that. I'm aiming for tomorrow evening.
-
Awesome! Actually putting off playing until Reunion R03b and these scenes are released. Really looking forward to it.
-
Hi folks,
It's done. I've uploaded the .7z to the repo. For the latest version, check out the
first message of the Released Gallery thread (http://forums.qhimm.com/index.php?topic=15975.msg224827#msg224827)
So right now I think a Victory Fanfare is in order :P
That pretty much concludes 15 months of work. I'm beat, time to hit the sack.
-
Wooooo I cannot wait to try this :evil:
Absolutely amazing work, now I have to try and remember where I put ff7 and how to get these puppies in the game
-
Great work man, keep it up! Your renders are very faithful to the originals, yet add more details and never fail to impress. :)
-
Good work :) What's next on the plan? You'll have done all of Midgar before long!
-
Good work :) What's next on the plan? You'll have done all of Midgar before long!
I wish! Sadly the reality is very different, considering it took me 15 months to complete 11 scenes. We would need about 10 more people with the same work rate for this project to go somewhere.
I'm not really decided about what I am going to do next, to be honest. I hope my works can motivate people to pitch in, because I can't make this work on my own.
That being said, I have a couple of ideas regarding what to do next, but they would be outside Midgar. Most likely, resuming my works on the Tiny Bronco, or starting some scenes in the Nibel mountains.