Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - insidious611

Pages: 1 2 3 [4] 5 6
76
https://www.youtube.com/watch?v=PvePGgdm_sc An *extremely* short demonstration of working resume support.

77
GOOD NEWS. With Covarr's help, I now understand the changes to manifests/file naming conventions/etc that higan v096/v097 brought. As such, I was able to get everything working.

Resume support is (preliminarily, I haven't done full testing) working in higan v097.


Now for the bad news:

I have to rewrite a significant section of the Installer to work with the changed file structure of the new higan. So, no test version with resume support yet, or for at least a few more days.

78
So I'm kind of stuck... I can't get the current patch to run *at all* in Higan v097, which is the one with resume support. It just seems to freeze near the beginning. Higan does *not* currently have a debugger so I can't figure out why this is happening and bsnes-plus v073+2, which does have a debugger, works fine, but it reports itself as MSU-1 version 1 which doesn't have resume support.

Resume support should be working but I am currently *unable to test this at all* and indeed, unable to do *any further testing or development* until I can figure out what's going on.

NOTE: This *includes* current alpha code prior to my working on resume at all, ie, code that *should* work because it did (and *does*) work on higan v094

79
Took a look at it. I'm intrigued, I might have to poke at this header some at some point and figure out if we can't find where the loop point value is and how it's calculated.

Unfortunately I'd have to basically back-convert our .pcms to OGGs, since they underwent heavy editing in some cases, and besides that the actual loop point is contained in the header of those, as Covarr and I didn't write the numbers down :P

For an update to the project, all's quiet on the bugtesting front, with only 5 bugs remaining, all of which were submitted months ago and 3 of which were known before I put in alpha-testing and have been held off until beta. I had someone apply to be an alpha tester but at this point it's a tad too late I'm afraid. I'll be poking for a new round of testers once I get ready to push into beta, which will be after I get Dancing Mad subsong switching and battle resume support working.

On a personal note, active work on the project on my part has been on an unofficial hiatus the last few months due to personal issues, in addition to preparations for an upcoming move from my current location in northern Illinois to southern Virginia. I felt the need to be honest with you guys on that front. But I promise that work will resume in March once I'm moved and settled in. Thanks your understanding.

I'd like to take an unofficial poll, actually, concerning resume support. Now that Higan has been updated to have official resume support, and presumably similar updates are in works to the SD2SNES's firmware (I wouldn't know as I don't have one and don't really check these things), should I actually continue with the hacky "faux resume" support I was working on, where the MSU-1 tracks would be muted temporarily to bring up the SPC battle/victory themes, and then brought back to full volume, or should I scrap that entirely in favor of cleaner code that focuses only on true resume support, falling back to current "no resume at all" functionality if it's absent?

The reason I ask is that the faux-resume support actually involves a lot of somewhat convoluted logic to implement, since I have to kind of bolt it onto the current code which is not designed to mute the MSU-1 without stopping it or switching tracks. It's not a lot of actual lines of code (the ASM patch in general is actually quite short) but it's a lot of stuff that makes the code harder to read and maintain, even when well commented.

True resume support on the other hand is the matter of setting a flag and then otherwise using the exact same logic I'm already using.

I'd also like to point out, not that I foresee this ever happening, but should something happen to me where I can no longer work on this project, I will release my source code in its current form to the public for someone to pick up on my work. Again, this is highly unlikely to happen, I'm enjoying working on this project and I fully intend to see it through to completion, but I wanted to make that clear.

80
I wasn't at all aware this was a possibility. I'll look into it, but tentatively at least I'm interested in trying. AFTER I've got this version done, anyway.

81
If you mean working on a hack yourself, I highly encourage trying it. The basic coding and hooking is surprisingly simple and yet the knowledge of ASM and of a game's inner workings you gain from it is very rewarding.

Honestly the hardest parts of this have been logistics and testing. Testing is kind of a bitch, you have to be very very patient with a debugger. (I suggest a recent version of bsnes-plus)

82
WIP / Re: [In Development] FF8 Voiceover Project
« on: 2015-12-21 04:09:44 »
Have considered auditioning for Irvine but I've got next to no voice acting experience.

83
Just wanted to say this is still in progress. I haven't gotten any bug reports in a while (which I choose to see as a good sign :P), I've been working on the battle theme support (both a prototype of the resume support and the fake-pause-and-play-the-SPC support) but I haven't gotten the chance to test it as yet.

84
We haven't come up against any show stoppers so far since then, knock on wood. I'm starting to work on some of the "safer" beta-target features, such as handling the last 2/5th of dancing mad, and the battle/victory theme handling, with muting (but not stopping) the MSU and playing the SPC music, then turning it back up to full.

To make use of the battle theme, since we're downloading it anyway (especially since eventually we will be using the new MSU spec's pause/resume support), for scenes that use the SPC music anyway like the wind-effects stuff at the beginning of the game, we'll trigger the MSU battle music. We'll only use the SPC battle music if the MSU music is already playing for the area. This is going to be a little inconsistent but imo it's the best way to use the resources available and show off what we can do.

85
So we've fixed the sound glitches on SD2SNES but have come up against some hard to explain graphical glitches


EDIT: ... Which have just been explained away as being glitches in the latest preview SD2SNES firmware. Whee.

86
Good news on that front. We have someone with considerably more experience than myself with development for the SD2SNES doing some preliminary testing and ready to give tips, so *hopefully* he can figure something out.

87
Glad you enjoyed it! The video is using my recommendation of music, which is a mix of sources.

Unfortunately we currently have a bit of a head scratcher bug with SD2SNES, so a public release may be a while down the road :/

88
Probably in BitBucket. I just gave you write access, lets see if that helps.

EDIT: Oops, I also hadn't turned the issue tracker *on* :D. Fixed that too.

89
WIP / Re: [In Development] FF8 Voiceover Project
« on: 2015-09-29 02:07:00 »
From one audio modder to another, keep it up! I'm rooting for you and for this awesome project. This will definitely give me a reason to replay FF8 when it's done.

90
The below is a stream I did earlier. It's a full demonstration of Dancing Mad in its current state, from the intro through to just past when Terra wakes up and remembers her name. I tried to keep the talking to a minimum so that you can hear the music in all its glory, but I did stop to address some questions and explain and point out some of the current known bugs.

https://www.youtube.com/watch?v=4WHT98HBgpE

91
Alright, the installer seems to be complete enough for an Alpha test. I'll be doing a bit of a stream (which will then be uploaded to youtube) of the mod in its current form (complete with bugs) today, going up to the scene where Terra wakes up in the Old Man's house. Following this I'll start sending the installer out to selected testers. For now, this is Covarr and, provided he's still willing to test (I'll be PMing him first so I can be sure) RetroDan, as well as possibly another person who has PMed me. Anyone else willing to test is encouraged to apply, though I require people with emulation experience, romhack experience, and experience with FF6. You require a Final Fantasy III US ROM, version 1.0, without a header. The installer should detect whether or not you have the right ROM. Optional desires of mine are testers with experience with MSU-1 hacks, experience with a SNES debugger, and/or who own an SD2SNES.


I'm going to put this in big bold text for the people who like to skip over the technical stuff.

The Private Alpha is complete and we are now accepting testers. Please read above for requirements.

Please PM or email me to apply.

My email is

If you are selected to test I am trusting that you will not leak my alpha-grade IPS patch or patched ROM to anyone. Please do not betray this trust.

Source code is available on my private BitBucket which will be opened to testers to view. When we reach public beta and beyond, this BitBucket will be made public and also duplicated on GitHub.

92
So at this point I've finished coding and testing the installer (after a 13 hour coding marathon). It is working 100%. Some polishing may be necessary, I'll check it again in the morning when I'm less tired.

I also need to turn it from its current state as a collection of .py files that depends on having Python3.4, PyQt5, and PyCurl installed into an EXE file that can be ran like a normal installer and doesn't depend on any of these being installed (instead bundling in compressed versions of the python interpreter and libraries), but this shouldn't be too difficult.

93
Good news!

http://board.byuu.org/phpbb3/viewtopic.php?f=16&t=700

The MSU-1 spec is going to be updated with Resume support! Which means I can potentially replace the battle/victory music and not interrupt the area music! I am not going to be implementing the version in bsnes-plus, I'm going to wait for a stable implementation in higan v096, so this is not in the cards for this project in the immediate future, and in the immediate future (after the alpha which should be coming out soonish) I'll be implementing the version that uses the SPC for battle/victory.

However, this means that yes, saftle, if things with the emulator developers go as they've said they should, we will eventually have a solution that doesn't need the workaround!

94
There's a few more things to do on the installer, mostly window dressing at this point, but it is now working.

I'm going to finish up the installer and do some basic testing on my end over the next couple days, and once I'm done I'm going to get this out to our Alpha testers.

Silly installer teaser:


95
Just wanted to pop in and confirm that leaving the battle/victory music using the SPC so that area songs can play all the way through is the direction we're going. The alpha I'll be giving out to select people soon may not have this feature (I'm planning on getting everything wrapped up in a nice package that can be easily tested so that I can make sure there's no bugs in the current implementation before I shake things up)

Progress is slow as I'm currently dealing with some issues in my life that make working on these things difficult. I'm currently on a two week vacation spending time with family away from these issues but it goes without saying I'm not working on my vacation :P. But I also wanted to assure people this is still a priority for me.                                           

96
With a couple minor exceptions, all audio work is now finished. I need to get the installer functioning 100% and then we'll see about getting it out to testers for feedback.

I have a bit of a problem though, and I'm going to be asking for a bit of feedback on this

The MSU-1 does not support seeking or resuming tracks. Therefore, if I stop one track and start another, it always starts at the beginning. This isn't an issue *except* during battle scenes, where you would maybe hear the first few strains of an area's melody, then the battle music because you got into a battle, then the victory fanfare, and then the area's melody would start over again. This leads to a large portion of different songs not really being hearable due to constantly being restarted. There's two basic solutions to this issue.

The first is basically to ignore the issue, have people just kind of deal with the restarting area music, maybe use a second track type to skip intros so it isn't as obvious.

The second is use the SPC (original style) music for battle and victory themes, merely fading out the MSU-1 area music before battle and then fading it back in afterwards. This isn't perfect either as it means you'll be "missing" a bit of the area melody when it fades back in (since it won't have stopped), but it's much better imo than restarting it every time.

The problem with the second idea off course being that that means the normal battle theme and victory fanfare will not be being replaced. The boss themes will be replaced as normal because they don't happen often enough for this to be a problem, but yeah.

My question is, which do you guys favor? Replacement of all tracks including battle/victory but area music restarting after battles, or leave the battle/victory music as is but have smoother music otherwise.

97
Thanks for the event commands table! I've actually been hooking around the routine where the song selection and volume are pushed from RAM into the SPC and haven't touched those yet *but* they're probably going to be useful in investigating the current problem-tracks of Dancing Mad and The Phantom Train, which appear to be using maybe the F7 command to transition to various different "subsongs" within one actual in-memory song. We can't actually implement that in MSU-1, *but* we can trap those calls and implement them as transitions to different tracks, enabling the proper handling of those songs.

I haven't touched the NMI stuff too much since mentioning it because I want to get people testing it without the fading first, to make sure we have no major bugs in the current implementation and provide kind of a baseline so I know that if I broke something it's due to the NMI stuff.

98
Progress continues slowly. I've been focusing on some other things lately, but the installer is properly downloading tracks from my servers. Once I've finished that, finalized the last of the music (including deciding upon my selection), and run a test on the current patch code to make sure it hasn't been overtaken by gremlins or something (that is to say, to make sure it's still functioning the way it was last time I tested it), I'll be sending out alpha copies.

As far as when this will happen, Douglas Adams said it best when he said  “I love deadlines. I love the whooshing noise they make as they go by.”. That said, I hope to have an Alpha patch out to testers by the end of August.

99
It'd be a rather grand undertaking in and of itself to redo the actual music for the opera. As Covarr said, we already have the redone music from the OCRemix album, the problem is entirely a programming one. Getting the time cues in the game to sync, not just the mouth flaps but the dialogue timings and the proper "you made the right choice, you made the wrong choice" stuff all working right is going to be a pain in the ass, as this is all stuff that the "new track" handler that we're overwriting in the game's code isn't used to handle, instead there's very specialized code in the game for handling that sequence musically.

This is the bit that's a pain, because we have to have our own very specialized code for this sequence. Our code for most other sequences is very simple and thus very easy to maintain. Dancing Mad and The Phantom Train are also going to need similarly specialized handling (though not as detailed) for only somewhat similar reasons.

To be honest, it's not quite accurate to say we're "still having trouble" with the Opera, though. Code-wise, I haven't even started with the opera. The way our code works, I can choose not to handle certain tracks and the game's native music code will handle it, so I'm having it to do that for the Opera for now, and haven't even begun the runs through the debugger required to set the stage for coding the opera stuff.


In other news, I've completed preliminary work on designing the installer, which is a vital step towards being able to get copies out to our alpha-testers.

100
Putting in a vote for 4.

Pages: 1 2 3 [4] 5 6