Regular glitch or ghost in the camera? DIAL UP WARNING - ABOUT 650K!

[Jari]

From Jari's archive of weird things - the very same that brought you the "Sky King, Sky King..."-nukular no-go message few years ago.

I'm sure that most of you have seen a photo or two that demonstrate some sort of purple artifacts on them. Often in the shape of vertical streaks, but I've also seen interlace-pattern type of artifact and some cases where there's no weird pattern at all - it's just that the highlights have taken on a purplish hue.

As far as I know, this is due to the sensor being overdriven, thanks to very bright source of UV, IR or visible light. ...I'm sure that someone can correct me on this, if I'm wrong... and specify what kind of emission it is, that the camera is sensitive to.

Anyway, I have something far cooler to show you. A camera with a built-in time machine. Or a variation of this artifact that I've never seen before.

Some background info first: following five frames are taken from a video which shows a 95mm recoilless rifle (I used to shoot one of those myself ) firing. The 640x480 resolution is very weird for a video camera, plus IIRC the video had a frame rate of either 15 or 30 fps, whereas real PAL video should have 25 - this leads me to believe that it's captured with a digital still camera in video mode, instead of real (digital or not) video camera. Of course, all this could be due to compression; the video is in .wmv-format.

Sadly I can't contact the author of the video anymore, so I could ask him about these things. And I've also archived the video to some DVD and can't find it right now, I took these screen grabs from it earlier.

I swear that these frames are consecutive and in order. I have done nothing at all to them.

Observe.



Frame #1 You can see the weapon itself, crew, some shell casings in the foreground and the supervising officer (dude in the red safety vest and helmet). Minor vertical purple streaking is visible - it looks like this is from the sun.




Frame #2 Nothing has happened - except some branches and leaves are very slightly moving (in the wind, I believe). Yet there is massive artifact in the frame. Where does it come from?




Frame #3 Weapon fires, you can see the backblast venting from it, and rather impressive puff of smoke on the front side, too. To give you some idea about how impressive it is in real life; the sounds is 208dB loud, equaling heavy artillery piece. The crew is wearing double hearing protectors, and the pressure shock has been known to knock pictures of walls a mile away from the shooting range.

There's about 5 pounds of propellant in each round, and the projectile weights more than that, IIRC. Muzzle velocity is a tiny bit over 700m/s, which should give you some impression about the power.




Frame #4 Observe the leafs and branches. See the pressure shock hitting them?




Frame #5 And due to human reaction time, the cameraman reacts now, and jerks the camera little bit.

That's one shot. The video has several, and each of them exhibit the same pattern; a massive artifact due to what looks like a sensor being overdriven in the frame previous to each shot. And even more uncannily - at the end of the video there are two shots where the camera is pointed at the target, with the RR behind it (and thus nowhere near the frame)... and guess what. The artifact is still there. The projectiles have a tiny tracer embedded into their ass, so that you can see where you shoot, but the artifact shows up before the projectile is in the frame.


So... why does the artifact show up before the shot?

Regardless of what kind of emission it is, I'm pretty certain that it can't beat the speed of light. So, it should hit the sensor at the same time with the regular light.

The sensor itself... can it react faster to some forms of emission than to others? This doesn't sound very plausible to me, since digital camera sensors can capture visible light with exposures as short as 1/8000 s.

Maybe it's a feature of the compression codec? Some kind of video equivalent for the pre-echo sometimes heard in recordings made with the early versions of aac and vqf-codecs? Note to people who are not familiar with this; pre-echo is exactly what it sounds like; an artifact you hear before the real sound - a temporal distortion caused by the codec.


Any thoughts? Ever seen anything like that? Perhaps it's just a glitch in the Matrix?

[ChaosControl]

Heh, at least you're not thinking its ghosts..
They tend to do that a lot here nowadays..

"Ohmigawd, a ghost in my camera!"

I think it's just a technical problem, maybe the sensor sucks? or is broken? could be anything really.

[Jari]

Yes, some sensors are more sensitive to such purple artifacts. In fact, I've heard that they are even mentioned in the manuals of some cameras.

But the question remains; how can the camera predict the RR firing, with unerring accuracy? Even when the weapon is nowhere to be seen. I think that there are five or six shots on that video, and every time the artifact shows up one frame before the shot (well... can't say for sure about the ones that have the gun out of sight, of course - but the artifact does show up in them, before the projectile comes into sight).

[ChaosControl]

Uhm.. perhaps the light of the shot sends some ultraviolet beams across which react on the sensor for some reason and it would be being a bit faster then the other light which could be pulled back a little because of the blow and thats why?

Really, really far fetched, i know.

[L. Spiro]

It always appears in front of the muzzle right?

The artifact is an inverse impression of the trees immediately in front of the muzzle—if you flip the glow up-side down it would snuggly fit behind the trees just above where it is now.

Not being an expert in artilary at all, I have only guesses at the moment, but I would like to know how the haze looks when the muzzle is not on the screen (and where the muzzle is in relation to the screen.

So far there is physical evidence linked to the muzzle position.
In the shots where the muzzle is not on screen, I expect that the muzzle is just a bit behind the screen or is still in some position directly related to the flash.


The idea has crossed my mind that the gun is fired in frame 2, followed by smoke and fire, which isn’t captured on film until frame 3 giving the illusion it was actually fired between frames 2 and 3.
I the haze appears at varying intensities and distances in front of the muzzle, the case is probably something along these lines.


L. Spiro

[Jari]

Uhm.. perhaps the light of the shot sends some ultraviolet beams across which react on the sensor for some reason and it would be being a bit faster then the other light which could be pulled back a little because of the blow and thats why?

Well... there is the problem of regular light traveling 299 792 458 m/s (that's what Google told me), which means that it travels 9 993 081.93 meters during one frame of that video, assuming that it's 30fps, twice that if it's 15fps.

The RR is about... I guess less than 20 meters from the camera - assuming that the camera is not zoomed to very long focal length (big zoom factor, in other words), and it probably is not, since it's fairly stable while being obviously held in a hand.

So, that would mean that the ultraviolet beams would have to travel lot faster than visible light, for that to happen.

It always appears in front of the muzzle right?

Behind the rear of the gun, actually. You are looking at it from the rear.

But yes, as far as I can remember it appears roughly in the same place in every shot.

The artifact is an inverse impression of the trees immediately in front of the muzzle—if you flip the glow up-side down it would snuggly fit behind the trees just above where it is now.

Yeah, I noticed. Well, some of the trees - it looks like some of the group got omitted. But roughly, yes.

Not being an expert in artilary at all, I have only guesses at the moment, but I would like to know how the haze looks when the muzzle is not on the screen (and where the muzzle is in relation to the screen.

I could try look for the video, I have it somewhere. I can grab those screens, if I find it. Sadly, it's not possible to know exactly were the gun is in those shots - except that it's behind the camera somewhere. It certainly is not directly behind it, since they wouldn't let people wander in front of it. I suppose that one could try to locate it from extrapolating the trajectory of the rounds, seeing that they are visible in multiple frames, but there isn't even a way of knowing how far they are from the gun. My guess would be that they are considerably farther away than in this shot - possibly have a small hill between them and the gun, to prevent stray shot from hitting them.

I haven't used that range myself, so I don't know where they like to "seat" the spectators.

The idea has crossed my mind that the gun is fired in frame 2, followed by smoke and fire, which isn’t captured on film until frame 3 giving the illusion it was actually fired between frames 2 and 3.
I the haze appears at varying intensities and distances in front of the muzzle, the case is probably something along these lines.

So, what is the purple artifact, then?

First thing that happens when such RR is fired is that the primer is ignited - before that it's entirely mechanical process. The primer itself should not give off light, at least visible light. After all, it is embedded in the rear of the shell, except for the little part that is actually hit to ignite it.

[L. Spiro]

So I can’t tell the front of the gun from the rear; I told you I’m definitely no artillery expert. 

Some of the trees are omitted in the artifact but the smoke in frame 3 omits the same trees, so it’s definitely something taking physical space at the same spot where the smoke would later follow.

But then I don’t know how well that holds true in the shots where the read of the gun is off-screen.


I would normally suggest some type of emition during the mechanical process of firing the weapon, but the M/58 (correct?) isn’t exactly a new weapon capable of emitting odd kinds of radiation during firing.  If it is from the Cold War era, it’s probably entirely basic mechanics, so I-

Actually I do have an idea now.
How does a recoilless rifle work?  To damp the backwards force of the bullet (which happens to be huge in this case) it probably has to send something forward with the same momentum as to counter-act the discharge of the bullet.
It probably is basic mechanics, but there is definitely going to be something happening in the rear of the gun, possibly with pre-fire discharges of its own.

With no special knowledge on the subject I couldn’t get into detail what it is, but it’s probably related to the artifact.
I imagine some kind of pre-fire discharge that moves a large block forward in the rear so that when the bullet is fired less than a frame later the machine is resting in equilibrium.


L. Spiro

[Jari]

I would normally suggest some type of emition during the mechanical process of firing the weapon, but the M/58 (correct?) isn’t exactly a new weapon capable of emitting odd kinds of radiation during firing.  If it is from the Cold War era, it’s probably entirely basic mechanics, so I-

Very good.

M58/61, technically. I believe that wheels were added to it during '61. You can shoot it from your shoulder (without the wheels), like most recoilless rifles, but the sucker weights like 200 pounds even without the wheels and is over 10 feet long... so, it's not very handy.

Yeah, there shouldn't be anything radiating there. Only thing I remember having a radiation warning was the tritium used on the assault rifle night sights. I think that they advised us not to eat the sights.

Actually I do have an idea now.
How does a recoilless rifle work?  To damp the backwards force of the bullet (which happens to be huge in this case) it probably has to send something forward with the same momentum as to counter-act the discharge of the bullet.
It probably is basic mechanics, but there is definitely going to be something happening in the rear of the gun, possibly with pre-fire discharges of its own.

That would be the big, bright backblast you see in the pictures. The breech of the weapon is basically open, and the rear of the weapon forms a Venturi-chamber. So, most of the propellant is actually "wasted" exploding backwards - to keep the weapon itself still. In exchange for wasting powder you get lot lighter weapon, regular anti-tank gun of the same caliber would weight at least five or six times as much.

Needless to say, the backblast is lethal at close range and extremely unpleasant bit farther away.

So, it shouldn't be pre-fire discharge, as it uses the very same explosion.

This is the method used by this particular recoilless weapon, and I believe most (all?) still in service today. It is not the only possible method, though. In fact, IIRC there are designs that indeed use some kind of counter mass and don't waste propellant, but I don't remember the details. ...and I've never seen or heard about anyone using it in production weapon.


Anyway, I found the video.

First, I have to apologize for my memory; the artifact does not appear when the gun fires, but rather when the projectile's HEAT-warhead goes off. Which happens on screen.

Still, there is the same phenomena; first the artifact shows up - while the same frame still shows the projectile in mid-flight, and the actual explosion is on the next frame.

Also, the artifact takes a different shape now - it clearly looks like it's originating from the center of the explosion. The shape of it is no surprise, I have a photo of exploding tank (from either of the Gulf Wars) that shows similar artifact - except that it appears during the explosion.

Bit of technical information; the video is 15fps, so there's quite a bit of time between frames.

There are three shots from this perspective, and all are similar.

Here's one:




Frame #1 Look closely little bit to left from the dark shape in the middle. That's the "bright" tracer carried by the round.




Frame #2 As you can see, the round is close to it's target now, easily visible over it. It still seems to be in mid-flight, because we can see the tracer, but no explosion. Yet the artifact has appeared.




Frame #3 Warhead has gone off, resulting in a nice little explosion.




Frame #4 ...and some smoke in the this frame.

[L. Spiro]

I have one last theory but it’s a bit hard to explain.

It seems to be up to the camera.
If the actual recording process was done at 15 frames per second, the camera would have to have a higher exposure time.
During the exposure of one frame it may be possible that a particularly bright area would over-expose and bleed onto one frame before the next is done exposing.
For this to make sense, I have to imagine a digital camera that temporarily stores the current exposure behind the scenes while the next frame is exposing, allowing the next frame to have an effect over the previous while it is still in memory.

From the photoreceptor to the digital image, anything can happen, including vertical inversion.

But just a guess.

I’m also not a camera expert. 


L. Spiro

[ChaosControl]

well, it isnt that far fetched.
People who photograph lighning do this on purpose to capture the flash.

[Jari]

Spiro: Your guess is as good as mine. I don't think that the cameras are supposed to do that, but it's not like I have an explanation, either.

People who photograph lighning do this on purpose to capture the flash.

Actually, I think that they just shoot with a really small aperture and super long exposures - usually using the bulb mode. Basically same thing you do to shoot fireworks. What Spiro is talking about is something different - sort of a (unintentional) double exposure, really. Something that could be done as a special effect with film cameras (at least with some trickery, or stuck film ), but I'm not sure if there are any digital cameras that let you do one.