The Canon C300 is quite the camera, and by all accounts it's a high-end professional camera (and we should refer to it as such since it's the most expensive camera Canon makes). But something strange is going on that could affect your footage in a very real and disastrous way as compared to other cameras. Paul Antico at NextWaveDV has discovered a very disturbing image artifact that appears in purple and green blocks on overexposed edges. He's not the only one, as others have replicated this exact same problem.

Here is the most telling example of the problem, with a comparison to the RED Scarlet, a camera which does not downscale footage in camera.


Scarlet-vs-c300_fringing

Scarlet-vs-c300_fringing-e1333063296336-616x395

Over at NextWaveDV they've got a couple more examples and people who've had issues with the same problem in their footage. Now this is troubling because of Canon's response, which seems to indicate that they've never seen this before in any of their testing:

The cyan/purple fringing seems to be more notable when the shooter is deliberately trying to blow out the highlights....Still, it is something that I will pass along to our senior engineers. We will investigate the situation, and see if there is anything else that can be done to reduce the fringing.

So what's really going on here - and is it something you should be worried about as a C300 shooter? It's definitely not chromatic aberration, because it's been shown to happen regardless of the lens (and it's very clearly solid blocks instead of a smooth green or purple gradient on highlight edges). It seems to only be appearing on highly overexposed areas of the image - and not anywhere else. This leads me to believe that groups of pixels start acting funny when they are over-saturated, and they essentially get locked in the "on" position - not too dissimilar to a stuck pixel that you'd find on an LCD monitor or TV (there is sometimes confusion between stuck and dead pixels - but dead pixels are always off, and will therefore appear black or white). I'm not an engineer so I can't be sure, but that seems like what is happening here, so the problem is all the way at the pixel level.

This probably isn't something Canon can fix very easily at the software (or firmware) level, this looks to be a permanent design flaw relating to the pixel response. Now, as I asked before, should you be worried? I think the answer really depends on what you're shooting. If you are shooting a feature film and your goal is to keep the image flat and not overexpose highlights, then you probably have nothing to worry about. If, on the other hand, you are shooting a documentary or footage where you need the image to be very similar to the final output, you've got to be very careful about your highlights. Consider underexposing slightly if a lot of your image has bright, saturated highlights. It will really depend on the situation whether someone will notice it or not - because obviously for the most part it seems like people have not noticed this problem (or we would have heard about it sooner). It's certainly not confined to one camera, as others have experienced the exact same issues.

Is it possible that only certain sensors have this problem? Maybe - but the way that sensors are fabricated, it's likely that all Canon C300 cameras will suffer from this problem at one point or another. Does this make the C300 a worthless camera? Absolutely not, and the fact that the issue has only been noticed now shows that it does its job correctly most of the time. Whether this turns out to be a hardware problem or not, it's a little troubling that not only does Canon not know anything about this issue, but that a $16,000 camera experiences any major issues like this at all. All cameras have flaws, as we know, but this seems inexcusable on a camera this expensive that's made for professionals.

I'm sure we'll get updates on this in the near future, and I'll stay on top of the issue to see if Canon updates their response in any way.

[via NextWaveDV]