There seems to be a constant battle between client, production, and the people who actually have to sit behind the computer and do the post-production. This battle comes in the form of a huge dissonance between expectations and budget. I don’t care how good the VFX artist is, crap footage is crap footage. I’m not talking artistically; I’m talking technically. This specific distinction is probably the crux of the issue. The only person who understands the technical ins and outs is the guy who sits behind the computer. Sure, the other groups understand how something like a green screen works, they understand that a computer can track markers and make CG fit into the shot, basic fundamental visual effects knowledge. Then you go a level deeper, do they understand how H.264 affects your footage, do they understand where and how many tracking markers you need? I chuckle to myself to think how few of these “professionals” actually do know. I could go deeper asking if they even know WHY we use something like green screen.

It is definitely not their job to know. It is their job to get things done under budget, on time, and of high quality. Yes, in that order of preference. I find that a lot of the time people think that budget and quality is unrelated. You might disagree with that sentence because budget and quality is obviously linked. It makes me wonder where this disconnection happens. Here’s a typical example that happens very often. There are these wonderful cameras out there (i.e. the 5D and 7D) that shoot beautifully in HD… to H.264. There’s nothing wrong with that, nothing at all, until you try to manipulate the image. Crank your colours too much and you’ll start to see the artefacts. This is even more obvious when you try to use the poor thing for green screen. A distinct halo will form around your subject (Where the stark green and the rest of the colours meet). In extreme cases you’ll even see the blocks of compression! So then why is this even used? If the problems are so obvious then surely we’d be using something else? Using a 5D to shoot vs. using a RED or an Alexa is cheaper, oodles cheaper. So much cheaper, in fact, producers will completely overlook the lack of quality. Most of the time you’ll shoot some footage, grade it lightly, and be done with it. This generally works out really well and looks very good. This in repetition will reinforce the fact that the H.264 cameras are just as good as the high-end cameras.

This is only a small part of the problem. The big part comes in where expectations are unrealistic because of this skewed view on what “works”. Once a shoot is done and in post only then do these problems crop up. You’ll have your producer and director look at the green screen or the grade and to their genuine surprise, it looks nothing like <insert Hollywood blockbuster here>. I sit there and laugh at questions like, “Where is all this noise coming from?”, “Why is the motion blur so blocky on the green screen?”, “How come the blacks are flickering?” Usually I’ll look up from my desk and ask, “You shot on a 5D didn’t you?” and it doesn’t seem to click.

If your device outputs to H.264 then your green screen WILL be shit.
I know some people might start bringing up Andrew Kramer and say that his green screen shots have been amazing and he shot on a 5D. There’s a simple answer to that: Andrew Kramer is the guy who sits behind the computer, as well as the director, as well as the producer, he knows the limitations of H.264 and works around them.

All I’m saying is there needs to be some sort of shift in thinking, imaginations need to be reeled in a bit, and maybe a bit more research done before choices are made resulting in highly unrealistic expectations. If you want your $10,000 production to look like a $1,000,000 then you need to put in $990,000 worth of preproduction, planning and research. Shooting it guerilla style will result in exactly that, something that looks like it was shot by a gorilla.

Here’s a very nice example of the different looks you can get out of different cameras, they’ve all been graded to match. Pay very special attention to the crispness of the hair, deepness of the black, noise, and artefacts. I’m still trying to find the source file, all I have to watch is… you guessed it… the HD H.264 download from Vimeo…

I thought I’d be nice and put down the rates for the different cameras as a comparison but I had a look at some rate cards and they don’t even mention the formats that these cameras output which just made me rage even more. I did however find the related article that I’ll give a read in the near future.

Cameras used in order of appearance (all shot at 24fps):
• Arri Alexa || HD 1920×1080 (Quicktime ProRes 444 LogC)
• Red Mx || 4K 16:9 3840×2160 (4K REDRAW, redgamma, redcolor)
• Canon EOS 7D || HD 1920×1080 (Canon H.264 mpeg-4, Technicolor Cinestyle)
• Sony F65 || 4K 4096×2160 (4K Sony RAW, S-Gamut, Slog2)
• Sony FS700 || HD 1920×1080 (NXCAM AVCHD mpeg-4)
• Canon C300 PL || HD 1920×1080 (Canon XF Codec 4:2:2 mpeg-2)

</endRant>