Newb/ Color Correction and Bit Depth

denzlite

denzlite

Active member
When you launch a new AE project, it defaults to 8 bit color. I've been thinking about learning a color correction workflow using AE and Prem and someone over in a my camera's forum (T2i) suggested importing a Prem project and working in 16 bit color.

I'm having a hard time wrapping my mind around bit depth, can someone point me to some tutorials or explain a little? Or maybe a link to somewhere else in this forum in general talking about utilizing bit depth in After Effects for good Color Correction.
 
Higher precision with increased color depth results in less banding /posteriztion added when you do adjustments. It doesn't "fix" artifacts that have already been baked in. Basically each color channel is represented by more "notches" at higher color depth. I explained it in more detail here
http://www.dvxuser.com/V6/showthread.php?t=203001

With your T2i, you won't gain very many benefits, unless you get rid of the .mov container. It's very limiting (at least on PC platform). Of course working at 16 or 32bpc is a lot slower because it's more CPU intensive

Basic color correction concepts in AE (but they apply to video in general)
http://www.peachpit.com/articles/article.aspx?p=1276351
 
Hmm interesting. So you must convert the T2i footage 'out' of it's .mov wrapper before working in 16bit color depth?

I've got AE and Prem CS5 and working on the native footage is something I enjoy immensely. I'm just not sure right now that the benefits warrant the amount of time and HD space it would take to convert all of my footage, but I'll do some tests, thanks!
 
You don't want to work on the native H264 T2i video, you should immediately convert it to something better. H264 is a delivery codec, it does not hold up to generational loss very well, etc. The benefits are definitely there to convert to something that's lossless.

Regarding bit depth, higher bit depths will give you better results, technically, but you may have a hard time seeing the difference. Most pro grading systems work in 32 bit (floating point) math to ensure the highest quality transformations to the footage.

Here are some posts from Stu Maschwitz, a very knowledgable source on the subject, required reading IMO:

http://prolost.com/blog/2007/2/13/mastering-in-the-nle.html

http://prolost.com/blog/2010/1/26/color-correcting-canon-7d-footage.html

http://prolost.com/blog/2005/1/26/color-correction-in-linear-vs-gamma-corrected-space.html
 
It's not necessary to work with different footage to use a higher bit depth within After Effects. Even if you read 8bpc (bits per channel) footage in and render and export to 8bpc files, working at higher bit depths within After Effects prevents losses of color information at each step of the process within After Effects.

There's an example here that shows how glows and blurs are less flat when you work with 32bpc color within After Effects, even when the output is to an 8bpc format.
 
Yes, it's definitely better to work at higher bpc , but only if your content "acts" like it's interpolated to higher bit depth.

BUT - The native T2i (or 7D) doesn't interpolate to higher bit depths in AE. If you work in 16 or 32 bpc mode, it acts as if it's 8bpc and you still get banding as if it was 8bpc .

All you are doing is wasting rendering time because it takes longer in 16/32bpc mode. The screenshots in my above post demonstrate what's happening. I didn't append a corresponding screenshot of the actual footage, but you can see it break up noticably (i.e it correlates with the actual footage). Try this out yourself, on footage that has gradients, like shadows, sky and do some levels/curves adjustments.

Even if you use the exact same footage, but re-wrap into .mp4 container, the banding disappears. The limitation is h.264 in mov container (at least on PC platform). But other issues appear , as AE does the Y'CbCr=>RGB conversion differently if you re-wrap it (you will see your levels shift as black & white are mapped differently.)

The point is , some formats respond well to working in 16/32bpc , some do not. The T2i/7D native footage do not respond as well, because of the .mov container. h.264 in other containers can be interpolated correctly and respond to higher bit depth in AE
 
Regardless of your source footage, work at 16bit for regular color correction. 8bit will introduce artifacts much faster and heavier.
 
You must log in or register to reply here.
Back
Top