Newby technical questions about Gamma

[Anton Shavlik]

I'm reading about ACES and realizing there's two terms I can't yet fully conceptualize. I figured Reduser is a good place to ask.

1. Is a "color gamut" a 2d plane of hues and saturations at middle grey and a "color space" a 3D shape that includes a luma dimension, as well as defining a white and black point in addition to the three color primary points?
If that's correct, is a color space also defined by a certain gamma or is gamma something completely external to color spaces? I understand sRGB and Rec709 have the same primaries and different gamma. Why?
Can does a color gamut define it's white point or is that something external?

2. Why do we have different gamma values? If the physical scene is linear light, and the camera sensor captures linearly, and my display is also linear, why do different gammas exist? Is it just to map the higher dynamics of the scene / capture to the lower dynamics of my display? If so why aren't gammas display dependant (since different screens have different possible contrast ratios) instead of a standard of ~2.2 with a well defined gamma function.

It seems like point #2 is not really relevant to an ACES workflow, as aces works in linear and will have a defined output display transform for each possible display format. But then it seems like the reference rendering transform takes the place of a gamma instead so I thought it seemed relevent. Have I misunderstood everything?

 

[Marc Wielage]

Read this as part of an explanation:

http://cinematiccolor.com

David Stump's Digital Cinematography: Fundamentals, Techniques, & Workflows goes into the details quite well, too.

http://www.amazon.com/Digital-Cinem...=sr_1_1?s=books&ie=UTF8&qid=1445840546&sr=1-1

Glenn Kennel's book Color & Mastering for Digital Cinema also goes into the nitty-gritty, though be warned his book is a bit dated (coming from 2006). Glenn was part of the team that created Cineon for Kodak in the early 1990s, and he's currently the head of Arri's U.S. Alexa division.

http://www.amazon.com/Mastering-Dig...=sr_1_1?s=books&ie=UTF8&qid=1445840840&sr=1-1

 

[Anton Shavlik]

Thank you very much, Marc.

 

[Marc Wielage]

The idea for ACES is that it accomplishes a couple of things:

1) it allows you to combine different cameras with different color space into one "even" playing field so that all of them can be corrected in similar ways

2) it allows sharing files with other artists (particularly VFX) to allow them to use files from many vendors interchangeably.

I think a lot of the advantages of ACES are theoretical from my point of view, but I understand why the industry is leaning in this direction. One hopes that it would help remove a lot of the uncertainty with multiple cameras and lots of artists working together on one project. I think if you have a closed system -- that is, one central person doing all the final work -- they can generally make everything match and work together. But big pipelines that involve dozens of people at different facilities, it can get pretty complicated very quickly.

I personally think if you just have a couple of Red cameras in one project, I see no benefit to ACES and I think you can correct them just fine in our existing color spaces. On the other hand, in a color-managed system, if you had to take one master and deliver it to the web (sRGB), broadcast (Rec709), theatrical (P3), wide-gamut (Rec2020), and HDR (Rec2084), ACES has the potential to handle all of those. I can't say how well these will work in the real world, but the idea has merit.

More here:

https://en.wikipedia.org/wiki/Academy_Color_Encoding_System

http://www.oscars.org/science-technology/aces/aces-documentation

My only caveat is I believe ACES is still in v1.0, so there are still some things being shaken out. Ultimately, it's very likely this is the way the industry will go in another couple of years.