CinemaDNG color management

I’m new to Davinci Resolve, and I’ve had hard time in understanding colour management in Resolve. So, maybe someone can enlighten me a bit.

To explain my issue, I made an example test with one .R3D clip, one CinemaDNG clip (BMPCC 4K) and one ProRes422HQ (BMPCC 4K). All clips shot in the same setting (lighting, ISO, aperture, shutter, color balance,…). I understand that there are differences between the (RED Scarlet-X & BMPCC 4k) chips and how they are handling color and gamma. However, I have difficulties understanding how DaVinci Resolve handles color management between all these formats (.R3D, CinemaDNG, ProRes422HQ). It seems that Rec.709 Gamma2.4 is baked in to ProRes422HQ, but it’s baked in also to CinemaDNG (RAW), but it is not baked in .R3D. Is this true?

An example:
Project Color Management, Camera Raw (CinemaDNG) and Camera Raw (RED)settings:

1+.jpg (139.24 KiB) Viewed 3459 times

.R3D clip with CST (upper) and CST disabled (lower):

2+.jpg (354.65 KiB) Viewed 3459 times

CinemaDNG (upper) and AppleProRes422HQ lower. Both Color Space: BMPCC 4K, Color Science Gen 4, with no corrections:

3+.jpg (372.33 KiB) Viewed 3459 times

As one can see, non-corrected CinemaDNG is very close to non-corrected ProRes422HQ (CDNG just slightly lighter). And corrected .R3D is quite close to both non-corrected CDNG and ProRes422HQ. And, as expected, non-corrected .R3D is far from being even near to non-corrected CDNG or ProRes422HQ.

In my understanding, ProRes422HQ has Rec.709 Gamma 2.4 baked in, right? And that’s why it shows “correctly” on a Rec.709 Gamma 2.4 timeline, right? However, in my logic, I’m puzzled why CinemaDNG RAW seems to have Gamma2.4 baked in also? I thought that RAW formats are not Rec.709Gamma2.4 (color or gamma). Have I missed something essential?

Thanks in advance!

However, I have difficulties understanding how DaVinci Resolve handles color management between all these formats

You're not using a color-managed project, so what you're seeing is just the result of the different color sciences of the cameras in a Rec. 709 timeline, with the CST nodes enabled and disabled. Blackmagic's film gamma curve is not that far from Rec. 709, which is why both RAW and ProRes don't change a huge amount. The Scarlet-X uses a much flatter curve, as you have seen.

Note you can also use a CST plug-in in the node tree to do the same thing. These are covered in the manual.

It's also possible to work completely without altered color management and just go with the normal daVinci YRGB color management and use the Raw controls to get the cameras close, then use curves and LGG to color-correct the images. The latter is what we did for 25-30 years before color management was possible.

It does help to shoot color charts with known values on them, and then you can assess how they look on scopes. That both helps to match cameras as well as to know where they normally fall with a "mid-range" adjustment.

Don't make this harder than it needs to be -- matching different cameras is a very standard job for colorists. Often, I'll get a feature where 10 shots were shot on a Tuesday, 5 other shots in the middle were shot on a Saturday, and then several reaction shots were done six months later by a different DP... and they all have to match. We just grab stills and turn knobs until the scene visually flows as smoothly as possible.

Thanks Marc and Jack for your replies!

I understand that I can match cameras in numerous ways (RAW controls, nodes, curves,…). And I do understand the help of using color charts, etc. No problem in that.

What I’m trying to achieve is some kind of a logic in the beginning. And one illogic thing for me is that, while the CST works well for .R3D, it does not work with CinemaDNG in the same manner. I mean that CinemaDNG looks (in non-color-managed YRGB color science) close to the ProRes422HQ by default (with no CST), but it looks very different when CST applied. In my logic the CST applied to CDGN should take it to Rec,709 Gamma2.4, right? (edit: close to the ProRes422HQ (Rec.709 Gamma2.4)) But it doesn’t. (CST lifts all tones up, but only the highlights can be lowered with max input value.)

Here’s an example:
ProRes422HQ

6.jpeg (173.74 KiB) Viewed 3251 times

CDNG without CST

5.jpeg (177.25 KiB) Viewed 3251 times

CDGN with CST

5+.jpeg (207.28 KiB) Viewed 3251 times

So, where's the logic?

Thanks again in advance!

It looks like you shot that ProRes clip in Blackmagic Film gamma, not in Rec. 709/Blackmagic Video. If you apply a Blackmagic Film to Rec. 709/Video LUT or CST, it should match the result of doing that to the DNG clip.

Lauri Astala, what was the conclusion here?

In the Camera Raw setting, if you select Decode Using > Clip, there will be other choices for Color Space and Gamma.

If instead of Blackmagic Design, you choose e.g. P3-D60/Linear, you can then add a CST as the first node to transform that to your preferred timeiine.

This article discusses in more detail the idea: https://blog.sigmaphoto.com/2020/fp-raw-workflow-with-pro-timur-civan/