Color Matching Color Charts With RAW and Log in DaVinci Resolve

So the gamma conversion is a similar to a lut its not taking into account the grayscale patches. It would be nice if you could specify the desired white and black code values and have them apply a custom curve for you.

RE: Gamma – I’m not sure I understand the comment about the Gamma Conversion being similar to a LUT? To me, the key for best success is to not let Color Match apply any gamma conversion. Our job is to simply define the gamma as it is (which is why I like to keep a debayered RedLogFilm in the Log space when doing Color Match).

I do think Color Match may have been designed to do both the Color and Gamma Match… but it does the latter poorly so I do everything I can do limit the work it has to do… and then it works wonderfully.

RE: White / Black targets: It would be good if Resolve showed us the target White and Black code values so we could properly place them before applying the Color Match… since those values are different for the different test charts (and it can be tough to find those values at the moment you need them).

Thanks Patrick.

RE: White / Black Targets – Actually, I’m reconsidering if knowing the Black / White Target values are worth anything… especially in Log / RAW workflows where you design the Color Match to happen before the contrast expansion. In that case, the proper White / Black Targets will depend on the camera and the specific Log gamma profile of that camera.

In Part 2 I’ll show you how to ‘intuit’ if you’ve got the proper code values set… it’s simple as pi (and you’ve already seen me figure it out using the middle gray values here in Part 1).

Thanks for the question!

Hello Patrick,
this inside is pure gold.
I just noticed that you didn’t care about the color temperature of the Raw file. It was set at 5600K instead of 6500k already set in the match menu. Was that irrelevant?

That’s a really good point. It depends. When the grayscale seems spot-on but the color chips are way off, then color temp will usually fix it.

In fact, a refinement of this workflow is to look for the most out-of-bounds color chip (in our example, blue) and then tweak the Color Temp to see if it brings it (and the others) closer to baseline. I hadn’t really thought of that as the final step in this process. Nice insight!

Hello Patrick,
I am just wondering you advice on using this technique in prep before a Job – theory is to shot a DSC one shot with all say 3 cameras bring it in to resolve – then check if the vector scope matches up so you know all the cameras match up – if needed adjust to match the main camera – when done could you then save this out as a grade and use it though out the shot on the first node before the LUT is applied to adjust for any colour differences? If not do you know any other way to to so this? or would you need to shot a colour chart in different lighting etc INT EXT etc or does it not matter.

In professional multi-cam truck shoots, they match all the cameras to charts before recording – and then live-switch (usually with an engineer riding the levels as cameras come in and out of matching each other).

What you’re talking about is something similar but in post production. You’d align each camera to the chart it shot, as a starting point, and use that as a starting node. But like an engineer on a live shoot, you’ll still find you’ll need to ride levels as the lighting changes.

Also remember, when aligning to the vectorscope, the targets shown in vectorscopes are for SMPTE bars. Most color charts don’t align to those targets (they’re too saturated for most printing techniques). So you want to tweak the waveforms between cameras to match each other.

Thanks Patrick!

Hi Patrick,

Quick question on selecting the “target color space” with blackmagic Cinema DNG files. There is no “Blackmagic Color Space” available and when choosing rec.709 it really messes up the footage in funky ways. RedColor2 seems ok, but I just want to make sure it’s ok to use that one or if there is a better choice that is closer to “Blackmagic Color Space.” Thanks

Is “color match” a destructive process? I feel like it is since there is no place to adjust the individual values given in the results.

So I’m more and more convinced that the colormatch is destructive, as it’s hard to fine tune things once it’s applied. Even when applying it with a BM Source and Target Gamma along with a Target Color Space of RedColor2 (as no BM color space is available in the drop down menu). For example, if I do a color match and then add the CST to cineon along with the Koji Lut, the contrast is way too strong. Does this mean that using the colormatch tool in resolve prevents using a LUT on top? Would really love some feedback on this. Thanks!

Do you know if it’s possible to use a custom color chart in Resolve, for example an IT-8 one? I have all the reference XYZ/RGB data, would be great if there was a way to import it somehow in some form.