Difference between gamma 2.6 and gamma DCI in CST

Hi,

I am trying to understand the difference between the 2.6 gamma and DCI gamma options in Color Space Transform.

My understanding is that both P3 DCI and DCI XYZ use a 2.6 gamma, however there is a noticeable shift between the two, DCI gamma appearing brighter than 2.6.

I am trying to find the correct transform to apply to render a project graded in DCI P3 to a DCDM DCI XYZ sequence for DCP production.

If I apply a transform from P3-DCI Gamma 2.6 to DCDM Gamma 2.6 and create the DCP from that and import it back into resolve, the image seems to be brighter when compared to the DSM exported in P3-DCI Gamma 2.6. Since resolve automatically applies its own color space adaptation for DCP sources though i am not sure that comparison can be trusted.

The DCP is created with EasyDCP Creator 4.0.10 using the XYZ (DCI) color space.

Should I rather apply a P3-DCI Gamma 2.6 to DCDM Gamma DCI tansform ?

Same problem here. Have you found out a solution?

Same issue on my end. Round tripping DCP back into Resolve has a lifted gamma. I am submitting to a festival that is projecting in DCI-P3 Gamma 2.6. Film is mastered in Rec.709 Gamma 2.4. Resolve project settings are DaVinci YRGB, Input timeline Color Space is ARRI LogC3, Output is Rec.709 Gamma 2.4.

Round trip steps are:
-CST output node on timeline node tree from Rec.709 G2.4 to DCI-P3 G2.6 (with no tone/gamut mapping or any other transform settings)
-DCP Render
-Import DCP MXF, lay it on its own Timeline with Input Color Space set to XYZ (CIE) Gamma 2.6, and Output Color Space set to Rec709 Gamma 2.4
-Apply CST to the DCP MXF on this timeline and set input to XYZ (CIE) / Gamma 2.6, and Output to Rec.709 / Gamma 2.4 (with no tone/gamut mapping or any other transform settings)
-Then take a still frame, switch to original Rec709/G2.4 timeline and lineup the frame and A|B. Color appears the same but Gamma is lifted.

If I change that final CST to input gamma to DCI instead of 2.6, the still frame reference lines up perfectly. So the only place in the chain I can think that the Gamma is shifting is the DCP export from Resolve. There is no step on the deliver tab to modify Gamma, so my assumption was Resolve is doing the DCP standard, which I believe is Gamma 2.6. Is this not true?

Is it a similar situation to Rec.709 and Gamma 2.4 where the real target is not quite 2.4, but people generally set their reference levels to 2.4. So technically its deviating from the SMPTE standard, but its become its own standard in a way. Is this another situation like that? Where DCI is the target, technically slightly different than 2.6, but 2.6 is more widely adopted or something? If that is the case, is DCP exporting from Resolve not going to work since it is outputting true DCI Gamma whereas the theatre I'm submitting too is asking specifically for a Gamma of 2.6 deliverable?

Any help would be greatly appreciated,
Spencer

Disregard my last post. I found the solution almost right after that post but it takes a day for the post to be approved.

First, to address the original poster's question more directly, my speculation is similar to the SMPTE standard of Rec.709. My understanding is that the target spec for Rec.709 gamma according to SMPTE's spec is not truly 2.4, but 2.4 has become widely adopted in most cases within the industry. Perhaps P3-DCI is similar, where DCI technically has a similar but slightly different Gamma than 2.6. But 2.6 has become the widely adopted standard by theatres/festivals for exhibition. Hence, Resolve CST offers both true DCI as an option, but also 2.6. This is my own speculation and would need to be verified. But if you look under Gamma in your CST, compare Rec.709 as an output Gamma and 2.4, there is also a shift as with DCI/2.6.

As for my round-trip gamma shift problem: User error, my mistake was the final CST steps needed to have XYZ (CIE) set to Linear and not Gamma 2.6. It goes to show that every setting in your conform chain needs to line up perfectly-- and how easily one setting will result in a shift in color/gamma. So, this doesn't help previous user's with the same question except to say perhaps make sure with absolute certainty that each CST step is correct! And my results have demonstrated that Resolve indeed renders DCP to linear faithfully.

The correct round trip steps were:
-Original camera negative was ARRI LogC3/AWG
-Project mastered in Rec.709/Gamma 2.4
-Master Project settings set to DaVinci YRGB, Timeline Color Space ARRI LogC3, Output Color Space Rec.709 Gamma 2.4
-CST output node on timeline node tree from Rec.709 G2.4 to DCI-P3 G2.6 (with no tone/gamut mapping or any other transform settings)
-DCP Render
-Import DCP file, lay it on its own Timeline with its Input Color Space set to XYZ (CIE) and Linear Gamma, and Output Color Space set to Rec709 Gamma 2.4 (adjust timeline settings to uncheck project settings and set these parameters instead).
-Apply CST to the DCP MXF on this timeline and set input to XYZ (CIE) / Linear, and Output to Rec.709 / Gamma 2.4 (with no tone/gamut mapping or any other transform settings)
-Then take a still frame, switch to original Rec709/G2.4 timeline and lineup the frame and A|B. Color matches in my case.