Hello all,
I need some guidance on how to properly roundtrip from DaVinci Resolve to After Effects and back again using ACEScg. Specifically, I want to be able to export from Resolve ACEScg footage to Ae and lay in some titles and then have those titles come back to Resolve as pure white like they were intended. I have successfully done this using ACEScct but am not able to do this using ACEScg as the titles come back with a slight off white or gray.
I have uploaded my documentation with screenshots attached to this message where I have successfully been able to do this using ACEScct but now I need it to work via ACEScg for full compositing flexibility.
Hi there!
I don’t see any additional information to your process, could you share it again?
We’d need to know a few things to better understand your intent. Does any of the ‘white text’ data need to be ‘integrated’ into a scene of sorts? Or can it be it’s own layer on top?
What is the format you’re exporting your ACEScg plates from Resolve to?
How is your AE color management set up?
What is the project working depth and which value are you using for the ‘white text’?
What is the format you’re exporting out of AE?
Hi Shebanjah thank you for taking the time to look at this. I have attached my documentation that shows how I can get it to work with ACEScct but am unable to get it to work with ACEScg. Am I missing to understand something?
And to answer your questions:
The ‘white text’ data can be it’s own layer on top.
A full 1 on each value as the “white text” is an After Effects generator element.
Hey Dave,
I think I understand where things get messy. I also have some comments on the layed out workflow itself. It looks like you throw DWG/DI as grading space and OpenDRT as the DRT in the pipe. This would mean that the workflow isn’t strictly an ACES workflow and while you can make use of the ACES mechanisms, I don’t really see the benefit from it if you have no intent to use their display transforms and work in ACEScct (in Resolve). But still regardless of following a true ACES pipeline or a sort of hybrid, for getting ‘white text’ into the mix the following should be considered.
If the workflow is set up in a way that the display rendering always happens last, the value from the white text needs to have a ‘scene’ value that lands at 1.0 in display space after the tone mapping occurs. In a project wide ACES managed this would be achieved by converting the inverse display transform back to the working space. Inverse display transforms are used to preserve the appearance of already display rendered imagery. (But that only works if it’s the same as the forward display transform, you can’t combine it with OpenDRT)
If you roll manual node-based color managed setup, it would be more practical to only apply the ODT to camera/cg footage and add Rec.709 white text elements as separate layers on top.
For getting plates to AE to serve as reference there would be no benefit in making them ACEScct or ACEScg. Either a base display rendered Rec.709 prores file or one that also includes the look would be more practical. There is no need to use ACES OCIO in AE if you only need to create text elements. I would set the footage as guide layer and export prores + alpha for the resulting white text animation.
To answer the actual issue you’re having. If you create a value of 1.0 in ACEScct space, it means a linear value of 222.861. This value is so high that after the display rendering it becomes 1.0 in display space. ACES 1 would require a scene-linear value of around 16 to become ‘white’, ACES 2 quite a bit higher but where your problem emerges is that in ACEScg which is a linear space, your created value of 1.0 remains 1.0 and thus becomes ‘grey’ after the display rendering as a DRT ‘makes room for specular highlights to make a pleasing image’.
I hope that makes sense.