EXR delivery from video game

(Charles Boileau) #1

Hi To all,

I was wondering if anyone could advise on a simple way to convert EXR Linear SRGB gamut images into ACES.

The bad part is I’d like to do this with Resolve…

Any ideas?

Thank you very much!

(Scott Dyer) #2

The conversion from linear sRGB to linear ACES is just a 3x3 matrix.

How’s DCTL?

Just put in the DCTL subfolder in the LUT directory, right click on the clips that need it, and apply like any other IDT (except of course it will show up in the ‘DCTL’ menu not ‘ACES Input Transform’).

(Nick Shaw) #3

Depends of course if the EXRs are scene referred linear, and include values >1. If they are display referred linear, and therefore restricted to 0-1, an inverse ODT might be required.

Resolve includes an inverse ODT for material with sRGB transfer and primaries, but not for linear sRGB. Writing one in DCTL would be quite complex.

(Scott Dyer) #4

@CharlesBoileau yes, what @nick said. I was assuming scene-referred linear sRGB, because you said they were in OpenEXRs. I have no idea why someone would put display-referred linear into an OpenEXR, but they could (and many probably have).
So if these are scene-referred linear (like what you might get from VFX), then the matrix should be all you need.
(Of course the other option is to just tell the VFX house to deliver ACES 2065-4 OpenEXR files, which would eliminate any need for IDT and also any ambiguity :wink:)

(Nick Shaw) #5

I have seen it done many times! It’s sort of what Nuke’s pre-OCIO 1D VLUT based pipeline might lead people to do. It’s also what you get from some 3D packages which have a linear output option. i.e. the artist is viewing through a 1D sRGB VLUT when animating/compositing, and makes the image “look right” on an sRGB display. Then they render a linear sRGB EXR.

(Charles Boileau) #6

I also thought it was scene referred but in the end I got test files and the are display referred… :frowning:

I think the guys at the video game editor only know how to output EXRs.

I’ll just apply the inverse ODT in Resolve.


(Charles Boileau) #7

Thanks for the DCTL @sdyer.

It’ll come in handy at some point!