Is there a way in Nuke to check the colorspace of the imported footage?

Discussions - Using ACES
21

Yes, I think that’s the intent of AMF.

22

There was an interesting conversation on the OCIO slack about exr and metadatas. Here it is :

Larry Gritz :

We had an OpenEXR TSC meeting today in which we thought some OCIO folks would attend to talk about the intersection of OCIO color space names, OpenEXR chromaticity attributes, and some practical (and contradictory) uses thereof. […] I think that in OpenEXR land, we feel like we’re not sure what to do next without some guidance from OCIO and want to align the two projects’ vision about what we can/can’t and should/shouldn’t do to document what color spaces a file’s pixels are assumed to be in. (There was a bit of a thread about this on the openexr dev mail list.)

[…] There are strong arguments that metadata can’t be trusted because it’s too easily wrong. But that seems unsatisfying to me – I still am faced with practical real-life problems like (for example) oiiotool in.exr -colorconvert foo bar -o out.exr and just trying to grapple with basic questions like “if in.exr had chromaticity attributes, do I propagate them to out.exr, or silently erase them, or write new ones that match the ‘bar’ colorspace (and can I even figure out bar’s primaries?).” And “should I write anything into the attributes of out.exr indicating that we think the pixels refer to ‘bar’?” And, oof, “are the names so arbitrary that it’s all meaningless if we aren’t sure it’ll later be interpreted using the very same OCIO configuration?”

[…] Rod Bogart made a good argument for “you can never know anything”, but I really want to believe we can do better than that somehow.

[…] There was also mention of the issue I believe we’ve talked about before, of possible future ability for OCIO API to allow basic queries about color spaces that would be useful to know: Is this named color space linear? What are the primaries? Is this color space really ACEScg underneath, despite its unique snowflake name?

Doug Walker :

The only easy part is that if you are doing a colorconvert you should clearly not propagate the chromaticities from the original file.
For the other parts, we have several features related to this thread on our OCIO 2.2 roadmap. At your next meeting I will plan to present what we are thinking of to try and help the situation.

Thanks,
Chris

2 Likes