I am quite new to ACES so please bear with me. I have just started at a new company and the current colour workflow doesn’t seem quite right to me. Their work flow goes as follows. Scans are read in as RAW and then converted from ACES 2067 to ACES CG. DMP is done in a simular way with CG rendered in ACES CG. This is then all comped together and rendered out as RAW. This doesn’t seem correct as the elements are not put together in the same colour space inside nuke and rendering out as RAW doesn’t seem right too. I would have thought that the correct workflow would be to pull in the scan in as ACES 2067, convert it to ACES CG, do the same with the DMP and this would mean that the scan, DMP and CG are all working together in ACES CG. Then either rendering it out as ACES CG or convert it back to ACES 2067. Any thought on this would be greatly appreciated.
Hi Liam, thanks for welcoming me. Ok by scans I mean footage shot on Sony Venice 4k camera. The footage is being ingested by editorial and written out RAW for us to pick up in nuke and setting it to RAW in the read node. It’s being set to RAW by selecting the check box in the read node and not by setting the colour space to utility - raw in the read node. The footage is being processed by technicolour before it reaches us so I’ll have to double check how it’s being outputted their end. The intended output our end is once we render it it is send back to the client.
First thing you need to know is that “RAW” is a “dangerous” term used to mean a lot of thing. In reality nothing is “Raw”, it always come from somewhere , but in this context we can use raw to mean “no-operation” (no colorspace transformation).
So if I understanded correcty your workflow:
You are importing a footage that is processed by someone else , could be usefull to know what did they do on it and what colorspace it has.
You are checking the Rawdata checkbox wich just mean “no-operation”
You convert the data from ACES-2065-1 (AP0) to ACEScg (AP1): this mean you assume the source is an ACES AP0 file . (and this could also been done directly on the read node without checking the raw checkbox)
Everything is comped in ACEScg (as you told cgi/dmp element are converted to ACEScg too)
Then on the write node you check RawData meaning again “no-operation” which mean you are exporting ACEScg data.
Which we can resume to:
in-> AP0; working-space-AP1; out-> AP1
And for now i don’t see any issue with this workflow (anybody, correct me on this one) if this what it was intended for. Just need to know where are you coming from and where are you going.
Hi Liam, thank you so much for this. Yeah I guess when you lay it out like this is does make sense. I think I was just getting confused by the fact that the elements in nuke were being set to RAW - no operation. The last place I worked at everything was read into nuke in Linear colour space, worked on and then outputted with Linear colour space set in the write node. From what your saying I could set the colourspace to ACES CG at this new company and it would give me the same result as RAW (no - operation)?
From what your saying I could set the colourspace to ACES CG at this new company and it would give me the same result as RAW
As Utility-Raw mean no operation, if your working space is ACEScg, picking Utility-Raw or ACEScg will give you the same result:
Working space: ACEScg (destination)
IDT:
Utility Raw: No operation
ACEScg: Mean the source is in ACEScg and shoudl be converted to the working space: in this case nothing happens.
was read into nuke in Linear colour space
Big issue here is that Linear is listed into the “Colorspace” dropdown while it’s not a colorspace and this has probably confuse a lot of people.
Lot of people could explain this way better than me but basically with the default Nuke color-management, the Linear option is like a “no-op” operation , nuke awaits linear data, you tell the data is linear so it doesn’t apply any transformation.
But in that case the Linear term should not be used alone and probably mean you are using sRGB - Linear data.
If you have some time I could only recommending you this ressource to learn mroe about color-science:
