So I think we’ve got a better spec now for the TransformIDs/filenames but what about the User Names?
Do people care as much about usernames or do you ignore them?
Now that we’ve flipped the TransformID definition ordering to be: LimitingRGB-W_Luminance_in_EncodingRGB-W_EOTF
it seems potentially confusing to have the UserName still be of the form: EncodingRGB-W EOTF Luminance (LimitingRGB-W Limited)
Example using output to Rec.2020 PQ 500 nit:
TransformID: urn:ampas:aces:transformID:v2.0:Output.Academy.P3-D65_500nit_in_Rec2020-D65_ST2084.a2.v1
UserName: Rec.2020 ST2084 500 nit (P3-D65 Limited)
Currently in Resolve, they use UserNames akin to what we recommend:
I don’t have a strong preference at the moment, but what I do like about Baselight’s list is that they use separators. It could be welcome possibly to also have this in some form.
It would also be very welcome if both the IDTs and ODTs would use consitent username conventions. Resolve’s ACESTransform list is quite a wild one especially on the IDT side. But I don’t know if these are the ACES suggested/provided names or not or if the team has any leverage on how they handle this.
Speaking as a user, one issue to be aware of is that very long names in an OCIO config often get cut off in the software display for many DCCs. This is in part because the display and view are combined currently in most DCCs, making them even longer than they are here.
With that in mind I do like that the more common ones like Rec.709 are just called Rec.709 as opposed to a verbose name like Rec.709 BT.1886 100 nits (D65 whitepoint) or whatever. IOW, I’m advocating for abbreviated user names where possible.
I feel like that is an issue such companies should preferably adress rather than put ACES in a place where they have to conform to such unoptimized implementations.
However I do agree that we don’t need a full descriptor. I would prefer at the minimum primaries and EOTF for the more common ones.
I recognize that no matter what we end up recommending as a UserName, we can’t always dictate what would be best for an implementor to put into their tool. They can decide how best to present the critical information to their users and places like ACES documenation will exist to fully describe all the technical characteristics associated with any simplified names. Also, ACESCentral or other places where users ask questions will exist where users can get clarity on what is best option for their use case - if the first time they go to select something it is labeled different than what they are used to.
I just want to try to get the UserNames that we include in the CTL into a form that is seems most workable as a base recommendation.
Rec.709 sure, but Gamma 2.2 alone doesn’t mean much right? Noting both primaries an EOTF will at least on the user level avoid most ambiguity.
Anyway, maybe I’m nitpicking. I’d also be fine with the current way it’s layed out in the Google sheet.
Understood. In the case of Resolve it looks like they will simply use the ACES UserNames since the current implementation also seems to match? (at least on the ODT side)
