I've been trying to make a push to integrate OpenColorIO heavily into
our pipeline, and I'm at a bit of a crossroads with regard to how to
proceed dealing with Read and Write nodes in Nuke, and I'm curious
what people think, and what other studios are doing.  My issue is that
the colorspaces in the Read/Write nodes only show luts from Nuke's
list of 1D luts.  I'd like a single place to control my input
colorspaces.  I see three options available:

1) Script the creation of a new set of Nuke 1D luts on startup, based
on the OCIO config.

2) Develop a fancy Read and Write node that incorporate an
OCIOColorspace node, either as a gizmo or some sort of C++ inheritance
thing that I'd rather not deal with.

3) Wait for Nuke to release a version that lets us opt into native
OCIO support for the Read and Write nodes.

Option #3 seems particularly compelling, and seems like it should be
coming soon, given that Hiero already relies entirely on OCIO for its
colorspaces, and it would obviously behoove The Foundry to have their
color management consistent across their own apps.

If option 3 is going to happen, but not right away, I suppose I'd opt
for option 1, just to avoid complexity and proprietary nodes that get
deprecated within the year.

But if it's going to be a long time or maybe never happen, option 2
would make sense.  I'm kind of assuming some people have already done
option 2, so if anyone has tips from that angle, or could point me in
a good direction, I'd be curious to hear them.  I know how to make a
basic gizmo, but getting all the nuanced behavior of a read and write
node seems harder.

A big disadvantage of option 1 is that it's completely limited to 1D
luts.  I could sort of live with that, as we're mostly working on
commercials, but it's far from ideal.  Especially since gamut is the
thing that confuses people the most.

Any thoughts?  Thanks!
_______________________________________________
Nuke-dev mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev

Reply via email to