To clarify this, it’s not that the viewer gain is being applied at a different
time, but rather that your OCIO nodes are evaluated at a different step in the
display pipeline based on whether they are in the node graph or part of an
input process. Correct me if this is not what you’re seeing, but unless you
have your viewer gain adjusted, the result will the same between the two setups.
The difference is due to the fact that the viewer gain is applied *before* the
input process and ViewerProcess (unlike the viewer gamma, which is the last
stop in the pipeline). Assuming the default input process application order, I
believe the resulting pipeline looks like this:
Node graph –> Viewer gain –> Input process –> ViewerProcess –> Gamma
You can see how moving your input process nodes into the node graph changes the
evaluation order.
-Nathan
From: samb
Sent: Friday, March 08, 2013 5:01 PM
To: [email protected]
Subject: [Nuke-users] Re: OCIOColorSpace reacts differently within aVIEWER_INPUT
Hey John,
What appears to be happening is the viewer gain is being applied at a different
time using the VIEWER_INPUT
VIEWER_INPUT is going:
gain > colorspace > colorspace
Whereas the the other viewer is:
colorspace > colorspace > gain
Adding a gain node above your colorspace nodes and matching what your doing in
the viewer output seem to make them match. Also when the gamma is 1 both images
appeared the same to me.
Hope that helps!
-Sam
[quote="John Stewart"]This is on the RGB. Here is a detailed screenshot of the
setup. I've paired things down. The 2 viewers are gained down to show the
difference easier. This is Nuke 7.0v3
John
Jeremy Selan wrote: [quote] Quote:
I would expect them to be the same. Is this on rgb channels, or a
different layer? Can you provide an example demonstrating this? What nuke
version are you in?
-- Jeremy
On Tue, Feb 5, 2013 at 11:03 AM, John Stewart ([email protected])
wrote:
Quote:
I've just encountered this and I wonder if anyone else has seen
this.
I've got a VIEWER_INPUT containing 3 nodes total, all of them OCIO
nodes.
If you were to compare:
A) an image with the IP on
vs.
B) an image with the 3 nodes of the VIEWER_INPUT connected directly
to the
plate
The B result are totally different from the A results. In theory,
these
two ways of applying color transforms SHOULD be the same, no?
Thoughts?
John
______________________________**_________________
Nuke-users mailing list
[email protected].**co.uk
([email][email protected].**co.uk[/email])
([email protected]),
[url=http://forums.thefoundry.co.**uk/]http://forums.thefoundry.co.**uk/[/url]
[url=http://support.thefoundry.co.**uk/cgi-bin/mailman/listinfo/**nuke-users]http://support.thefoundry.co.**uk/cgi-bin/mailman/listinfo/**nuke-users[/url]
--------------------------------------------------------------------------------
_______________________________________________
Nuke-users mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users_______________________________________________
Nuke-users mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
