The problem is that 3D luts doesn't work with pixel data above 1.0. To get around this you need to apply the lut in a a compressed range, for example in log. So basically you have 2 luts, first the shaper lut (1d) that converts your linear values to log (for example) and a second lut that goes from log to destination.
As you seem to be using OCIO, have you had a look at the command line tools it offers? With ociobakelut <http://opencolorio.org/userguide/tool_overview.html#ociobakelut> you could create a lut and also set the shaper space. I guess you would need to create a OCIO colorspace with your four transforms first though. As a quick and simple workaround in Nuke, you could try the method below. It actually never made any sense to me but it does work. set cut_paste_input [stack 0] version 9.0 v8 CMSTestPattern { inputs 0 name CMSTestPattern1 selected true xpos -6625 ypos 3022 } Colorspace { colorspace_in AlexaV3LogC name Colorspace2 selected true xpos -6625 ypos 3168 } Grade { black 0.034 white {1.254999995 1 1 1} multiply 3 gamma 0.81 name Grade4 selected true xpos -6625 ypos 3393 } GenerateLUT { file C:/Users/Simon/Desktop/tmp.cube file_type .cube name GenerateLUT1 selected true xpos -6625 ypos 3571 } StickyNote { inputs 0 name StickyNote5 label "Input colorspace is set to (Alexa) log. This basically does a lin > log converion before applying the lut." selected true xpos -6454 ypos 3793 } StickyNote { inputs 0 name StickyNote6 label "Increse resolution if needed." selected true xpos -6477 ypos 3051 } StickyNote { inputs 0 name StickyNote7 label "Set 1d pre lut" selected true xpos -6473 ypos 3184 } push $cut_paste_input Vectorfield { vfield_file C:/Users/Simon/Desktop/tmp.cube version 1 file_type cube colorspaceIn AlexaV3LogC name Vectorfield2 selected true xpos -6629 ypos 3799 } StickyNote { inputs 0 name StickyNote4 label "Add color transforms here (assumes linear input, no log > lin stuff)" selected true xpos -6468 ypos 3387 } ------------------------------- Simon Björk Compositor/TD +46 (0)70-2859503 www.bjorkvisuals.com 2016-04-21 17:04 GMT+02:00 Erik Johansson <erik.johans...@fido.se>: > So I can't figure this out myself so i'll try it here. > > I am trying to generate a lut for a shot that would subsitute four other > nodes. 2x OCIOColorSpace, 1x OCIOCdlTransform and 1x OCIOFileTransform. > > I put down a CMSTestPattern and connect the nodes to it and then a > GenerateLut to write out a new lut. > > When I then try and use the generated Lut on the original plate everything > looks the same for values in the plate that are <1 but in the overbright > areas the newly generated lut clamps the output. > > There is some threads on it on the forum that says it's because of the > CMSTestPattern only containing values between 0 and 1 but no info on how to > work around that. > > Does anyone happen to know? > > Cheers, > Erik > > > > > -- > > > *Erik Johansson**Pipeline TD* > > > *Fido*Rosenlundsgatan 40 > 118 53 Stockholm, Sweden > www.fido.se > > _______________________________________________ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >
_______________________________________________ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
