Preview LUT for linear Nuke workflow(?)

[Sean Cunningham]

Greetings. I've got some Helium footage (will be getting a lot more of it) that will require VFX. Transcoding to EXR is straightforward enough but I'm hoping that there's a way to generate a LUT that I can use for preview and Quicktime generation in Nuke that will apply the tone mapping and highlight roll-off that's used by Redcine-X, Scratch or Adobe if the R3D files are viewed directly.

I get why this grade-before-the-grade step wouldn't be desirable in linear or log exports for post production outside an R3D workflow. The colorist who will receive our final composites as EXR sequences doesn't need this but I'd like to offer some continuity for the offline versions of our shots provided to the editor and director as rec709 Quicktime files. Curiously, the Nuke preview LUT that I have for Alexa LogC->rec709 is almost dead on, apart from some saturation spikes, particularly in blue, but it nails the highlight roll-off. I imagine this would be a lot more straightforward if Nuke was updated to support Helium files.

Many thanks

 

[John Schick]

Hey, one of the forum members created a nuke gizmo for converting between linear light and Log3G10:
http://www.reduser.net/forum/showthread.php?152798-Log2Lin-(RLF-amp-3G10)-for-RED-Nuke-Gizmo

From there (assuming you’re using REDWideGamutRGB as your colorspace), you can apply one of RED’s official IPP2 Rec709 transform LUTs (they include various versions of contrast & rolloff):
https://www.red.com/downloads/58de98eff77f300ebe00053f

If using these LUTs for previewing, you’ll most likely want to disable nuke’s default sRGB/Rec709 viewer gamma.

 

[Sean Cunningham]

Thank you, I will give this a shot.

Per our colorist, the EXR are exported with "Dragoncolor". They were shot "Dragoncolor2", per the meta, but the colorist noticed a lot of blue clipping. When you export EXR or DPX from Redcine-X, as told to me, it ignores the gamma specified and always uses "REDLogFilm", but the primary colorspace does matter. I've got a possibly similar workflow for viewing Arri footage.

 

[M Most]

Thank you, I will give this a shot.

Per our colorist, the EXR are exported with "Dragoncolor". They were shot "Dragoncolor2", per the meta, but the colorist noticed a lot of blue clipping. When you export EXR or DPX from Redcine-X, as told to me, it ignores the gamma specified and always uses "REDLogFilm", but the primary colorspace does matter. I've got a possibly similar workflow for viewing Arri footage.

Not quite. It exports in Linear (NOT RedlogFilm), as it should for Open EXR (DPX is exported using whatever settings you have set). The color matrix is, as you note, applied. So if you export with Dragoncolor applied, you get a different result than with, say, RedWG. The only real option in RedcineX for Open EXR export (other than compression options) is the use of ACES primaries, which I assume forces AP0.

 

[Sean Cunningham]

What I mean is the gamma of the linear export is effectively the same as exporting a standard Cineon DPX which is REDLogFilm. I import an EXR sequence from Redcine-X and a DPX sequence saved from either Redcine-X or Resolve and do a standard log->lin on the DPX sequence and they're identical. However, neither exhibit what I'm assuming is the SDK's nominal application of the tone mapping curves for contrast and highlight roll-off, which aren't a part of either the exr or dpx transcode process (and rightly so).

The colorist doesn't need this and he's already signed off on the basic linear conversion after the switch to "DragonColor", avoiding blue clipping. But I anticipate the editor and/or director will notice the difference between preview Quicktimes we send based on the linear conversion and what they're replacing in their timeline. Their files will be QT export from Redcine-X, which will have the tone mapping and roll-off curves applied. It's a subtle difference but they've already unwittingly assumed it was a difference in colorspace, or gamma, and were naively trying to force us to accept Quicktime files as plates with the notion that "it's 444, so it's all there".