Need help understanding RedCineX Color Space and Gamma Space

[Omar N]

Just to be clear, I am understanding that shooting and grading in redcolor3 is the way to go, and working in redgamma3 or redlogfilm is best for for gamma settings.

What is the difference between redgamma3 and redlogfilm? Is the gamma curve just a little bit different?

 

[Uli Plank]

It's massively different. RG3 is for viewing and REDlogFilm is for grading. It will look milky in direct viewing.

 

[Graeme Nattress]

Think of RG3 as a pre-built beauty grade on your RAW footage, and RLF as a neutral starting point for you to grade from.

Graeme

 

[Nick Shaw]

You can think of REDgamma3 as an S-Curve put on top of a REDlogFilm image. It just happens to be an S-Curve designed by RED to complement the image from the camera in what they consider the optimum way for a 'one light' dailies grade.

I have made an approximation (a very good approximation IMHO) of that curve, and it is available to buy from tomorrow as either a plugin for FCP7 (demo here) or LUTs for Color and Resolve (see this ReadMe for more info.)

UPDATE: My LUTs and plugin are available NOW

 

[Elias Heuninck]

Hi forum!

I have a question about color consistency in and beyond the RED workflow.
I'm not familiar with color management and I'm probably about to say something silly, but currently I have the idea that color- and gammaspace—rather than a look—work as a reference to a standard so the image would stay the same through an entire workflow, swapping applications without too much color differences.

For example: I want to pull stills from my R3D clips, import them into Photoshop to edit and finally print them from Photoshop on paper.
I tried several ways to get the stills into Photoshop (the new r3d Photoshop plugin included), with and without attached color profiles, but the colors in Photoshop where always a lot different.

Workflow
-- r3d file from Red One: raw, but with the look we had when shooting (already the case when you open up RedCine-X i guess)
-- first color correction in RedCine-X (default (meta) settings: colorspace: REDcolor, gammaspace: REDgamma)
-- transfer to Photoshop (with as little color or luminosity shift as possible)
-- print (maintaining color consistency by using a monitor that's calibrated to the printer's output)

I'm confused by the colorspaces. Should I set the colorspace in RedCine-X to Photoshop's working colorspace?
I know this is a complex topic, but maybe you guys can point me in a direction?

thanks!
Elias

 

[Chris Bouchard]

For linear light integer formats, it's important to set the ISO down to 320 to ensure it doesn't clip or go into non-linear FLUT gain, and remember there's a black offset of 800/65535 in the data that must be subtracted in processing to get pure linear light. But there's better ways to get linear light... Always use a 16bit format for this kind of output.

1) export EXR as Rob suggests above. That is utterly linear, no offset, never anything non-linear and no ISO/FLUT range restrictions or anything.
2) export with REDLogFilm because it linearizes perfectly back to linear light using the standard cineon equation in Nuke, Shake etc.

Graeme

If exporting REDLogFilm for compositing in true linear light, should the ISO be set to 800 or 320?

thanks!
Chris

 

[Graeme Nattress]

RLF is happy anywhere in the 320ISO - 3200ISO range.

Graeme

 

[Chris Bouchard]

RLF is happy anywhere in the 320ISO - 3200ISO range.

Okay. But what about for VFX use when you want get real-world linear light data from the RAW - surely only one specific ISO/gain will result in actual original values (without adding digital gain)...?

 

[Graeme Nattress]

Any ISO in that range will produce scene referred linear light values upon linearization of the RLF. I don't know why you'd not want any gain applied as then you'd not have a defined rendering intent, showing the image as either too bright or too dark, depending on how it was shot.

Graeme

 

[Chris Bouchard]

Thanks Graeme! No problem with exported files - all is linear in Nuke and gain is reversible! I realized the FLUT/ISO range restrictions appear to be still active in the RCX histogram when gamma is set to linear. So that was what confused me.

You said earlier in this thread that FLUT/ISO 'range restrictions' are disabled when exporting EXR. May I ask if this is also the case when you export RedLogFilm?

 

[Graeme Nattress]

Thanks for the feedback that it's working.

As EXR can handle a vast range, we don't restrict ISO gain, and FLUT and ISO are pure gain at that point for pure linearity.

In our integer formats, like TIFF and DPX, there will be some compression of linearity when exporting a redlogfilm at some point above a gain equivalent to 3200ISO, which although will make the extreme highlights less linear, it's better than a hard clip.

Graeme