Hugh
Active member
Hi all,
I've been doing a load of testing recently on some footage that we're going to be doing VFX on, and would love to hear if any of you have been seeing the same kind of things as I am...
Any screenshots below have been exported using REDline (Build 365), with the colour version and then the three separate channels. They've been exported with a Gamma curve of RedFilmLog. They are half their original resolution.
Before I start, the main issue that I'm concerned about is the over-saturation, leading to banding, or even bottom-end clipping in some channels.
Before I post the actual images, I just want to point out that these are from *all* of the colour spaces. Some of them I would never consider using anyway.
CameraRGB:
RedSpace
RedColor
RedColor2
sRGB
Adobe1998
rec709
Now, all of these, apart from CameraRGB, show, to a greater or lesser extent, the effect that I'm not so keen on - primarily in the saturated yellow in the bottom right (noticeable in the blue channel).
From these, the only one that I'm particularly comfortable using is CameraRGB, as it's the only one that doesn't seem to have had something done to it to increase saturation. Which would make sense, if it's the raw RGB from the camera.
CameraRGB, however, feels very desaturated. This is concerning some people here over what is going to happen with the CG - the textures are going to be from DSLR stills.
Now, I see the following options:
1) Pass the textures through a 3D Lut prior to rendering, which will match the DSLR colourspace to CameraRGB. Allow lighters and Compers to use a viewer lookup which increases saturation slightly
2) Tweak the saturation up on the lighting backplates and then let compositors desaturate the CG when they get it.
3) Up the saturation in a controlled way on all the plates, for compositors/lighters/everyone.
Right now, 1) is probably my preferred method. 2) Just sounds horrible, and liable to have all sorts of problems, and 3) doesn't fill me with a huge amount of confidence...
What would people suggest in this situation? I'm thinking Viewer LUTs would be far preferable to modifying the actual plates, and doing a conversion on textures before rendering....
I've been doing a load of testing recently on some footage that we're going to be doing VFX on, and would love to hear if any of you have been seeing the same kind of things as I am...
Any screenshots below have been exported using REDline (Build 365), with the colour version and then the three separate channels. They've been exported with a Gamma curve of RedFilmLog. They are half their original resolution.
Before I start, the main issue that I'm concerned about is the over-saturation, leading to banding, or even bottom-end clipping in some channels.
Before I post the actual images, I just want to point out that these are from *all* of the colour spaces. Some of them I would never consider using anyway.
CameraRGB:
RedSpace
RedColor
RedColor2
sRGB
Adobe1998
rec709
Now, all of these, apart from CameraRGB, show, to a greater or lesser extent, the effect that I'm not so keen on - primarily in the saturated yellow in the bottom right (noticeable in the blue channel).
From these, the only one that I'm particularly comfortable using is CameraRGB, as it's the only one that doesn't seem to have had something done to it to increase saturation. Which would make sense, if it's the raw RGB from the camera.
CameraRGB, however, feels very desaturated. This is concerning some people here over what is going to happen with the CG - the textures are going to be from DSLR stills.
Now, I see the following options:
1) Pass the textures through a 3D Lut prior to rendering, which will match the DSLR colourspace to CameraRGB. Allow lighters and Compers to use a viewer lookup which increases saturation slightly
2) Tweak the saturation up on the lighting backplates and then let compositors desaturate the CG when they get it.
3) Up the saturation in a controlled way on all the plates, for compositors/lighters/everyone.
Right now, 1) is probably my preferred method. 2) Just sounds horrible, and liable to have all sorts of problems, and 3) doesn't fill me with a huge amount of confidence...
What would people suggest in this situation? I'm thinking Viewer LUTs would be far preferable to modifying the actual plates, and doing a conversion on textures before rendering....
I'm seeing "Camera Firmware Version: 30.6.0#30"