Only apply white scaling if linear_scale_factor == 1.0
#144
Conversation
|
If we decide that white fitting is not needed for HDR encodings, due to the headroom inherent in the 10,000 nits of ST.2084, a different mechanism to signal "don't apply white fitting" would be needed. Maybe it should be an explicit boolean parameter. It's not a problem with the current list, but if we add D60 sim HDR versions it will be. |
|
A bool has been added to determine if or if not to apply the white scaling 414b4bf This allows the user to control when the white is or isn't applied, and is clearer and more robust than relying on a mismatch in the white point chromaticities. For example, a DCDM P3D65 ODT already has white point fitting handled by a 48/52.37 scale factor, so wouldn't need that additional white scaling that would otherwise be applied because the encoding white and limiting white would differ. This commit relies on a corresponding change to the Output Transform to add a "scale_white" bool and include it in the call to the display_encoding/decoding functions. |
|
@scottdyer I think there is a small typo on the encode function, it says |
|
Yes, thanks. I also caught this while running tests but haven't yet pushed the fix |
The scaling to fit peak white where the limiting and encoding white points do not match is not needed for DCDM encoding, since the necessary headroom is already provided by the 48/52.37 normalisation. A linear scale factor less than 1.0 seems to be a suitable way to identify that headroom is already being provided in the encoding. Without the changes in this PR, the DCDM Output Transforms will have white fitting scaling applied based on the difference between the limiting white and equal energy white, when no scaling is actually required.