Re:I accidentally Blender VSE
by celegans25 on Thursday February 08, 2024 @07:09PM
(#39309609)
> and replacing them with just straight up floating point calculation
Not only that, the diff replaces 100-150 lines of code with 3. That's the kind of pull request I can only dream of making
Hardcoding gamma=2 is doing the heavy lifting, though.
Re:I accidentally Blender VSE
by olddustytrail on Friday February 09, 2024 @04:13PM
(#39320617)
Gamma always equals 2, so it's a natural optimization. Assuming of course you are 0 indexing the alphabet.
Yeah a nice optimization, but is "return c * c;" really a good approximation of sRGB gamma?
No it's not. But, the previous code was already effectively doing "c * c" for the last 15 years. So for now, just keep doing that, a bit faster.
A more proper way would be to do proper color-space aware luma calculation. Which under default settings is sRGB indeed, but not necessarily so (VSE can be set to operate in some other color space). Someday!
Yes, it's weird they used 2.0 in the original code too. Doing proper gamma for any regular space (sRGB, YUV Rec.705, etc.) isn't actually that heavy (even without LUTs).
My guess is that the code was written by someone in 1995 back when no one understood color spaces, or something (it's hard to track down who and when wrote it exactly due to all file moves and refactors etc.)
This change [1] was my favorite, optimizing an image processing step by removing lots (and I do mean lots) of table-based "optimizations" and replacing them with just straight up floating point calculation, and making it ~4X faster on the OP's machine. Fantastic!
[1]: https://projects.blender.org/blender/blender/pulls/115801/fi...