28 Mar
2023
28 Mar
'23
8:20 p.m.
but I fear that something still needs care here: digging a bit further on the 2% distance: out of 96*96*3 bytes compared, 5834 have an error which is more or equal to 2 (and some greater than that), which looks suspicious from what should pure rounding discrepancy. Didn't look further.
Fwiw I don't think such low differences matter very much. I think it's likely some diff in color sampling, and I think windows does a bad job at it with a lot of color bleeding. There's already some differences between windows versions in some tests.
Also the test produces a .bmp file that can be used to visually compare the results (for YUV images, the .bmp is only a representation of the data, which is appended to the the .bmp data).