> On Mon, Sep 20, 2021 at 10:21:23PM -0400, Manuel Stoeckl wrote: > > This resolves a problem where conversions from YUV to X2RGB10LE > > would produce color values a factor 4 too small. > > > > The variable 'yval' used later in the switch case 30 has range > > [0,255], but the color channel values in X2RGB10 have two more > > bits than 'yval' and can go up to 1023. Increasing (r|g|b)base > > by 2 effectively multiplies yval by 4 and fixes this discrepancy. > > does white have 1023 ? or 1020 ? > a multiplication by 4 would not in general produce a correctly scaled > 10bit per channel value from 8bit
I've replied to this email with a new version of the patch set, that now ensures the channel values in the YUV->RGB table created in ff_yuv2rgb_c_init_tables go up to 1023. However, this does not mean the code will always convert 1.0 white to 1.0 white; for example, pure 0xffff RGBA64 white gets converted to X2RGB10 grey level 1019. Presumably there is an off-by-one rounding error somewhere in the RGB->YUV->RGB conversion path, but I don't think tracking it down is worth it at the moment. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
