On 12/30/2019 8:06 PM, Derek Buitenhuis wrote: > On 30/12/2019 19:20, James Almer wrote: >> + ff_side_data_set_encoder_stats(pkt, x265pic_out.frameData.qp * >> FF_QP2LAMBDA, NULL, 0, pict_type); > > Does multiplying by FF_QP2LAMBDA even make sense here? That contsant > is for H.263's QP scale, not HEVC's.
Pretty much every call to ff_side_data_set_encoder_stats() in lavc does it, including libx264, so i figure it's expected. For example, encoding with -qp 28 i get > out= 0 st= 0 frame= 1 q= 25.0 f_size= 2726 s_size= 3kB time= > 0.010 br= 523.4kbits/s avg_br= 2180.8kbits/s type= I > out= 0 st= 0 frame= 2 q= 28.0 f_size= 181 s_size= 3kB time= > 0.010 br= 34.8kbits/s avg_br= 2325.6kbits/s type= P > out= 0 st= 0 frame= 3 q= 29.0 f_size= 61 s_size= 3kB time= > 0.010 br= 11.7kbits/s avg_br= 2374.4kbits/s type= B > out= 0 st= 0 frame= 4 q= 30.0 f_size= 27 s_size= 3kB time= > 0.042 br= 5.2kbits/s avg_br= 573.9kbits/s type= B with it, but > out= 0 st= 0 frame= 1 q= 0.2 f_size= 2726 s_size= 3kB time= > 0.010 br= 523.4kbits/s avg_br= 2180.8kbits/s type= I > out= 0 st= 0 frame= 2 q= 0.2 f_size= 181 s_size= 3kB time= > 0.010 br= 34.8kbits/s avg_br= 2325.6kbits/s type= P > out= 0 st= 0 frame= 3 q= 0.2 f_size= 61 s_size= 3kB time= > 0.010 br= 11.7kbits/s avg_br= 2374.4kbits/s type= B > out= 0 st= 0 frame= 4 q= 0.3 f_size= 27 s_size= 3kB time= > 0.042 br= 5.2kbits/s avg_br= 573.9kbits/s type= B without it. _______________________________________________ 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".
