tis 2019-08-06 klockan 13:09 +0000 skrev Gaullier Nicolas: > > > +FATE_SAMPLES_AUDIO-$(call DEMDEMDEC, WAV, S337M, DOLBY_E) += > > > +fate-dolby-e-wav > > > +fate-dolby-e-wav: CMD = dolbye2pcm16 -i > > > +$(TARGET_SAMPLES)/dolby_e/512.wav > > > +fate-dolby-e-wav: CMP = oneoff > > > +fate-dolby-e-wav: REF = $(SAMPLES)/dolby_e/512.wav.pcm > > > > This does a floaty comparison, right? Since the Dolby-E decoder is > > float at the moment.. > The test is similar to the current "fate-dolby-e", this new one is > only different because the input dolby E stream is 24 bits (instead > of 16), contains 5.1+2 (instead of 5.1)... and uses wav wrapper of > course. > In both case, the comparison is based on the truncated 16-bit output > of the Dolby-E decoder (but I think the inner precision of the codec > is higher) AND a tolerance of +/-1.
There being a tolerance is what I was getting at. Some tests expect bitexact output, which is obviously problematic with a float decoder :) /Tomas _______________________________________________ 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".
