Clément Bœsch: > On Mon, Aug 10, 2020 at 08:36:13PM +0200, Andreas Rheinhardt wrote: >> Clément Bœsch: >>> Source: Videolan sample database (samples/sub/dvbsub) >>> --- >>> Can someone upload >>> https://streams.videolan.org/samples/sub/dvbsub/dvbsubtest.ts to >>> fate-samples/sub? >> >> A 20MB sample? Can't one use a smaller sample or filter out the >> interesting PIDs from this sample? (If I remux only the subtitle stream >> to Matroska with mkvmerge, the sample is 33 kB.) > > Yup, you're correct. New patch attached. > > For the sample: > curl 'https://0x0.st/i35H.mks' -o fate-samples/sub/dvbsubtest.mks > > sha256sum: > 3fccc784eb61e701f1d6789253f637a32374ce6c81539d44a4e74075246ff02b > fate-samples/sub/dvbsubtest.mks > If one strips the video and the audio PID away, one gets a 77KB transport stream. And as it turns out, this file is unusable since commit a1b4f120c031e6697bac9fd8c725d9c37ee36d13; see ticket #8844 for details. Maybe it would be better to go with the stripped down transport stream (the file I attached to #8844) in order to make sure that we don't regress again (after the regression has been fixed, of course)?
But anyway, the Matroska file uses zlib compression (that's why it's so small -- the actual data size is 51KB) and therefore your test needs to check for zlib. - Andreas _______________________________________________ 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".
