[FFmpeg-user] '-bsf setts=dts' has problems

[Paul B Mahol]

On Sat, Jun 8, 2024 at 8:36 AM Mark Filipak <markfilipak.imdb at gmail.com>
wrote:

> '-bsf setts=pts' seems to always work, but '-bsf setts=dts' has problems.
> Look:
>

It is evident that you never looked at actual setts documentation.

# ffmpeg -h bsf=setts
ffmpeg version N-115950-g1f63f014ff Copyright (c) 2000-2024 the FFmpeg
developers
  built with clang version 17.0.6
  configuration: --cc='ccache clang' --as='ccache clang' --disable-indevs
--disable-outdevs --disable-autodetect --enable-sdl2 --enable-gpl
--enable-indev=lavfi --enable-indev=alsa --enable-outdev=alsa
--enable-nonfree --enable-zlib --enable-openssl --disable-static
--enable-shared --enable-clap --enable-ladspa --enable-libflite
--enable-lv2 --enable-version3 --enable-libmp3lame --enable-libmysofa
--enable-libx264 --enable-fontconfig --enable-libfreetype --enable-libass
--enable-librubberband --enable-libzimg --enable-libharfbuzz
--disable-stripping --disable-htmlpages --enable-libsoxr
--samples=../fate-suite
  libavutil      59. 21.100 / 59. 21.100
  libavcodec     61.  6.100 / 61.  6.100
  libavformat    61.  3.104 / 61.  3.104
  libavdevice    61.  2.100 / 61.  2.100
  libavfilter    10.  2.102 / 10.  2.102
  libswscale      8.  2.100 /  8.  2.100
  libswresample   5.  2.100 /  5.  2.100
Bit stream filter setts
setts_bsf AVOptions:
  -ts                <string>     ...VAS..B.. set expression for packet PTS
and DTS (default "TS")
  -pts               <string>     ...VAS..B.. set expression for packet PTS
  -dts               <string>     ...VAS..B.. set expression for packet DTS
  -duration          <string>     ...VAS..B.. set expression for packet
duration (default "DURATION")
  -time_base         <rational>   ...VAS..B.. set output timebase (from 0
to INT_MAX) (default 0/1)


>
> C:\Windows\System32>ffmpeg -copyts -i "g:\FANNY AND ALEXANDER
> [1982(1983)]\00305.m2ts" -map 0
> -bsf noise=drop='lt(pts\,2854113)'
> -c copy -sn -dn -muxdelay 0 "g:\FANNY AND ALEXANDER [1982(1983)]\foo.m2ts"
> 0,    2846606,    2854113,     3753,   640646, 0x3a5a0c45, S=1,        1
> 0,    2854113,    2861621,     3753,   640646, 0x02208727, F=0x0, S=1,
>     1
>
> C:\Windows\System32>ffmpeg -copyts -i "g:\FANNY AND ALEXANDER
> [1982(1983)]\00305.m2ts" -map 0
> -bsf
> noise=drop='lt(pts\,2854113)',setts=dts='if(eq(DTS\,2846606)\,2850360\,DTS)'
> -c copy -sn -dn -muxdelay 0 "g:\FANNY AND ALEXANDER [1982(1983)]\foo.m2ts"
> The command provokes this: "Invalid DTS: 2850360 PTS: 2846606, replacing
> by guess"
> 0,    2846606,    2846606,     3753,   640646, 0x3a5a0c45, S=1,        1
> 0,    2854113,    2854113,     3753,   640646, 0x02208727, F=0x0, S=1,
>     1
> There is nothing invalid with DTS: 2850360 if PTS is 2854113, but FFmpeg
> appears to change PTS on
> it's own, without a command to do it.
>
> The second command should not change PTSes, but it does. In fact, all
> frame PTSes are being set to
> their DTSes.
>
> I have documented a litany of DTS problems now. Are they being taken
> seriously?
>
> This is unrelated to my investigation of splices and my search for a
> workaround because of FFmpeg's
> inability to set MPEG tag 'closed_gop' to '1'. Or maybe it is related, I
> have no way to know.
>
> Will someone please work with me so I can share the source with them so
> they can verify what I'm
> seeing? I cannot continue to work alone on this. It's just too nerve
> wracking.
>

Nobody can take you and your mails seriously.


> --Mark.
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-user
>
> To unsubscribe, visit link above, or email
> ffmpeg-user-request at ffmpeg.org with subject "unsubscribe".
>