[MPlayer-dev-eng] Re: [PATCH] fix audio detection in libmpdemux/demux_ts.c

Andrew Cohen cohen at andy.bu.edu
Sun Jan 28 18:11:56 CET 2007


>>>>> "Nico" == Nico Sabbi <nicola_sabbi at fastwebnet.it> writes:

    Nico> Andrew Cohen wrote:
    >> The recent modification of demux_ts.c to use ts_add_stream for
    >> unified creation of sh_audio and sh_video fails to correctly
    >> detect the audio on some dvb streams. This one-line patch seems
    >> to fix the problem.
    >> 
    >> 
    >> Index: libmpdemux/demux_ts.c
    >> ===================================================================
    >> --- libmpdemux/demux_ts.c (revision 22035) +++
    >> libmpdemux/demux_ts.c (working copy) @@ -293,7 +293,7 @@
    >> if(priv->ts.streams[es->pid].sh) return; -
    >> if((IS_AUDIO(es->type) || IS_AUDIO(es->subtype)) &&
    priv-> last_aid+1 < MAX_A_STREAMS)
    >> + if(((IS_AUDIO(es->type) || (es->type==SL_PES_STREAM &&
    >> IS_AUDIO(es->subtype)) || (es->type == PES_PRIVATE1)) ||
    >> IS_AUDIO(es->subtype)) && priv->last_aid+1 < MAX_A_STREAMS) {
    >> sh_audio_t *sh = new_sh_audio_aid(demuxer, priv->last_aid,
    >> es->pid); if(sh)
    >> 

    Nico> an undetected ac3 stream? If so what was previously PES_PRIVATE1 is
    Nico> supposed to have been changed to AUDIO_A52, but it seems
    Nico> it's not done, yet.

Yes, I guess so. These streams still report as PES_PRIVATE1

    Nico> Can you retry without your patch and with a larger value for
    Nico> -tsprobe, please?

OK, I've tried with many different values larger than
TS_MAX_PROBE_SIZE (as large as 100 times) and see no difference. Still
reports the stream as PES_PRIVATE1.

-- 
Andrew Cohen		Office:	(617) 353-6051
Physics Department	Fax:	(617) 353-9393
Boston University	Email:	cohen at bu.edu
Boston, MA 02215





More information about the MPlayer-dev-eng mailing list