[FFmpeg-soc] [soc]: r4240 - wmapro/wma3dec.c

[Sascha Sommer]

Hi,

On Dienstag, 28. April 2009, Benjamin Larsson wrote:
> Sascha Sommer wrote:
> > Hi,
> >
> > On Sonntag, 26. April 2009, Benjamin Larsson wrote:
> >> banan wrote:
> >>> Author: banan
> >>> Date: Sun Apr 26 22:53:28 2009
> >>> New Revision: 4240
> >>>
> >>> Log:
> >>> Channel layout matching 8_Channel_ID.wma
> >>>
> >>> Modified:
> >>>    wmapro/wma3dec.c
> >>>
> >>> Modified: wmapro/wma3dec.c
> >>> =======================================================================
> >>>== ===== --- wmapro/wma3dec.c	Sun Apr 26 22:36:20 2009	(r4239)
> >>> +++ wmapro/wma3dec.c	Sun Apr 26 22:53:28 2009	(r4240)
> >>> @@ -390,6 +390,7 @@ static av_cold int wma3_decode_init(AVCo
> >>>      }
> >>>
> >>>      dump_context(s);
> >>> +    avctx->channel_layout = channel_mask;
> >>>      return 0;
> >>>  }
> >>
> >> Looks like 5.1 has a non dwChannelMask layout. Will fix that later. T2
> >> sample from WMV HD Content Showcase decode with lots of jumps.
> >
> > Did you try the 5.1 stream without my manual channel reordering?
> > Unfortunatelly I do not know what is wrong with the T2 sample.
> >
> > Regards
> >
> > Sascha
>
> Ahh, didn't see that. Ok, then I guess it's final. Wmapro uses
> dwChannelMask, which makes it easy to support multichannel files. I'll
> fix up the code later. Regarding T2 it might be a demuxer problem. Can
> you try with the binary code and mplayer ?
>

It is not a demuxer problem. It is one of the cases where certain binary 
decoders also have their problems. This is probably caused by wrong handling 
of the 24 bits sample depth information. I commited a workaround that should 
improve the playback a bit.

Regards

Sascha