[FFmpeg-devel] [PATCH 2/6] avformat/s337m: Consider container bit resolution

Tomas Härdin git at haerdin.se
Tue Feb 21 11:47:25 EET 2023


fre 2023-02-17 klockan 09:44 +0000 skrev Nicolas Gaullier:
> > > @@ -56,6 +56,12 @@ static int s337m_get_offset_and_codec(void
> > > *avc,
> > >              avpriv_report_missing_feature(avc, "Data type %#x in
> > > SMPTE 337M", data_type & 0x1F);
> > >          return AVERROR_PATCHWELCOME;
> > >      }
> > > +    if (container_word_bits &&
> > > +        !(container_word_bits == 16 && word_bits == 16) &&
> > > +        !(container_word_bits == 24 && word_bits == 20) &&
> > 
> > I presume 20/24 is intentional. Does WAV not support signalling 20-
> > bit?
> > 
> > The rest looks OK enough
> > 
> > /Tomas
> 
> You are true, I meant "container_word_bits" as "block_size" rather
> than "valid bits per sample" and
> I think this should be clarified is the latter integration code in
> "wavdec: s337m support" patch where
> I use par->bits_per_coded_sample...
> But I should have rather coded "AV_CODEC_ID_PCM_S16LE ? 16 : 24"
> This is in case a wav file would be detected as 20 bits in a 24 bits
> container (I don't think it is supported yet, but it could as
> bitspersample and validbitpersample are two different fields in
> WAVE_FORMAT_EXTENSIBLE).
> Are you ok with this ?

I haven't worked enough with S377m to really know, but I do know it's a
mess. Is there a way to differentiate "regular" packed 20-bit audio
from S377m in wav?

/Tomas



More information about the ffmpeg-devel mailing list