On 09/04/14 00:04, Vittorio Giovara wrote: > On Thu, Apr 3, 2014 at 5:48 PM, Luca Barbato <[email protected]> wrote: >> On 03/04/14 17:03, Vittorio Giovara wrote: >>> From: Michael Niedermayer <[email protected]> >>> >>> --- >>> libavformat/riff.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/libavformat/riff.c b/libavformat/riff.c >>> index f6787ea..0a9e081 100644 >>> --- a/libavformat/riff.c >>> +++ b/libavformat/riff.c >>> @@ -39,6 +39,7 @@ const AVCodecTag ff_codec_bmp_tags[] = { >>> { AV_CODEC_ID_H264, MKTAG('V', '2', '6', '4') }, /* CCTV >>> recordings */ >>> { AV_CODEC_ID_H264, MKTAG('G', 'A', 'V', 'C') }, /* GeoVision >>> camera */ >>> { AV_CODEC_ID_H264, MKTAG('U', 'M', 'S', 'V') }, >>> + { AV_CODEC_ID_H264, MKTAG('I', 'N', 'M', 'C') }, >>> { AV_CODEC_ID_H263, MKTAG('H', '2', '6', '3') }, >>> { AV_CODEC_ID_H263, MKTAG('X', '2', '6', '3') }, >>> { AV_CODEC_ID_H263, MKTAG('T', '2', '6', '3') }, >>> @@ -85,7 +86,6 @@ const AVCodecTag ff_codec_bmp_tags[] = { >>> { AV_CODEC_ID_MPEG4, MKTAG('H', 'D', 'X', '4') }, >>> { AV_CODEC_ID_MPEG4, MKTAG('D', 'M', 'K', '2') }, >>> { AV_CODEC_ID_MPEG4, MKTAG('D', 'I', 'G', 'I') }, >>> - { AV_CODEC_ID_MPEG4, MKTAG('I', 'N', 'M', 'C') }, >>> /* Ephv MPEG-4 */ >>> { AV_CODEC_ID_MPEG4, MKTAG('E', 'P', 'H', 'V') }, >>> { AV_CODEC_ID_MPEG4, MKTAG('E', 'M', '4', 'A') }, >>> >> >> Looks fishy, is it MPEG4 or H264? should we try to decode its content to >> figure out which is which ? > > I have a sample that works if INMC is mapped to h264, none working > when INMC is on mpeg4. > The only mention of a mpeg4 INMC sample is here > http://spreadys.wordpress.com/2013/07/27/inmc-codec-update/ and the > author is unresponsive of my requests. > So, what should we do? >
The proper way would be trying to have a look at the extradata if present and match the right codec there. Or we can just write down a note and tell the hypothetical user to force the coded in that specific case. I'm for the latter. lu _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
