On Wed, 25 Jan 2012, Martin Storsjö wrote:
On Wed, 25 Jan 2012, Justin Ruggles wrote:
On 01/25/2012 01:10 PM, Alex Converse wrote:
On Wed, Jan 25, 2012 at 7:14 AM, Justin Ruggles
<justin.rugg...@gmail.com> wrote:
On 01/25/2012 05:37 AM, Martin Storsjö wrote:
@@ -762,9 +773,13 @@ static int mov_find_codec_tag(AVFormatContext *s,
MOVTrack *track)
{
int tag = track->enc->codec_tag;
- if (track->mode == MODE_MP4 || track->mode == MODE_PSP ||
track->mode == MODE_ISM)
+ if (track->mode == MODE_MP4 || track->mode == MODE_PSP)
+ tag = mp4_get_codec_tag(s, track);
+ else if (track->mode == MODE_ISM) {
tag = mp4_get_codec_tag(s, track);
- else if (track->mode == MODE_IPOD)
+ if (!tag && track->enc->codec_id == CODEC_ID_WMAPRO)
+ tag = MKTAG('w', 'm', 'a', ' ');
why is this separate fallback needed? does wmapro in mp4 not have an
object type that could be added to ff_mp4_obj_type?
WMAPro does not have an assigned objectTypeIndication. Probably
because objectTypeIndication is only relevant inside an elementary
stream descriptor.
Then maybe mp4_get_codec_tag() shouldn't return right away if the object
type is not found. Then the wma tag could be added to the existing
series of checks there.
I'm ok with changing it that way too, if you'd prefer that. What do others
think?
Are there sufficient checks elsewhere that makes sure it doesn't silently try
to mux any unsupported codec into mp4?
Justin, Alex - what's your opinion on this? Is the code simplification
worth risking muxing unsupported codecs into mp4? (Also, should it be ok
to mux wmapro into a normal mp4 file?)
// Martin
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
