On 02/26/2012 05:32 PM, Martin Storsjö wrote:
> On Sun, 26 Feb 2012, Justin Ruggles wrote:
>
>> ---
>> libavformat/movenc.c | 2 +-
>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/libavformat/movenc.c b/libavformat/movenc.c
>> index e2781ac..5d9b2d7 100644
>> --- a/libavformat/movenc.c
>> +++ b/libavformat/movenc.c
>> @@ -2215,7 +2215,7 @@ static int mov_write_tfhd_tag(AVIOContext *pb,
>> MOVTrack *track,
>> if (flags & MOV_TFHD_BASE_DATA_OFFSET)
>> avio_wb64(pb, moof_offset);
>> if (flags & MOV_TFHD_DEFAULT_DURATION) {
>> - track->default_duration = track->audio_vbr ? track->enc->frame_size
>> : 1;
>> + track->default_duration = track->audio_vbr ?
>> get_cluster_duration(track, 0) : 1;
>> avio_wb32(pb, track->default_duration);
>> }
>> if (flags & MOV_TFHD_DEFAULT_SIZE) {
>> --
>> 1.7.1
>
> You could remove the audio_vbr?: stuff altogether - the default duration 1
> for other formats is no better than the duration of the first packet
> either. For CFR with a time base identical to the frame rate, it normally
> is 1, but for other time base setups, the duration can be something else
> than 1, and then the first packet duration is a better guess.
Ok, I tested this in QT. Even for pcm audio, the packet duration here is
supposed to be the number of samples, not 1. I'll send a new patch that
always uses the first cluster duration for the default. With the current
code, we write the duration of each packet in 'trun', so it still works
but it has unnecessary overhead.
Thanks,
Justin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
