On 02/26/2012 04:47 PM, Martin Storsjö wrote:
> On Sun, 26 Feb 2012, Justin Ruggles wrote:
>
>> For encoding, frame_size is the number of input samples to send to the
>> encoder and is not necessarily the same as the duration of all output
>> packets.
>> ---
>> libavformat/movenc.c | 56 ++++++++++++++++++++++++++++++++++---------------
>> 1 files changed, 39 insertions(+), 17 deletions(-)
>
> Looks good in general, but IMO needs a bit of splitting.
ok, i'll split it.
>> @@ -1107,9 +1141,7 @@ static int mov_write_stts_tag(AVIOContext *pb,
>> MOVTrack *track)
>> av_malloc(track->entry * sizeof(*stts_entries)) : /*
>> worst case */
>> NULL;
>> for (i=0; i<track->entry; i++) {
>> - int64_t duration = i + 1 == track->entry ?
>> - track->track_duration - track->cluster[i].dts +
>> track->start_dts : /* readjusting */
>> - track->cluster[i+1].dts - track->cluster[i].dts;
>> + int duration = get_cluster_duration(track, i);
>> if (i && duration == stts_entries[entries].duration) {
>> stts_entries[entries].count++; /* compress */
>> } else {
>
> This changes the data type of duration fron int64_t to int, but I guess it
> shouldn't make any difference (int makes more sense).
I suppose the function could do with a sanity check for the dts
difference being between 0 and INT_MAX.
-Justin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
