Ronald S. Bultje:
> Hi,
>
> On Tue, Mar 2, 2021 at 11:20 AM Andreas Rheinhardt <
> andreas.rheinha...@gmail.com> wrote:
>
>> Ronald S. Bultje:
>>> ---
>>> libavformat/ivfenc.c | 11 ++++++++---
>>> 1 file changed, 8 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/libavformat/ivfenc.c b/libavformat/ivfenc.c
>>> index 0951f56..e22625b 100644
>>> --- a/libavformat/ivfenc.c
>>> +++ b/libavformat/ivfenc.c
>>> @@ -23,7 +23,7 @@
>>>
>>> typedef struct IVFEncContext {
>>> unsigned frame_cnt;
>>> - uint64_t last_pts, sum_delta_pts;
>>> + uint64_t last_pts, sum_delta_pts, first_duration;
>>> } IVFEncContext;
>>>
>>> static int ivf_init(AVFormatContext *s)
>>> @@ -86,6 +86,8 @@ static int ivf_write_packet(AVFormatContext *s,
>> AVPacket *pkt)
>>> avio_write(pb, pkt->data, pkt->size);
>>> if (ctx->frame_cnt)
>>> ctx->sum_delta_pts += pkt->pts - ctx->last_pts;
>>> + else
>>> + ctx->first_duration = pkt->duration;
>>> ctx->frame_cnt++;
>>> ctx->last_pts = pkt->pts;
>>>
>>> @@ -97,12 +99,15 @@ static int ivf_write_trailer(AVFormatContext *s)
>>> AVIOContext *pb = s->pb;
>>> IVFEncContext *ctx = s->priv_data;
>>>
>>> - if ((pb->seekable & AVIO_SEEKABLE_NORMAL) && ctx->frame_cnt > 1) {
>>> + if ((pb->seekable & AVIO_SEEKABLE_NORMAL) &&
>>> + (ctx->frame_cnt > 1 || (ctx->frame_cnt == 1 &&
>> ctx->first_duration))) {
>>> int64_t end = avio_tell(pb);
>>>
>>> avio_seek(pb, 24, SEEK_SET);
>>> // overwrite the "length" field (duration)
>>> - avio_wl32(pb, ctx->frame_cnt * ctx->sum_delta_pts /
>> (ctx->frame_cnt - 1));
>>> + avio_wl32(pb, ctx->frame_cnt > 1 ?
>>> + ctx->frame_cnt * ctx->sum_delta_pts / (ctx->frame_cnt
>> - 1) :
>>> + ctx->first_duration);
>>> avio_wl32(pb, 0); // zero out unused bytes
>>> avio_seek(pb, end, SEEK_SET);
>>> }
>>>
>> Shouldn't we not always use the duration of the last packet for the
>> duration and not only when it is the only packet?
>
>
> That assumes the duration is always set, which according to the
> documentation is not guaranteed.
>
Then why don't you use it if it is set?
- Andreas
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
