Andreas Rheinhardt:
> Paul B Mahol:
>> Hi,
>>
>> patch attached.
>>
>> return ret;
>> }
>>
>> + pkt->duration = av_rescale_q(pict->duration, (AVRational){ 1, 100 },
>> + avctx->time_base);
>> pkt->size = outbuf_ptr - pkt->data;
>> if (s->image || !avctx->frame_number)
>> pkt->flags |= AV_PKT_FLAG_KEY;
>
> Where does the magic 1/100 timebase come from? For encoding, the
> AVFrame's time base is supposed to be AVCodecContext.time_base. This
> could actually be done generically for all video-encoders without delay.
> (I know that AVFrame has a time_base of its own that is currently unset
> and unused.)
>
Strangely, if I do exactly as above, the gifenc tests show weird
durations like:
+0, 0, 0, 10, 1368, 0x6cf0befd
+0, 1, 1, 10, 158, 0xcd173bb4, F=0x0
+0, 2, 2, 10, 163, 0x4f7a451d, F=0x0
+0, 3, 3, 10, 152, 0x17723839, F=0x0
+0, 4, 4, 10, 160, 0x67854056, F=0x0
So somewhere a timebase is wrong or AVFrame.duration is not converted to
a new timebase somewhere.
- 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".
