[FFmpeg-devel] About guess_correct_pts / AVFrame.best_effort_timestamp

Alexander Strange astrange
Fri Feb 18 05:53:32 CET 2011


On Feb 17, 2011, at 11:39 PM, Luca Barbato wrote:

> On 02/18/2011 01:43 AM, Baptiste Coudurier wrote:
>> On 02/17/2011 09:27 AM, Ronald S. Bultje wrote:
>>> Hi,
>>> 
>>> [...]
>>> 
>>> 
>>> But like I said, my motivation for this subject is gone. The patch is
>>> rejected, I've explained why, and that's all I'll say about it.
>> 
>> Nice to see how the dictatorship only switch heads.
>> Good job guys.
> 
> Mind reading the whole threads and figure out why ronald (and just him)
> decided to quit discussing?
> 
> We said that the patch as is, is wrong. We told why is wrong, we told
> how to get the results in an acceptable way and we explained the long
> run plan that would lead to a possibly cleaner result w/out this number
> of side effects and corner cases.

What side effects?
I saw nothing in this thread except questions about how libavformat works, all concrete proposals for client code were wrong.

I think more valid times _could_ be made to work in libavformat, with better AVParsers that don't quite exist, but that would leave something like this:
avcodec_decode_video(..)
pts = avctx->pkt_pts ? avctx->pkt_pts : avctx->pkt_dts;

Even this one line deserves to be moved into lavc, since the fact that ffplay/ffmpeg were wrong for 5+ years means nobody can be expected to get it right.



More information about the ffmpeg-devel mailing list