[FFmpeg-devel] [PATCH] av_get_delay

Nicolas George nicolas.george at normalesup.org
Wed Jun 29 12:25:26 CEST 2011


Le decadi 10 messidor, an CCXIX, Michael Niedermayer a écrit :
> with the current API audio/video output "muxers" should "present"
> the data to the user at the pts of the given packet.

You mean: according to the system's realtime clock rather than some
device-specific internal clock?

I do not think this is a good idea. To account for the skew in a sound
card's clock, it would require slightly resampling the incoming data, while
most application would be just happy to just sync on the sound card's clock.

> Maybe something along the following lines is clearer:
> the delay is the duration in timebase units of all buffered data.
> buffered in the sense of between what goes into the muxer and what
> is presented to the user or transmitted over the net.

I adapted it that way:

 * The delay is the duration, in timebase units, of all buffered data,
 * i.e. data that has been submitted to the muxer but not yet presented to
 * the user of transmitted over the net.
 * The exact meaning of that depends on the muxer.
 * It is mostly relevant for live streams, like devices or network
 * protocols.

> some value that is not a valid delay would allow distinguishing it from
> valid delays. some error code representing "Not implemented" maybe

Would simply AV_NOPTS_VALUE be ok?

> also the code should consider the buffering required for interleaving
> (by dts for example)

You mean packets that have been submitted to av_interleaved_write_frame but
that lavf has not yet submitted to oformat->write_packet?

I see the concern, and I will think about it to try and come with a
satisfactory solution.

Regards,

-- 
  Nicolas George
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20110629/0c6ab619/attachment.asc>


More information about the ffmpeg-devel mailing list