why audio is used as a reference in AV sync.

[amitpandya]

Hi all,

I have a question regarding AV synchronization.

Normally when we deal with both audio+video data processing in gstreamer,
we takes audio as a reference for AV synchronization and not the video frames.

Why is it so?
What is the advantage of doing so?
Why shouldn't we tak video frames as our reference while synchronization at muxer?  

Thanks,
Amit Pandya.

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

[Martin Bisson]

It's generally because a skip in audio playback is much more noticeable
than skipping (or repeating) a frame.  Also, it is possible to use the
sound card as a clock generator for synchronization based on the
supposed playrate (might not be EXACTLY 44100 Hz...  hence the use of
the card as a clock generator) and the time taken to "consume" a number
of bytes.

Martin


[hidden email] wrote:

> Hi all,
>
> I have a question regarding AV synchronization.
>
> Normally when we deal with both audio+video data processing in gstreamer,
> we takes audio as a reference for AV synchronization and not the video frames.
>
> Why is it so?
> What is the advantage of doing so?
> Why shouldn't we tak video frames as our reference while synchronization at muxer?  
>
> Thanks,
> Amit Pandya.
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev 
> _______________________________________________
> gstreamer-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
>
>  

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

[Robert Powell-2]

[hidden email] wrote:
> Hi all,
>
> I have a question regarding AV synchronization.
>
> Normally when we deal with both audio+video data processing in gstreamer,
> we takes audio as a reference for AV synchronization and not the video frames.
>
> Why is it so?
> What is the advantage of doing so?
> Why shouldn't we tak video frames as our reference while synchronization at muxer?
>

Suppose you were to use video at 50fps with audio sampled at 48k.  For the first frame you have one frame and 960 audio samples.  How will you space the 960 samples out evenly if you only have one clock tick?

Hope that helps

Rob

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel