[vlc-devel] Re: VLC audio output FIFO questions ...

Derk-Jan Hartman d.hartman at student.utwente.nl
Wed May 14 21:27:41 CEST 2003


Hi there,

This is a known issue
http://bugzilla.videolan.org/cgi-bin/bugzilla/show_bug.cgi?id=115

I know fenrir solved it for some other stuff, it might be worth looking 
into CVS logs of those.
Fenrir himself might be able to give you some more help too.
He's often in #videolan on Freenode IRC.

DJ

On woensdag, mei 14, 2003, at 20:42 Europe/Amsterdam, j.zorko at att.net 
wrote:
> Hello, all ...
>
> In my explorations of the VLC source, as well as adding functionality 
> and observing how things
> work, i've now come to the audio output stuff.  There is some behavior 
> i'm seeing that i'm trying to
> determine the cause of, and while i've a theory (and started looking 
> at the audio output source), I
> wanted to ask the people here, as well as perhaps get some additional 
> pointers on where to look.
>
> The behavior I see is this: I've added functionality to VLC to make it 
> do fast-forward and rewind,
> and I implemented this by seeking forward or backward in the file 
> (either purely on a temporal
> basis, or looking for the nth occurence of a sequence header / MP3 
> frame and setting the new file
> position there).  This works well with the audio and video sources 
> i've tested (MP3, MPEG2 VOBs).
> The idea is that, when VLC is acting as a server to other VLCs, this 
> is sort of like RTSP i.e. a control
> protocol -- some external process can send the ff / rewind commands to 
> the server VLC, and the
> clients receiving the stream don't need to know anything.
>
> However, after several ff / rewind commands, what i'm observing is 
> that, with MP3 audio, the
> audio output 'wobbles' i.e. gets faster, then slower, then finally 
> settling in.  Since i'm not altering
> the number of packets the input thread sends to the FIFO (only where 
> those packets are i.e.
> seeking forward or backward in the file before reading the packets), 
> one can infer from this that
> some FIFO is based on time in the buffer, not # of packets in the 
> buffer.  However, so far i've not
> found this code, and i'm frankly a bit confused (the VLC code seems to 
> have FIFOs everywhere, i'm
> not sure which one is the one that can result in wobbly audio).
>
> So -- can someone give me some pointers, where to look, whether this 
> mystical, magical FIFO is
> indeed based on the spacing of packets in time vs. the quantity of 
> packets in it?  Which FIFO is
> this?  Where is it?
>
> Regards,
> John, VLC hacker (in training) <smile>
---
Videolan - VLC media player
Derk-Jan Hartman (thedj at users.sourceforge.net)
Co-Developer of the MacOS X port of vlc
http://www.videolan.org/vlc

-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://developers.videolan.org/lists.html
If you are in trouble, please contact <postmaster at videolan.org>



More information about the vlc-devel mailing list