Hi Tom,

On Sat, 2008-02-16 at 18:21 -0800, Tom Jennings wrote:
> Not absolutely certain, as there is a "random" element that
> means unknown or hard to produce trigger, but switching from
> the generic VESA driver to i965 chipset driver seems to have
> stopped the interface lockup. Odd, as overall the display is
> fine, all windows except mixxx are perfect. It's got to be a
> library function write-to-window-something deadlocking. That's
> a relief.  I'll continue trying to tune this.
> 
> 
> There is still a very noticable and repeatable pause in audio
> output when hitting CUE while a player is playing, eg. jumping
> back to a set cue point. The delay seems congruous with the
> waveform display generation -- a large jump in time requires
> a full erase/rewrite.
> 
> It's quite repeatable, if I hit CUE... CUE... CUE... once per
> second or so, there's an audio pause/gap and full waveform
> rewrite every time.

I get this too on my computer, but as Robin said, it only happens once
every 10 CUEs. I've found that lowering my latency lessens the chance
that this will happen because it looks like the waveform is "redrawing"
faster. (By redrawing, I mean that rendering it does where it builds
itself from the centre outwards in each direction.) 

I always thought the pause in the audio was an xrun caused by the spike
in CPU usage associated with seeking (and redrawing the waveform at the
new position), and it could be some combination of that and another
problem, or just another problem altogether. 

I've never looked at fixing this because none of us understands how the
waveform view works. The code is _very_ messy, and needs a rewrite...

> 
> It's a 2.4GHz dual core intel with 4gb of 1066MHz RAM, running
> only mixxx and an xterm or two. ALSA driver from mixxx,
> not jackd.

If you compile with the optimize=2 (Intel P4), optimize=3 (Intel Core),
or optimize=4 (Intel Core 2) flag appropriately, you'll should get a
respectable speed boost, which might help you lower your latency to
battle this problem.

Thanks,
Albert




-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Mixxx-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mixxx-devel

Reply via email to