Re: [ecasound] Consistent buffer overrun problem (?)

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [ecasound] Consistent buffer overrun problem (?)
From: Arno van Amersfoort (arnova_AT_eld.physics.leidenuniv.nl)
Date: Thu Feb 26 2004 - 21:45:31 EET

Ok. I already had a feeling that I had something to do with that. But what I don't understand is why does it get better when I increase the buffersize? I'd expect that changing the buffersize shouldn't make a difference then...

Kai Vehmanen wrote:
On Thu, 26 Feb 2004, Arno van Amersfoort wrote:

  
I've set up a Fedora/Planet CCRMA system to record with 4 soundcards (8 
channels) simultaniously. And yes, I know, everybody recommends not to 
do this because of sync problems but I'm kinda stoburn ;-). Anyway, I 
    
[...]
  
seems to work a little now. I can record about 2m45 without any problems 
or error messages but after that  ecasound starts to fill my screen with 
"warning! playback overrun - samples lost!" messages.
    
[...]
  
minutes). I don't really understand what the real cause is. Can somebody 
    

This is caused by the very reason why people recommend against using 
multiple cards: clock drift.

Although all cards nominally run at the same speed (for example 44100 or 
48000 samples a second), in practise they each run at slightly different 
speeds. Unfortunately Ecasound (and other apps) only know the nominal 
speed, so they cannot easily compensate for the speed difference.

Now what happens is that sooner or later, the soundcard with the fastest 
clock will run out of data and you'll get xruns errors. Similarly a slow 
client will block the engine as it takes longer than other to process the 
given samples.

In theory you could try to measure the clock differences in software and
try to compensate, but in practise it is easier to just buy a real
multichannel card.

  

-- 
---------------------------------------------------------------------------
ing. Arno van Amersfoort
Department Of Electronics (ELD, k1007)
Huygens Laboratory
Leiden University
P.O. Box 9504
Niels Bohrweg 2
2333 CA Leiden
The Netherlands
---------------------------------------------------------------------------
Phone     : +31-(0)71-5275959
Fax       : +31-(0)71-5275819
---------------------------------------------------------------------------
E-mail    : a.c.j.van.amersfoort_AT_eld.physics.leidenuniv.nl
---------------------------------------------------------------------------
Arno's (Linux IPTABLES Firewall) Homepage:
http://rocky.molphys.leidenuniv.nl

New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Thu Feb 26 2004 - 21:42:06 EET