On Thu, 2007-09-13 at 09:43 +0200, Clemens Ladisch wrote: > Jonathan Stowe wrote: > > I'm getting some strange behaviour between LinuxSampler, JACK and the > > USB audio driver and I'm not quite sure where to look to narrow it down > > further. > > > > What *appears* to be happening is that when LinuxSampler is being output > > to JACK and JACK is using a usb audio device (with or without an > > intervening client) then there is a weird "arpeggiation" of the notes > > being played which sounds like something is messing with the events in > > the ALSA sequencer, the additional notes are the fifth and octave of the > > original note it sounds like though I haven't tried it right across the > > keyboard, ( this happens whether the MIDI input is from e.g. a USB > > keyboard or aplaymidi.) Other clients (e.g. AlsaModularSynth ) in > > exactly the same configuration play fine as well as does ardour and so > > forth. LinuxSampler outputting to JACK when using the internal > > (hda-intel) audio does not exhibit this behaviour. LinuxSampler > > outputting directly to the usb-audio doesn't do it either. > > So it's _only_ the combination LinuxSampler+Jack+USB? > Well I thought so, but two months of holiday appear to have sent my diagnostic skills down the toilet. I've absolved the USB driver as I can replicate this with the hda-intel as well now. > I don't know how LinuxSampler could change its behaviour depending on > Jacks output driver. The only possibility is that using another driver > changes the period size, i.e., Jack's buffer size. > > What is the exact configuration when you start jackd with either driver? I got obsessed with the "signal path" rather than the jack configuration, I was starting jackd with jackd -p128 -dalsa -dhw:1,0 -r44100 -p4096 -n2 -i2 -o2 -zs for the USB (which works fine for everything else) but because the hda-intel performs so badly it is unuseable with jackd I was simply starting it with jackd -d alsa (I haven't yet found any configuration that doesn't give rise to an unacceptable number of xruns,) however on starting it with the same command line as for the USB (substituting hw:0,0 for the hw:1,0 obviously) I got the same behaviour. Changing the period to 2048 fixes the problem as you suggest, but I'm not quite sure why and who if at all to report the "bug" to - does a jack client need to take into account the buffer size? Anyway not an ALSA problem. Thanks a lot for your help. /J\ ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Alsa-user mailing list Alsa-user@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/alsa-user
