Hi list, replying to my own mail from Jan 1st - > today I have found some time to play with Linuxsampler, gigedit and others, > and noticed 2 things that I did not understand yet: > > Software used: > - Ubuntu Intrepid (8.10) on AMD64 > - kernel 2.6.27-9 > - JACK 0.116.1 > - Linuxsampler, libgig, liblscp, qsampler from CVS 20081203 > > 1) > I created a simple .gig in gigedit, loaded it into QSampler, connected > everything in QJackCtl (external MIDI keyboard, coming in through a Yamaha > USB MIDI interface) and started playing. Everything is fine for a while, but > at some point (seems to be correlated to event density) sound output stops. > After 2-4 seconds, the queued notes are played out simultaneously (and, due > to that, this produces a rather loud sound mix). Then normal playback is > possible for a while, and then the same thing happens again. > > This sounds similar to this report from the BB: > https://bb.linuxsampler.org/viewtopic.php?f=6&t=151 > > I used aseqdump to make sure that it's not the keyboard, interface or kernel > who is delaying things - the events seem to come in just fine, which is also > reflected in QSampler's "MIDI activity" LED (the green spot to the left of > the used/active voice indicator). It blinks as soon as MIDI activity happens, > but the corresponding notes are coming late. > > I remember having seen this with LinuxSampler before, so I am pretty sure > that it's not connected to my recent 8.04->8.10 Ubuntu update.
Additionally I can now say after some days of testing and testing again that this does not happen always - in one case it did not happen at all during a whole evening. I have a feeling as if it depends on the boot process somehow - sometimes I boot the system, and everything works fine, the next day I boot again in exactly the same manner, and the problem is there again. I also tested this today on another computer, an Asus eeePC 1000H, running the same Ubuntu 8.10, and had no problems here so far. That machine is only a 32-bit architecture, though. Apart from that, its configuration is similar to the desktop PC: - kernel 2.6.27-8 - JACK 0.116.1 - Linuxsampler, libgig, liblscp, qsampler from CVS 20081220 - ALSA 1.0.17 (same for the 64bit system - had forgotten to mention that in the previous mail) Question for the experts: Could this be an ALSA driver issue? Like, could the ALSA sequencer subsystem create, say, "erroneous" event packages with something like invalid timestamp data so that LinuxSampler thinks it has to delay the playing of events? As far as I understood the code in srv/drivers/midi/MidiInputDeviceAlsa.cpp there is no "timestamping" for (inbound) events, so this is perhaps not the problem - just checking.. Any suggestions on how I could best trace/debug this are most welcome. As I had said above - the events appear to come in through ALSA just fine. Is my 64-bit architecture a possible problem? Thanks for reading, Frank ------------------------------------------------------------------------------ Check out the new SourceForge.net Marketplace. It is the best place to buy or sell services for just about anything Open Source. http://p.sf.net/sfu/Xq1LFB _______________________________________________ Linuxsampler-devel mailing list Linuxsampler-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxsampler-devel
