On Fri, 28 Nov 2003 09:03 pm, Joerg Anders wrote: > Als already stated: I'd use TiMidity instead. But > TiMidity loads the instruments not until they are really needed. And this causes audible glitches in it's playback for you ? > That is: If the clarinet keeps silent for - say - 20 measures > the clarinet waves are not loaded. In measure 20 TiMidity > starts loading the clarinet. This requires processor time > and leads to a short gap. (Also with my 2.4 GHz Intel) > > Perhaps anybody knows how to change this behaviour? *If* you happen to be using kernel 2.4.20 and ext3 with Timidity then that could've been the source of the problem. > On http://rnvs.informatik.tu-chemnitz.de/DROP/gap.html there > is now an MP3 file (594K) Excellent, thanks... any chance of the FluidSynth string envelope problem and dropouts from Timidity in mp3 format too :) > The gaps appear after: 12, 39, 43 and 47 seconds. Wow, yes, very obvious. Enough to make composing unbearable. > I used pmidi. But this does not matter. Yiu can use kmid or > tse3play or whatever you want: The gaps appear after > randomly choosen time :-(( Well if it doesn't happen with Timidity in alsaseq mode and, as you point out being across many apps, then it must obviously be something to do with the emu10k1 driver. Perhaps there might be some buffer fragments and buffer size adjustments but I'm not sure what values to suggest (if possible at all). --markc