Observations from Luis Jure...
Dave Phillips escribió: > > This is definitely not happening for me. However, I am running a > low-latency kernel, with Csound compiled with the setuid flag. hi dave, yet another mail about this babo business! i nailed down the problem: the -b flag. i don't have an extremely fast machine, and don't have the low-latency patched kernel either, so i have to set a bigger buffer size when using the babo opcode in real-time. that seems to cause problems in the unofficial linux csound. i haven't been too systematic, but i can state the following: 1) using for example -b4096 for buffer size produces temporal deviations in the real-time output of orchestras using babo opcode (the pulses of the fof opcode in the example, that should be regular, become irregular) 2) that doesn't happen with lower buffer sizes, or not using the -b flag at all (but for the problems i mentioned above, in my machine i get plops in the output, since the compilation isn't fast enough) 3) i haven't been able to reproduce this curious behavior with other opcodes except babo (but as i said, i haven't been systematic) 4) this doesn't happen in the "official" version, either the buffer is big enough for a clean output or not, but the tempo doesn't change 5) the same behavior was reproduced in two different machines: at home (suse 7.2, kernel 2.4.7, yamaha pci card, alsa drivers 0.9.0beta6), and at school (mandrake 8.0, soundblaster live, whatever kernel and alsa drivers that come with the distribution). both machines seem to work ok. BTW, where is that kernel patch available? is it available for all kernel versions? i'd like to give it a try... best, lj
