At Fri, 04 Jun 2004 14:25:01 -0400, Dave Phillips wrote: > > Takashi Iwai wrote: > > >I'm not sure in which situation it hogs the CPU on Dave's system. > >But, in general, it might be because of the sample interpolation. > >The recent timidity uses an expensive method (gaussian). When many > >voices are used, it costs high CPU power. > >Using a cheaper one like the linear interpolation (either -N0 or > >-Efresamp=l) will help in such a case. > > > > > I'll try it and see how it works over time. I'm using TiMidity 2.11.3, > it's all right when I start using it, but if I leave it running > eventually the CPU usage climbs to a very high amount, sometimes close > to 90%, while idling.
do you mean timidity 2.13.0, or really 2.11.3? > Btw, I don't see the resamp option listed in 'timidity -h', perhaps I > need a newer version ? I'd like JACK support with TiMidity anyway, I > think the latest version now does, yes ? yes, grab the latest one from timidity.sf.net. Takashi
