fluid-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[fluid-dev] FluidSynth 1.0.1


From: Josh Green
Subject: [fluid-dev] FluidSynth 1.0.1
Date: 10 Apr 2003 22:21:37 -0700

I've been doing some testing with the new FluidSynth 1.0.1 and find that
it still has some rather serious problems. I'm not sure if these are
directly related to the ALSA audio output driver or not. Just wanted to
post my initial findings, just in case they have already been fixed.

- First off I found that the original CPU usage problem was still there
(constant high CPU spikes at regular intervals). I traced it down to the
removal of the snd_pcm_nonblock call. Perhaps this was removed because
"SND_PCM_NONBLOCK" is used when the device is opened, but I seem to
recall this being only for the open and not for access. After adding the
snd_pcm_nonblock call back, it cleared up the constant CPU usage
problem.

- Another CPU usage problem now occurs, not sure if it is related (this
is a constant 100% CPU usage, rather than regular interval spikes). If I
stop playing notes on the Swami piano, FluidSynth will start consuming
all CPU time after a dozen seconds or so. Playing notes when in the max
CPU consumption state will eventually cause it to return to normal. The
longer I leave it in this state though, the longer it takes playing
notes to bring it back to normal.

- Every once in a while (very rarely) the 100% CPU consumption will
occur permanently during playing of notes. It remains this way until
FluidSynth is restarted.

Needless to say, I don't run it as root :)

I notice that the filter kaaabooooom! problem is indeed fixed, but it
seems at the cost of high Reverb width values. Is this necessary to fix
the problem? I really did like playing with those harsh reverb sounds
(useful for stunning insects and small mammals).

I'll see if I can provide more useful info on these problems. Cheers.
        Josh





reply via email to

[Prev in Thread] Current Thread [Next in Thread]