Re: [fluid-dev] FluidSynth sequencer test

[Pedro Lopez-Cabanillas]

David Henningsson wrote:
> Pedro Lopez-Cabanillas skrev:
> > David Henningsson wrote:
> > Using a big period size (either 2048 or 4096 is enough) instead of the
> > default size (64 in Linux) makes the test program to sound like a drunken
> > drummer. Of course, using the right function: fluid_settings_setint().
> > I've attached the modified sources.
> >
> > Now, the plan is to fix the problem. I guess you are going to apply a fix
> > similar to your patch for the MIDI player.
>
> Sure, if that also means my patch(es) will be committed to the SVN tree.
> The longer we wait, the larger the possibility that some other change in
> the tree will break my patch.

There are several options: either Josh or somebody else commits your patch, or 
Josh gives you write permission to the SVN so you can commit it yourself, or 
I will do it after the 1.0.9 version is released. I can't give you 
permissions or release FluidSynth myself. It is something that only Josh can 
do, and he is not answering mails right now.

I don't fear any problem if we have to wait a bit. OTOH, I've already 
explained that I'm not comfortable with some bits of your patch, and if I 
would change it slightly before applying, in addition to fix the MSVC 
compiler problems. Don't worry about that, I'm happy taking care of this 
issue.

> > I also would like to test an
> > additional approach, overhauling the current wall clock implementation.
>
> Okay, can you elaborate on what changes that would be and if/how they
> will work with faster-than-realtime rendering?

Faster than realtime has no sense for the sequencer. It has sense, though, if 
the sequencer is driven by a slave timer, like the one you have implemented. 
This design is already inside (for instance) the ALSA sequencer, which allows 
several timer sources: slave timers linked to the PCM streams, and wall clock 
timers using the RTC, HPET and System devices. There is some documentation 
about that: http://www.alsa-project.org/~frank/alsa-sequencer/node5.html

The user should be able to choose between wall clock and slave timer, the same 
option for the sequencer and the MIDI player.

My plan is to replace the Windows timer implementation by the multimedia timer 
in that platform, and add the ALSA timers in Linux.  I've already implemented 
something similar for other projects: 
http://plcl.users.sourceforge.net/timer/

Regards,
Pedro