[iiwusynth-devel] Re: Pitch bend, filter, sysex, LADSPA,...

[Josh Green]

On Sat, 2002-04-06 at 16:46, Peter Hanappe wrote:
> On Sat, 2002-04-06 at 19:21, Josh Green wrote:
> 
> Actually, that's a good idea Josh, about using the ladspa plugins for
> the chorus. I integrated the chorus effect from sox into the iiwusynth
> and it doesn't sound good. Maybe I made errors in the conversion or
> don't use the best parameters. However, I guess that Steve Harris'
> chorus plugin is a better option. Instead of copy/pasting the code, it
> might be smarter to simple load the plugin instead.
> 

Yeah, but this would also add another dependency. I don't see a problem
with ripping the code :) Having pluggable LADSPA modules for the
different effects would be nice though.

> 
> > 
> > What would be really cool next is to implement real time modulation of
> > as many of the sound font parameters as possible :) Sound font
> > modulators could then be implimented on top of this. How easy would it
> > be to do this? (Peter?) 
> 
> Are you thinking about modulating them by sending MIDI events or
> controlling them with the LFO used to modulate the pitch, etc...? 
> 

I was thinking modulating them with MIDI events or a GUI slider, etc. If
there was an iiwusynth API call, it wouldn't really matter what was
generating the modulation (mouse, joystick, etc).

> 
> It is currently impossible to address a particular voice inside a
> channel. This is somewhat annoying. Imagine we're able to stream large
> samples from disk. It would be nice to control these streamed samples
> separately. That way, you could actually use the synth as a nice mixer
> or audio file sequencer. But you'd need to be able to control voices
> individually otherwise the user is limited to 16 samples, one per MIDI
> channel.
> 

I wonder about this 16 MIDI channel thing. Seems like we could get
around it (at least in software). The routing to external hardware would
of course still need to be 16 channel, but I don't see any reason why we
can't have more in software. I wonder if the ALSA sequencer has any
particular limit.

I'm still trying to figure out whether I should have my own drivers for
Swami (PCM, MIDI, etc) or whether I should just require iiwusynth and
help to make its drivers flexible enough for all the little bells and
whistles I'd like to have (ability to record; playback of audio, from
sample files, etc; setting loop points while listening).

It seems kind of stupid to re-implement all the nice things in iiwusynth
(LADSPA host, etc) when the API could just be strengthened for use by
Swami. I don't really see any problems with requiring iiwusynth.
Can you see iiwusynth supporting all those features I mentioned? It
would basically need to have some way to stream audio (callback based?)
and mix it into the output (for playback) and provide a callback for
reading recorded data. I've heard you mention disk streaming before, so
some sort of API would most likely be developed to allow streaming data.
Simple audio files could then just be played by setting up a voice. If
the loop points could be modified in real time, then that basically
covers most of my needs (besides recording). Recording might be a
problem since not all cards do full duplex, so the output would need to
be stopped in those cases :( Perhaps I should just leave that up to
other applications, although I would like a nice level meter with a
record program (anyone know of any?).

Anyways.. Cheers!
        Josh Green