Re: [fluid-dev] fluidsynth~

[Peter Hanappe]

Norbert Schnell wrote:
> Hi Peter!
> Hi others!

Hi Norbert,
Hi mailing list,

> As you might have heard, finally I did a Max/MSP object fluidsynth~
> based on your code.

So I heard!

> The technique I used to integrate FluidSynth with Max/MSP is to grab
> the files into a CodeWarrior project (simply requiring a
> fluidsynth directory containing the FluidSynth sources next to the
> directory of the CodeWarrior project) and add one file faking some
> functions of the other files (essentially fluid_sys.c for the threads,
> timers and logs). This has the advantage of avoiding stupid #ifdef
> PP-instructions and messing up the code with Max/MSP API related stuff.

>    1. Would you have a better idea for handling the FluidSynth Max/MSP 
> integration?
>    2. Shell we integrate fluidsynth~ somehow (and how?) into the 
> FluidSynth sources?

I'd suggest to make a directory called 'bindings', or something similar,
in the fluidsynth root directory. The projects that 'bind' fluidsynth to
an application can go in there. For example, I have some Java
classes and JNI code to use the synth in Java that I would like to add
to CVS. So, if we take that option the codewarrior project could move to
the directory 'bindings/max_msp'.

About fluid_sys.c. It would be nice if you could reuse the existing
code, despite the #ifdef's (I don't like them neither, will find a
better solution). However, there is really not much you need from
fluid_sys.c. I don't think you need timers, threads, nor sockets
for the fluidsynth~ object. The log function is platform independent.
In fact, we could put that code in a new file. There are two functions
that you may need: fluid_curtime() and fluid_utime(). The synth only
uses them for profiling and to estimate the CPU load. You could
get away with returning 0 in both functions to get started.

So my suggestion would be to add a file 'config_maxmsp.h' in which you
put some of the defines you need and do keep the existing fluid_sys.c
with some minor tweaking.


> Here is a web page of the fluidsynth~ I published some minutes ago on
> our ATR IRCAM site:
>    http://www.ircam.fr/equipes/temps-reel/maxmsp/fluidsynth.html

Fancy!


Cheers and see you soon!
Peter


>    Norbert