Python music sequencer timing problems

[John O'Hagan]

On Sun, 14 Dec 2008, Bad Mutha Hubbard wrote:
> John O'Hagan wrote:
> > On Wed, 10 Dec 2008, badmuthahubbard wrote:
[...]
> > from time import time, sleep
> >
> > start = time()
> > for event in music:
> >     duration=len(event) #Really, the length of the event
> >     play(event)
> >     while 1:
> >         timer = time()
> >         remaining = start + duration - timer
> >         if remaining < 0.001:
> >             break
> >         else:
> >             sleep(remaining / 2)
> >     stop(event)
> >     start += duration
> >

> Very interesting approach, halving the remaining duration.  Right now
> I'm not working with note-offs, Csound takes care of the duration.  I
> just need to be able to call the notes at the right times.  
[...]

I'm also using the above code without the "stop(event)" line for playing 
non-midi stuff (fixed-length samples for example), which from the sound of it 
also applies to your requirement - IOW, you _can_ use it just to start notes 
at the right time, because the note-playing loop sleeps till then.
 
[...]
> I also like and respect Fluidsynth, but I'm working on a pretty
> microtonal system, and MIDI doesn't have enough microtonal support.
>
[...]
I'm certainly not trying to convert you to Fluidsynth - I'm more likely to 
convert to Csound - but just for the record, you can do microtonal stuff 
fairly easily using Fluidsynth's "settuning" command. In my program I can use 
any equal-tempered division of the octave by preceding each set of events 
which require it with this:

mesg = [ "select " + channel + " " + font
                +  " " + bank+ " " + program +
                "  \ntuning name 0 " + program + "\n"]

for key in xrange(128):
    pitch = str((((key - 60) * 12.0 / divisor) + 60) * 100)
    mesg.append( "tune  " + bank + " " + program 
                          + " " + str(key) + " " + pitch + "\n")

mesg.append("settuning " + channel + " " 
                       + bank  " " + program + "\n")
mesg = ''.join(mesg)
fluidsynth_socket.send(mesg)

where "divisor" is the number you want to divide the octave by (the other 
names are self-descriptive). The reason for the "60"s in there is just to 
keep middle C where it is. 

For more complex/non-linear pitch-warping, the rest of the "pitch = " line 
could be replaced with any pitch-altering function taking "key" as an 
argument; but I haven't tried that yet. 

I'd be interested to hear about the microtonal system(s) you're using.

Regards,

John