Hi

I looked at the .py files in examples and it mostly makes sense. But 
supposed I have recieved (or generated) a midi message in python, how do 
I send it to csound?

I understand Iain is doing something like what is in wxController.py:

self.csound.InputMessage("i 1 0 8 %i 70" % self.pitch)

However I'd like to use the dynamic voice allocation facility of csound, 
so I would rather pass on the midi messages...

-- 
peace, love & harmony
Atte

http://www.atte.dk

Istvan Varga wrote:

> or send MIDI messages using some sort of MIDI
> loopback device.

Do you mean alsa -> python -> alsa -> csound?

I can create a midi device from pyton that's seen by csound:

The available MIDI in devices are:
    0: Midi Through Port-0 (ALSA)
    1: MCS Out (ALSA) <- Mine

So simply starting csound from within python with "-M 1", is that what 
you mean?

-- 
peace, love & harmony
Atte

http://www.atte.dk

> I looked at the .py files in examples and it mostly makes sense. But
> supposed I have recieved (or generated) a midi message in python, how do
> I send it to csound?
> 
> I understand Iain is doing something like what is in wxController.py:
> 
> self.csound.InputMessage("i 1 0 8 %i 70" % self.pitch)
> 
> However I'd like to use the dynamic voice allocation facility of csound,
> so I would rather pass on the midi messages...

If you want to handle midi input from your host, as I do, I personally
suggest C++, but maybe things are better since I last looked. When I
hunted around about six months ago none of python real time midi
libraries looked ready, that may have changed. I found it not too
difficult to get real time midi input working with portmidi and C++.

Iain

> Currently, you can only send MIDI messages to, and receive from Csound with
> a C or C++ host, as these require using callback functions for which Python
> and Java wrappers are not implemented yet.
> So, you can either use real time score events with Csound::InputMessage() or
> Csound::ScoreEvent() (note also the software bus interfaces for sending and
> receiving control signals), or send MIDI messages using some sort of MIDI
> loopback device.

Is there much of a performance difference on the csound end between the
two? Seems like input message would require more parsing, but I don't
know the internal well enough to know if that matters.

Thanks
Iain

Atte, what kind of voice allocation are you wanting? That is always one
of the problems of real time control, even with hardware. ( droped midi
notes yay! hung midi notes, yay! ) In order to always *know* that we are
within our cpu use goals we need some sort of hard polyphony top out,
but how that is handled is pretty open.

I was thinking I could implement one for an api example using midi
input. If you are a pianist, I guess the most pianistic is:

- only one instance of an active note per pitch
- subsequent notes of a pitch cut off previous ones
- notes release after finger release ( imagine pedalling )
- potentially all playable notes could be releasing at the same time
  ( sweep keyboard with pedal down )

You also said you want to be able to switch sounds of a split correct?

Iain

Atte André Jensen wrote:
> Hi
> 
> I looked at the .py files in examples and it mostly makes sense. But
> supposed I have recieved (or generated) a midi message in python, how do
> I send it to csound?
> 
> I understand Iain is doing something like what is in wxController.py:
> 
> self.csound.InputMessage("i 1 0 8 %i 70" % self.pitch)
> 
> However I'd like to use the dynamic voice allocation facility of csound,
> so I would rather pass on the midi messages...
> 

Istvan Varga wrote:

> Some new code has just been added to the CVS sources that makes it possible
> to send MIDI messages to, and receive from Csound using the Python wrappers.
> The attached simple example shows sending MIDI input from Python.

That looks very nice, thanks! I'll look into it asap. I tried to run the 
attached example like this:

[atte@aarhus ace]$ python pymidi.py -o devaudio pymidi.csd
Localisation of messages is disabled, using default language.
time resolution is 0.417 ns
Exception in thread Thread-1:Traceback (most recent call last):
   File "/usr/lib/python2.3/threading.py", line 442, in __bootstrap
     self.run()
   File "/usr/lib/python2.3/threading.py", line 422, in run
     self.__target(*self.__args, **self.__kwargs)
   File "pymidi.py", line 60, in guiThread
     w.mainloop()
   File "/usr/lib/python2.3/lib-tk/Tkinter.py", line 965, in mainloop
     self.tk.mainloop(n)
RuntimeError: Calling Tcl from different appartment

This plays 4 C's in accending octaves, so I guess the midi part of the 
example is working. Should I just forget about the tkinter stuff? I'm 
using wxpython, so unless there's a point that I should understand 
besides tkinter, I might forget about the errors, or...?

-- 
peace, love & harmony
Atte

http://www.atte.dk

Istvan Varga wrote:

> Probably yes. The GUI was added only to show using a separate audio
> thread with CsoundPerformanceThread and redirecting the Csound message
> output.

Ok. Thanks again!

-- 
peace, love & harmony
Atte

http://www.atte.dk