Yes, i statements work the same as f statements in inputMessage.

Yes, use the C/C++ documentation for Python as well. In addition, you
can in python do this:

import csnd
import CsoundAC
print help(csnd)
print help(CsoundAC)

Hope this helps,
Mike

On 1/4/10, Anthony Palomba  wrote:
> I am using the csound/python API to load a csd file. I would like to be
> able to send events to an instrument in this csd in real time via python.
>
> I have been analyzing the example drone.py, it looks like you can send
> messages and set values via python:
>
>     self.csound.SetChannel("gkDistortFactor",
> float(self.gkDistortFactor.get()))
>
>      # Apply initial control channel values before actually starting
> synthesis.
>      f = self.configuration.gkHarmonicTableFactor
>      message = 'f  1  0  65536  10  3  0  %f  0  0 %f\n' % (f * 1.0, f *
> 2.0)
>      self.csound.inputMessage(message)
>
>
> Would I be able to send an i-statement using inputMessage?
>
> Also, is there some documentation that lists all the functions
> supported by the csound/python interface? Does it support
> all the functions in the csound API reference manual?
>
>
>
> Thanks,
> Anthony
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"

Your program is ambitious, and also one that I have thought about myself.

First of all, CsoundAC does a lot of what Common Music does, plus
other things that Common Music does not know how to do.

Second, Christopher Ariza has developed athenaCL, a Python module that
also does a lot of stuff that Common Music and CsoundAC do, plus other
things that neither of them do. He is working on a new version that
should be easier to use by scripting Python (currently athenaCL is
supposed to be used in an interactive command line mode).

Third, if I were to take this on, I would use Lua not Python for
various reasons (you can build Lua in, don't need to install it;
LuaJIT is faster than Python; etc.).

Feel free to talk about this at length with me. Do you have a
specification document or notes on what you want to achieve?

P.S. yes, you can use a timer to call in events. What you might want
is to set up a processKsmps function in Python that dequeues events
and dispatches them to Csound before calling csoundProcessKsmps
proper. Then your timer callbacks can enqueue events on your Csound
event queue.

There are lots of facilities for doing this kind of work -- have you
looked at Max/MSP or Pure Data? Or have you used Common Music itself?

Hope this helps,
Mike







On 1/4/10, Anthony Palomba  wrote:
> Thanks Micheal, I will give that a try. Another quick question...
>
> I want to be able to play gestures in real time. The gesture
> would be created by some python function that takes as its parameters
> a time value.
>
> Is there a way you can define a timer in python that would allow me to call
> my gesture
> function at a constant rate?
>
> Ultimately, I want to develop a system like Common Music but python based.
>
>
>
> Anthony
>
>
>
> On Mon, Jan 4, 2010 at 2:23 PM, Michael Gogins
> wrote:
>
>> Yes, i statements work the same as f statements in inputMessage.
>>
>> Yes, use the C/C++ documentation for Python as well. In addition, you
>> can in python do this:
>>
>> import csnd
>> import CsoundAC
>> print help(csnd)
>> print help(CsoundAC)
>>
>> Hope this helps,
>> Mike
>>
>> On 1/4/10, Anthony Palomba  wrote:
>> > I am using the csound/python API to load a csd file. I would like to be
>> > able to send events to an instrument in this csd in real time via
>> > python.
>> >
>> > I have been analyzing the example drone.py, it looks like you can send
>> > messages and set values via python:
>> >
>> >     self.csound.SetChannel("gkDistortFactor",
>> > float(self.gkDistortFactor.get()))
>> >
>> >      # Apply initial control channel values before actually starting
>> > synthesis.
>> >      f = self.configuration.gkHarmonicTableFactor
>> >      message = 'f  1  0  65536  10  3  0  %f  0  0 %f\n' % (f * 1.0, f *
>> > 2.0)
>> >      self.csound.inputMessage(message)
>> >
>> >
>> > Would I be able to send an i-statement using inputMessage?
>> >
>> > Also, is there some documentation that lists all the functions
>> > supported by the csound/python interface? Does it support
>> > all the functions in the csound API reference manual?
>> >
>> >
>> >
>> > Thanks,
>> > Anthony
>> >
>> > Send bugs reports to this list.
>> > To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>> > csound"
>>
>>
>> --
>> Michael Gogins
>> Irreducible Productions
>> http://www.michael-gogins.com
>> Michael dot Gogins at gmail dot com
>>
>>
>> Send bugs reports to this list.
>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>> csound"
>>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"

Just my 2 cents on this.
Yes, I think Python should be fast enough for most compositional
algorithms. There's also optimization methods (psyco, Cpython, etc)
that you could use to gain efficiency, these are very effective.
What I do in ImproSculpt is run Csound in a ksmps loop, and then
increment the clock inside this loop. I use a separate module for
timing and queing, where each event is time stamped, and dispersed
when it's time is come. The timer starts at zero when the application
starts, and is incremented in 1/kr seconds once every ksmps loop
iteration (if you want to keep time in seconds). This is not a perfect
solution, as any event calls that takes more then 1/kr seconds to
complete will leads to Csound dropouts. But, it ensures k-rate sync
between Python and Csound though.

I would not use Max for timing functions, at least not the standard
metro objects etc. These may deviate in timing precision depending on
system load (the same will happen to Python time.time() btw.) If Max
has a sample precice clock, you might consider using it.

best
Oeyvind

2010/1/4 Anthony Palomba :
> So I think you are saying that python might not be fast enough to handle
> things?
> Perhaps it would be best to leave the timing to something more reliable like
> Max?
>


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Sounds the same as what I was saying.

MKG from cell phone

On Jan 5, 2010 3:33 AM, "Oeyvind Brandtsegg" <obrandts@gmail.com> wrote:

Just my 2 cents on this.
Yes, I think Python should be fast enough for most compositional
algorithms. There's also optimization methods (psyco, Cpython, etc)
that you could use to gain efficiency, these are very effective.
What I do in ImproSculpt is run Csound in a ksmps loop, and then
increment the clock inside this loop. I use a separate module for
timing and queing, where each event is time stamped, and dispersed
when it's time is come. The timer starts at zero when the application
starts, and is incremented in 1/kr seconds once every ksmps loop
iteration (if you want to keep time in seconds). This is not a perfect
solution, as any event calls that takes more then 1/kr seconds to
complete will leads to Csound dropouts. But, it ensures k-rate sync
between Python and Csound though.

I would not use Max for timing functions, at least not the standard
metro objects etc. These may deviate in timing precision depending on
system load (the same will happen to Python time.time() btw.) If Max
has a sample precice clock, you might consider using it.

best
Oeyvind

2010/1/4 Anthony Palomba <apalomba@austin.rr.com>:

> So I think you are saying that python might not be fast enough to handle > things? > Perhaps it wo...

Send bugs reports to this list. To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubsc...

Actually, I've got a question on this...
Is there any difference in explicitly writing a python loop with
performKsmps and some function...
i.e.

while True:
      csound.performKsmps(True)
      somefunc()

As opposed to simply using using the SetProcessCallback method on the
CsoundPerformanceThread class?

Thorin

On Tue, Jan 5, 2010 at 11:02 PM, Michael Gogins
 wrote:
> Sounds the same as what I was saying.
>
> MKG from cell phone
>
> On Jan 5, 2010 3:33 AM, "Oeyvind Brandtsegg"  wrote:
>
> Just my 2 cents on this.
> Yes, I think Python should be fast enough for most compositional
> algorithms. There's also optimization methods (psyco, Cpython, etc)
> that you could use to gain efficiency, these are very effective.
> What I do in ImproSculpt is run Csound in a ksmps loop, and then
> increment the clock inside this loop. I use a separate module for
> timing and queing, where each event is time stamped, and dispersed
> when it's time is come. The timer starts at zero when the application
> starts, and is incremented in 1/kr seconds once every ksmps loop
> iteration (if you want to keep time in seconds). This is not a perfect
> solution, as any event calls that takes more then 1/kr seconds to
> complete will leads to Csound dropouts. But, it ensures k-rate sync
> between Python and Csound though.
>
> I would not use Max for timing functions, at least not the standard
> metro objects etc. These may deviate in timing precision depending on
> system load (the same will happen to Python time.time() btw.) If Max
> has a sample precice clock, you might consider using it.
>
> best
> Oeyvind
>
> 2010/1/4 Anthony Palomba :
>
>> So I think you are saying that python might not be fast enough to handle >
>> things? > Perhaps it wo...
>
> Send bugs reports to this list. To unsubscribe, send email
> sympa@lists.bath.ac.uk with body "unsubsc...


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

..Looking at your code, I'm guessing the reason it doesn't sound is
because you invoke a complete Csound performance in the same thread
from which you are sending events. You either need to send the events
from a different thread, or else not use threads at all, but just
handle events in between calls to csoundPerformKsmps, in a loop.

The CsoundPerformanceThread class in the examples/python/shapes.py
program does the former, and the examples/python/drone.py program does
the latter. Both of these approaches work well enough. The loop
approach in drone.py is simpler, and has worked very well for me,
without dropouts or glitches.

Note that in drone.py, the events are function table statements ('f'
statements), and they do cause dropouts because it takes time to
recalculate the function tables. If the events were note statements
('i' statements), they would not cause dropouts; I have other code
that demonstrates this (the examples/python/PythonDemoApp.py program
is one, but you need wxPython to run it).

Note that if you use threads, you have to keep Csound from running at
the instant you are sending events to Csound, otherwise Csound could
become corrupted and crash. The CsoundPerformanceThread class has code
internally that takes care of this.

Hope this helps,
Mike

 On Mon, Jan 25, 2010 at 6:21 PM, Anthony Palomba
 wrote:
> So I have made some progress in investigating the possibility
> of sending realtime procedural  events from python to csound, but I seem
> to have hit a snag...
>
> The goal is to load a csd then send a message to it to play a note.
> I found a nice threading class that manages timed operations.
> It forms the basis of my timing mechanism for sending events to csound.
>
> I created two events, one that starts the csound performance and one that
> plays a note every second. What I am seeing is that nothing gets played.
> Perhaps I am sending the wrong message type? What is the proper way to
> send a note event to csound?
>
> I have attached my .py and .csd files. Any feedback would be greatly
> appreciated.
>
>
>
> Thanks,
> Anthony
>
>
>
>
> On Tue, Jan 5, 2010 at 5:55 PM, Peiman Khosravi 
> wrote:
>>
>> I would not use Max for timing functions, at least not the standard
>>
>> metro objects etc. These may deviate in timing precision depending on
>>
>> system load (the same will happen to Python time.time() btw.) If Max
>>
>> has a sample precice clock, you might consider using it.
>>
>> Yes I agree but you can very easily make a signal-based MSP clock and then
>> convert its output to max messages. That way you are not relying on max for
>> timing so everything is sample-accurate with no fluctuations.
>> Best,
>> Peiman
>



-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

----- Original Message -----

From: Anthony Palomba

To: csound@lists.bath.ac.uk

Sent: Thursday, January 28, 2010 6:40 PM

Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Handling realtime events with python/csound...

Okay, I took a look at shapes.py and have a better idea of what is going on.
But I am still having trouble. The lack of documentation on the python interface
makes it very hard to figure things out.

I now use the CsoundPerformanceThread, like so...

csound = csnd.Csound()
res = csound.Compile("pytest3.csd")
perf = csnd.CsoundPerformanceThread(csound)
perf.Play()

This should create a thread that will run the performance, yes?
But when I know start my message process, which is in another thread, I get
an exception...

<Process(Thread-1, initial daemon)>
playing note!
Unhandled exception in thread started by <bound method Process.run of <Process(Thread-1, initial daemon)>>


Is it possible to send messages to a csound thread from another thread?
I have attached my source file for reference.



Thanks,
Anthony

On Tue, Jan 26, 2010 at 12:04 PM, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:

Why don't you use the CsoundPerformanceThread class? It's there exactly for this purpose.

Victor

On 26 Jan 2010, at 16:50, Anthony Palomba wrote:

You will have to forgive me, I am relatively new to python. I discovered
that my event scheduler was not running the csound performance.
I have fixed the problem, but am now seeing a different problem...

If you look at the source, you will see that each processes is started
as its own thread. When I create the process that starts the csound
performance, it looks like csound.perform() blocks and does not let
anything else run.

I am not sure why it is doing that, because the csound performance should
be running in a separate thread. I have attached my source file for
reference.



-ap

On Mon, Jan 25, 2010 at 5:51 PM, Michael Gogins <michael.gogins@gmail.com> wrote:

..Looking at your code, I'm guessing the reason it doesn't sound is
because you invoke a complete Csound performance in the same thread
from which you are sending events. You either need to send the events
from a different thread, or else not use threads at all, but just
handle events in between calls to csoundPerformKsmps, in a loop.

The CsoundPerformanceThread class in the examples/python/shapes.py
program does the former, and the examples/python/drone.py program does
the latter. Both of these approaches work well enough. The loop
approach in drone.py is simpler, and has worked very well for me,
without dropouts or glitches.

Note that in drone.py, the events are function table statements ('f'
statements), and they do cause dropouts because it takes time to
recalculate the function tables. If the events were note statements
('i' statements), they would not cause dropouts; I have other code
that demonstrates this (the examples/python/PythonDemoApp.py program
is one, but you need wxPython to run it).

Note that if you use threads, you have to keep Csound from running at
the instant you are sending events to Csound, otherwise Csound could
become corrupted and crash. The CsoundPerformanceThread class has code
internally that takes care of this.

Hope this helps,
Mike

 On Mon, Jan 25, 2010 at 6:21 PM, Anthony Palomba
<apalomba@austin.rr.com> wrote:
> So I have made some progress in investigating the possibility
> of sending realtime procedural  events from python to csound, but I seem
> to have hit a snag...
>
> The goal is to load a csd then send a message to it to play a note.
> I found a nice threading class that manages timed operations.
> It forms the basis of my timing mechanism for sending events to csound.
>
> I created two events, one that starts the csound performance and one that
> plays a note every second. What I am seeing is that nothing gets played.
> Perhaps I am sending the wrong message type? What is the proper way to
> send a note event to csound?
>
> I have attached my .py and .csd files. Any feedback would be greatly
> appreciated.
>
>
>
> Thanks,
> Anthony
>
>
>
>
> On Tue, Jan 5, 2010 at 5:55 PM, Peiman Khosravi <peimankhosravi@gmail.com>
> wrote:
>>
>> I would not use Max for timing functions, at least not the standard
>>
>> metro objects etc. These may deviate in timing precision depending on
>>
>> system load (the same will happen to Python time.time() btw.) If Max
>>
>> has a sample precice clock, you might consider using it.
>>
>> Yes I agree but you can very easily make a signal-based MSP clock and then
>> convert its output to max messages. That way you are not relying on max for
>> timing so everything is sample-accurate with no fluctuations.
>> Best,
>> Peiman
>

--

Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

<pytest3.py>

----- Original Message -----

From: ilterzouomo

To: csound@lists.bath.ac.uk

Sent: Thursday, January 28, 2010 6:52 PM

Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Handling realtime events with python/csound...

Hi Mr. Anthony,

this work to me:

 

 

import csnd

 

cs = csnd.Csound()
p = csnd.CsoundPerformanceThread(cs)

 

cs.Compile("piping01.csd")
p.Play()

 

# now csd is running (with a dummy ftable) and i can send messages like ...

 

p.InputMessage("i1 0 1")

 

p.InputMessage("i1 2 1\
                         i1 3 1 ")

 

 

a mini help,

ciao,

fran.

 

----- Original Message -----

From: Anthony Palomba

To: csound@lists.bath.ac.uk

Sent: Thursday, January 28, 2010 6:40 PM

Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Handling realtime events with python/csound...

Okay, I took a look at shapes.py and have a better idea of what is going on.
But I am still having trouble. The lack of documentation on the python interface
makes it very hard to figure things out.

I now use the CsoundPerformanceThread, like so...

csound = csnd.Csound()
res = csound.Compile("pytest3.csd")
perf = csnd.CsoundPerformanceThread(csound)
perf.Play()

This should create a thread that will run the performance, yes?
But when I know start my message process, which is in another thread, I get
an exception...

<Process(Thread-1, initial daemon)>
playing note!
Unhandled exception in thread started by <bound method Process.run of <Process(Thread-1, initial daemon)>>


Is it possible to send messages to a csound thread from another thread?
I have attached my source file for reference.



Thanks,
Anthony

On Tue, Jan 26, 2010 at 12:04 PM, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:

Why don't you use the CsoundPerformanceThread class? It's there exactly for this purpose.

Victor

On 26 Jan 2010, at 16:50, Anthony Palomba wrote:

You will have to forgive me, I am relatively new to python. I discovered
that my event scheduler was not running the csound performance.
I have fixed the problem, but am now seeing a different problem...

If you look at the source, you will see that each processes is started
as its own thread. When I create the process that starts the csound
performance, it looks like csound.perform() blocks and does not let
anything else run.

I am not sure why it is doing that, because the csound performance should
be running in a separate thread. I have attached my source file for
reference.



-ap

On Mon, Jan 25, 2010 at 5:51 PM, Michael Gogins <michael.gogins@gmail.com> wrote:

..Looking at your code, I'm guessing the reason it doesn't sound is
because you invoke a complete Csound performance in the same thread
from which you are sending events. You either need to send the events
from a different thread, or else not use threads at all, but just
handle events in between calls to csoundPerformKsmps, in a loop.

The CsoundPerformanceThread class in the examples/python/shapes.py
program does the former, and the examples/python/drone.py program does
the latter. Both of these approaches work well enough. The loop
approach in drone.py is simpler, and has worked very well for me,
without dropouts or glitches.

Note that in drone.py, the events are function table statements ('f'
statements), and they do cause dropouts because it takes time to
recalculate the function tables. If the events were note statements
('i' statements), they would not cause dropouts; I have other code
that demonstrates this (the examples/python/PythonDemoApp.py program
is one, but you need wxPython to run it).

Note that if you use threads, you have to keep Csound from running at
the instant you are sending events to Csound, otherwise Csound could
become corrupted and crash. The CsoundPerformanceThread class has code
internally that takes care of this.

Hope this helps,
Mike

 On Mon, Jan 25, 2010 at 6:21 PM, Anthony Palomba
<apalomba@austin.rr.com> wrote:
> So I have made some progress in investigating the possibility
> of sending realtime procedural  events from python to csound, but I seem
> to have hit a snag...
>
> The goal is to load a csd then send a message to it to play a note.
> I found a nice threading class that manages timed operations.
> It forms the basis of my timing mechanism for sending events to csound.
>
> I created two events, one that starts the csound performance and one that
> plays a note every second. What I am seeing is that nothing gets played.
> Perhaps I am sending the wrong message type? What is the proper way to
> send a note event to csound?
>
> I have attached my .py and .csd files. Any feedback would be greatly
> appreciated.
>
>
>
> Thanks,
> Anthony
>
>
>
>
> On Tue, Jan 5, 2010 at 5:55 PM, Peiman Khosravi <peimankhosravi@gmail.com>
> wrote:
>>
>> I would not use Max for timing functions, at least not the standard
>>
>> metro objects etc. These may deviate in timing precision depending on
>>
>> system load (the same will happen to Python time.time() btw.) If Max
>>
>> has a sample precice clock, you might consider using it.
>>
>> Yes I agree but you can very easily make a signal-based MSP clock and then
>> convert its output to max messages. That way you are not relying on max for
>> timing so everything is sample-accurate with no fluctuations.
>> Best,
>> Peiman
>

--

Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

<pytest3.py>

I did test it quickly, and here's a few things you can change:
- Use capitalized "I" on InputMessage, not inputMessage. It's an easy
error to make
- don't allow your main thread to finish, as all the child threads
will have nowhere to go...
Add an
import time
and
time.sleep(5)
to keep it running for a short time
- Stop and Join the csound perf thread before exiting the main thread,
this will clean up nicely before stopping.
Add these to the end of the python script:
perf.Stop()
perf.Join()

best
Oeyvind

2010/1/28 Anthony Palomba :
> Hey Fran, thanks for the examples. It looks like you are sending messages
> from the python executing process to the csound performance thread.
>
> In my example, I create a second thread which then sends the message
> to the csound performance thread.
>
> My intention is to be able to have algorithmic processes that run on their
> own, that send messages or update parameters of a csound performance.
> These processes could expire after a given period of time or play
> infinitely.
>
> Does anyone know if it is possible to send messages to a csound thread
> from another thread? My example code was posted previously.
>
>
>
> -ap
>
>
> On Thu, Jan 28, 2010 at 12:02 PM, ilterzouomo 
> wrote:
>>
>> I forget to say that if You run the file from some editor (i use scite)
>> You need to import time and add at the end of file  time.sleep(5), that
>> is wait for 5 sec.
>>
>> ciao,
>> fran.
>>
>> ----- Original Message -----
>> From: ilterzouomo
>> To: csound@lists.bath.ac.uk
>> Sent: Thursday, January 28, 2010 6:52 PM
>> Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:
>> Handling realtime events with python/csound...
>> Hi Mr. Anthony,
>> this work to me:
>>
>>
>> import csnd
>>
>> cs = csnd.Csound()
>> p = csnd.CsoundPerformanceThread(cs)
>>
>> cs.Compile("piping01.csd")
>> p.Play()
>>
>> # now csd is running (with a dummy ftable) and i can send messages like
>> ...
>>
>> p.InputMessage("i1 0 1")
>>
>> p.InputMessage("i1 2 1\
>>                          i1 3 1 ")
>>
>>
>> a mini help,
>> ciao,
>> fran.
>>
>>
>> ----- Original Message -----
>> From: Anthony Palomba
>> To: csound@lists.bath.ac.uk
>> Sent: Thursday, January 28, 2010 6:40 PM
>> Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:
>> Handling realtime events with python/csound...
>> Okay, I took a look at shapes.py and have a better idea of what is going
>> on.
>> But I am still having trouble. The lack of documentation on the python
>> interface
>> makes it very hard to figure things out.
>>
>> I now use the CsoundPerformanceThread, like so...
>>
>> csound = csnd.Csound()
>> res = csound.Compile("pytest3.csd")
>> perf = csnd.CsoundPerformanceThread(csound)
>> perf.Play()
>>
>> This should create a thread that will run the performance, yes?
>> But when I know start my message process, which is in another thread, I
>> get
>> an exception...
>>
>> 
>> playing note!
>> Unhandled exception in thread started by > >
>>
>>
>> Is it possible to send messages to a csound thread from another thread?
>> I have attached my source file for reference.
>>
>>
>>
>> Thanks,
>> Anthony
>>
>>
>>
>> On Tue, Jan 26, 2010 at 12:04 PM, Victor Lazzarini
>>  wrote:
>>>
>>> Why don't you use the CsoundPerformanceThread class? It's there exactly
>>> for this purpose.
>>> Victor
>>> On 26 Jan 2010, at 16:50, Anthony Palomba wrote:
>>>
>>> You will have to forgive me, I am relatively new to python. I discovered
>>> that my event scheduler was not running the csound performance.
>>> I have fixed the problem, but am now seeing a different problem...
>>>
>>> If you look at the source, you will see that each processes is started
>>> as its own thread. When I create the process that starts the csound
>>> performance, it looks like csound.perform() blocks and does not let
>>> anything else run.
>>>
>>> I am not sure why it is doing that, because the csound performance should
>>> be running in a separate thread. I have attached my source file for
>>> reference.
>>>
>>>
>>>
>>> -ap
>>>
>>>
>>>
>>> On Mon, Jan 25, 2010 at 5:51 PM, Michael Gogins
>>>  wrote:
>>>>
>>>> ..Looking at your code, I'm guessing the reason it doesn't sound is
>>>> because you invoke a complete Csound performance in the same thread
>>>> from which you are sending events. You either need to send the events
>>>> from a different thread, or else not use threads at all, but just
>>>> handle events in between calls to csoundPerformKsmps, in a loop.
>>>>
>>>> The CsoundPerformanceThread class in the examples/python/shapes.py
>>>> program does the former, and the examples/python/drone.py program does
>>>> the latter. Both of these approaches work well enough. The loop
>>>> approach in drone.py is simpler, and has worked very well for me,
>>>> without dropouts or glitches.
>>>>
>>>> Note that in drone.py, the events are function table statements ('f'
>>>> statements), and they do cause dropouts because it takes time to
>>>> recalculate the function tables. If the events were note statements
>>>> ('i' statements), they would not cause dropouts; I have other code
>>>> that demonstrates this (the examples/python/PythonDemoApp.py program
>>>> is one, but you need wxPython to run it).
>>>>
>>>> Note that if you use threads, you have to keep Csound from running at
>>>> the instant you are sending events to Csound, otherwise Csound could
>>>> become corrupted and crash. The CsoundPerformanceThread class has code
>>>> internally that takes care of this.
>>>>
>>>> Hope this helps,
>>>> Mike
>>>>
>>>>  On Mon, Jan 25, 2010 at 6:21 PM, Anthony Palomba
>>>>  wrote:
>>>> > So I have made some progress in investigating the possibility
>>>> > of sending realtime procedural  events from python to csound, but I
>>>> > seem
>>>> > to have hit a snag...
>>>> >
>>>> > The goal is to load a csd then send a message to it to play a note.
>>>> > I found a nice threading class that manages timed operations.
>>>> > It forms the basis of my timing mechanism for sending events to
>>>> > csound.
>>>> >
>>>> > I created two events, one that starts the csound performance and one
>>>> > that
>>>> > plays a note every second. What I am seeing is that nothing gets
>>>> > played.
>>>> > Perhaps I am sending the wrong message type? What is the proper way to
>>>> > send a note event to csound?
>>>> >
>>>> > I have attached my .py and .csd files. Any feedback would be greatly
>>>> > appreciated.
>>>> >
>>>> >
>>>> >
>>>> > Thanks,
>>>> > Anthony
>>>> >
>>>> >
>>>> >
>>>> >
>>>> > On Tue, Jan 5, 2010 at 5:55 PM, Peiman Khosravi
>>>> > 
>>>> > wrote:
>>>> >>
>>>> >> I would not use Max for timing functions, at least not the standard
>>>> >>
>>>> >> metro objects etc. These may deviate in timing precision depending on
>>>> >>
>>>> >> system load (the same will happen to Python time.time() btw.) If Max
>>>> >>
>>>> >> has a sample precice clock, you might consider using it.
>>>> >>
>>>> >> Yes I agree but you can very easily make a signal-based MSP clock and
>>>> >> then
>>>> >> convert its output to max messages. That way you are not relying on
>>>> >> max for
>>>> >> timing so everything is sample-accurate with no fluctuations.
>>>> >> Best,
>>>> >> Peiman
>>>> >
>>>>
>>>>
>>>>
>>>> --
>>>> Michael Gogins
>>>> Irreducible Productions
>>>> http://www.michael-gogins.com
>>>> Michael dot Gogins at gmail dot com
>>>>
>>>>
>>>> Send bugs reports to this list.
>>>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>>>> csound"
>>>
>>> 
>>
>
>


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

You only need to call csound.PerformKsmps() if you're not using the
performanceThread.
The performanceThread is a separate C thread for running Csound.
The other (older) method is requires Python to call PerformKsmps once
for each k-rate pass in Csound (csound performs one k-rate pass and
produces ksmps number of samples each time PerformKsmsps is called).

best
Oeyvind

2010/1/29 Anthony Palomba :
> YAY! It works!
>
> Thanks Oeyvind for your help. Another question, under what circumstances
> would I need to call csound.PerformKsmps()?
>
> I have attached the working version of my py file for everyone's reference.
>
>
>
> Thanks again,
> Anthony
>
>
>
>
> On Thu, Jan 28, 2010 at 3:59 PM, Oeyvind Brandtsegg 
> wrote:
>>
>> I did test it quickly, and here's a few things you can change:
>> - Use capitalized "I" on InputMessage, not inputMessage. It's an easy
>> error to make
>> - don't allow your main thread to finish, as all the child threads
>> will have nowhere to go...
>> Add an
>> import time
>> and
>> time.sleep(5)
>> to keep it running for a short time
>> - Stop and Join the csound perf thread before exiting the main thread,
>> this will clean up nicely before stopping.
>> Add these to the end of the python script:
>> perf.Stop()
>> perf.Join()
>>
>> best
>> Oeyvind
>>
>> 2010/1/28 Anthony Palomba :
>> > Hey Fran, thanks for the examples. It looks like you are sending
>> > messages
>> > from the python executing process to the csound performance thread.
>> >
>> > In my example, I create a second thread which then sends the message
>> > to the csound performance thread.
>> >
>> > My intention is to be able to have algorithmic processes that run on
>> > their
>> > own, that send messages or update parameters of a csound performance.
>> > These processes could expire after a given period of time or play
>> > infinitely.
>> >
>> > Does anyone know if it is possible to send messages to a csound thread
>> > from another thread? My example code was posted previously.
>> >
>> >
>> >
>> > -ap
>> >
>> >
>> > On Thu, Jan 28, 2010 at 12:02 PM, ilterzouomo
>> > 
>> > wrote:
>> >>
>> >> I forget to say that if You run the file from some editor (i use scite)
>> >> You need to import time and add at the end of file  time.sleep(5), that
>> >> is wait for 5 sec.
>> >>
>> >> ciao,
>> >> fran.
>> >>
>> >> ----- Original Message -----
>> >> From: ilterzouomo
>> >> To: csound@lists.bath.ac.uk
>> >> Sent: Thursday, January 28, 2010 6:52 PM
>> >> Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:
>> >> Handling realtime events with python/csound...
>> >> Hi Mr. Anthony,
>> >> this work to me:
>> >>
>> >>
>> >> import csnd
>> >>
>> >> cs = csnd.Csound()
>> >> p = csnd.CsoundPerformanceThread(cs)
>> >>
>> >> cs.Compile("piping01.csd")
>> >> p.Play()
>> >>
>> >> # now csd is running (with a dummy ftable) and i can send messages like
>> >> ...
>> >>
>> >> p.InputMessage("i1 0 1")
>> >>
>> >> p.InputMessage("i1 2 1\
>> >>                          i1 3 1 ")
>> >>
>> >>
>> >> a mini help,
>> >> ciao,
>> >> fran.
>> >>
>> >>
>> >> ----- Original Message -----
>> >> From: Anthony Palomba
>> >> To: csound@lists.bath.ac.uk
>> >> Sent: Thursday, January 28, 2010 6:40 PM
>> >> Subject: [Csnd] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:
>> >> Handling realtime events with python/csound...
>> >> Okay, I took a look at shapes.py and have a better idea of what is
>> >> going
>> >> on.
>> >> But I am still having trouble. The lack of documentation on the python
>> >> interface
>> >> makes it very hard to figure things out.
>> >>
>> >> I now use the CsoundPerformanceThread, like so...
>> >>
>> >> csound = csnd.Csound()
>> >> res = csound.Compile("pytest3.csd")
>> >> perf = csnd.CsoundPerformanceThread(csound)
>> >> perf.Play()
>> >>
>> >> This should create a thread that will run the performance, yes?
>> >> But when I know start my message process, which is in another thread, I
>> >> get
>> >> an exception...
>> >>
>> >> 
>> >> playing note!
>> >> Unhandled exception in thread started by > >> >
>> >>
>> >>
>> >> Is it possible to send messages to a csound thread from another thread?
>> >> I have attached my source file for reference.
>> >>
>> >>
>> >>
>> >> Thanks,
>> >> Anthony
>> >>
>> >>
>> >>
>> >> On Tue, Jan 26, 2010 at 12:04 PM, Victor Lazzarini
>> >>  wrote:
>> >>>
>> >>> Why don't you use the CsoundPerformanceThread class? It's there
>> >>> exactly
>> >>> for this purpose.
>> >>> Victor
>> >>> On 26 Jan 2010, at 16:50, Anthony Palomba wrote:
>> >>>
>> >>> You will have to forgive me, I am relatively new to python. I
>> >>> discovered
>> >>> that my event scheduler was not running the csound performance.
>> >>> I have fixed the problem, but am now seeing a different problem...
>> >>>
>> >>> If you look at the source, you will see that each processes is started
>> >>> as its own thread. When I create the process that starts the csound
>> >>> performance, it looks like csound.perform() blocks and does not let
>> >>> anything else run.
>> >>>
>> >>> I am not sure why it is doing that, because the csound performance
>> >>> should
>> >>> be running in a separate thread. I have attached my source file for
>> >>> reference.
>> >>>
>> >>>
>> >>>
>> >>> -ap
>> >>>
>> >>>
>> >>>
>> >>> On Mon, Jan 25, 2010 at 5:51 PM, Michael Gogins
>> >>>  wrote:
>> >>>>
>> >>>> ..Looking at your code, I'm guessing the reason it doesn't sound is
>> >>>> because you invoke a complete Csound performance in the same thread
>> >>>> from which you are sending events. You either need to send the events
>> >>>> from a different thread, or else not use threads at all, but just
>> >>>> handle events in between calls to csoundPerformKsmps, in a loop.
>> >>>>
>> >>>> The CsoundPerformanceThread class in the examples/python/shapes.py
>> >>>> program does the former, and the examples/python/drone.py program
>> >>>> does
>> >>>> the latter. Both of these approaches work well enough. The loop
>> >>>> approach in drone.py is simpler, and has worked very well for me,
>> >>>> without dropouts or glitches.
>> >>>>
>> >>>> Note that in drone.py, the events are function table statements ('f'
>> >>>> statements), and they do cause dropouts because it takes time to
>> >>>> recalculate the function tables. If the events were note statements
>> >>>> ('i' statements), they would not cause dropouts; I have other code
>> >>>> that demonstrates this (the examples/python/PythonDemoApp.py program
>> >>>> is one, but you need wxPython to run it).
>> >>>>
>> >>>> Note that if you use threads, you have to keep Csound from running at
>> >>>> the instant you are sending events to Csound, otherwise Csound could
>> >>>> become corrupted and crash. The CsoundPerformanceThread class has
>> >>>> code
>> >>>> internally that takes care of this.
>> >>>>
>> >>>> Hope this helps,
>> >>>> Mike
>> >>>>
>> >>>>  On Mon, Jan 25, 2010 at 6:21 PM, Anthony Palomba
>> >>>>  wrote:
>> >>>> > So I have made some progress in investigating the possibility
>> >>>> > of sending realtime procedural  events from python to csound, but I
>> >>>> > seem
>> >>>> > to have hit a snag...
>> >>>> >
>> >>>> > The goal is to load a csd then send a message to it to play a note.
>> >>>> > I found a nice threading class that manages timed operations.
>> >>>> > It forms the basis of my timing mechanism for sending events to
>> >>>> > csound.
>> >>>> >
>> >>>> > I created two events, one that starts the csound performance and
>> >>>> > one
>> >>>> > that
>> >>>> > plays a note every second. What I am seeing is that nothing gets
>> >>>> > played.
>> >>>> > Perhaps I am sending the wrong message type? What is the proper way
>> >>>> > to
>> >>>> > send a note event to csound?
>> >>>> >
>> >>>> > I have attached my .py and .csd files. Any feedback would be
>> >>>> > greatly
>> >>>> > appreciated.
>> >>>> >
>> >>>> >
>> >>>> >
>> >>>> > Thanks,
>> >>>> > Anthony
>> >>>> >
>> >>>> >
>> >>>> >
>> >>>> >
>> >>>> > On Tue, Jan 5, 2010 at 5:55 PM, Peiman Khosravi
>> >>>> > 
>> >>>> > wrote:
>> >>>> >>
>> >>>> >> I would not use Max for timing functions, at least not the
>> >>>> >> standard
>> >>>> >>
>> >>>> >> metro objects etc. These may deviate in timing precision depending
>> >>>> >> on
>> >>>> >>
>> >>>> >> system load (the same will happen to Python time.time() btw.) If
>> >>>> >> Max
>> >>>> >>
>> >>>> >> has a sample precice clock, you might consider using it.
>> >>>> >>
>> >>>> >> Yes I agree but you can very easily make a signal-based MSP clock
>> >>>> >> and
>> >>>> >> then
>> >>>> >> convert its output to max messages. That way you are not relying
>> >>>> >> on
>> >>>> >> max for
>> >>>> >> timing so everything is sample-accurate with no fluctuations.
>> >>>> >> Best,
>> >>>> >> Peiman
>> >>>> >
>> >>>>
>> >>>>
>> >>>>
>> >>>> --
>> >>>> Michael Gogins
>> >>>> Irreducible Productions
>> >>>> http://www.michael-gogins.com
>> >>>> Michael dot Gogins at gmail dot com
>> >>>>
>> >>>>
>> >>>> Send bugs reports to this list.
>> >>>> To unsubscribe, send email sympa@lists.bath.ac.uk with body
>> >>>> "unsubscribe
>> >>>> csound"
>> >>>
>> >>> 
>> >>
>> >
>> >
>>
>>
>> Send bugs reports to this list.
>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>> csound"
>


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"