OSX CsoundVST issues

classic Classic list List threaded Threaded
21 messages Options
12
Reply | Threaded
Open this post in threaded view
|

OSX CsoundVST issues

Victor Lazzarini
| At 15:31 27/10/2005, you wrote:
| What would need to be changed to build and run it on OS X?

I'm still looking at it, but I think the problem lies in the
way the same dynamic lib is used twice in the program. In other
systems, this is not a problem as the same type of dynamic lib
is used as an ordinary shared library and a loadable module.

The problem on OSX lies in the fact that, and that seems to be
a python-specific issue, python loadable modules are bundles,
and not shared libs.

There is no way that the same file can be used as a python module
and a dynamic-link library. So I tried creating two: this  generates two
separate instances of csound. Then, as Istvan suggested, I tried
linking the bundle to the dynamic library: python can't find the
entry point. Finally, moved the wrapper object code to the bundle
from the dynamic library. Now python can see the entry point,
but CsoundVST crashes. Gdb tells me that the program
received EXC_BAD_ACCESS, could not access memory (in
__dynamic_cast()).

When running python and then loading the module, gdb tells me
it is stuck in _semaphore_wait_signal_trap().

So I am not sure now, given what gdb is telling me, what problems
are there. But one thing is sure, the wrapper code works, because
if I build a bundle with all the csoundVST sources statically in it,
the module loads OK in Python.

I noticed that System.cpp was throwing a lot of warnings. One of them
was a linking problem of symbol redefinition (basename(), which is
found in libpthread and I renamed to basename_(), but that did not
change anything).

The answer to your question is probably I don't know. However,
you could consider making a CsoundVST program that is not a
python module, since there is now a python module in _csnd.so.
I'm not sure it would be what you want, though.

Victor



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Michael Gogins
My plan is for CsoundVST to load and use the csnd module as one Python extension, and the Silence algo-comp classes as another Python extension module. But, I may simply drop the Python support from CsoundVST and let it be purely a VST plugin. I prefer to code and compose in a Python code editor anyway.

So, do you think it would work to run Csound from Python importing "csnd" on OS X?

Thanks,
Mike

-----Original Message-----
From: Victor Lazzarini <[hidden email]>
Sent: Oct 27, 2005 11:03 AM
To: [hidden email]
Subject: [Cs-dev] OSX CsoundVST issues

| At 15:31 27/10/2005, you wrote:
| What would need to be changed to build and run it on OS X?

I'm still looking at it, but I think the problem lies in the
way the same dynamic lib is used twice in the program. In other
systems, this is not a problem as the same type of dynamic lib
is used as an ordinary shared library and a loadable module.

The problem on OSX lies in the fact that, and that seems to be
a python-specific issue, python loadable modules are bundles,
and not shared libs.

There is no way that the same file can be used as a python module
and a dynamic-link library. So I tried creating two: this  generates two
separate instances of csound. Then, as Istvan suggested, I tried
linking the bundle to the dynamic library: python can't find the
entry point. Finally, moved the wrapper object code to the bundle
from the dynamic library. Now python can see the entry point,
but CsoundVST crashes. Gdb tells me that the program
received EXC_BAD_ACCESS, could not access memory (in
__dynamic_cast()).

When running python and then loading the module, gdb tells me
it is stuck in _semaphore_wait_signal_trap().

So I am not sure now, given what gdb is telling me, what problems
are there. But one thing is sure, the wrapper code works, because
if I build a bundle with all the csoundVST sources statically in it,
the module loads OK in Python.

I noticed that System.cpp was throwing a lot of warnings. One of them
was a linking problem of symbol redefinition (basename(), which is
found in libpthread and I renamed to basename_(), but that did not
change anything).

The answer to your question is probably I don't know. However,
you could consider making a CsoundVST program that is not a
python module, since there is now a python module in _csnd.so.
I'm not sure it would be what you want, though.

Victor



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel





-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Victor Lazzarini
In reply to this post by Victor Lazzarini
I have tested python importing csnd on OSX and I could
run a script that Istvan suggested, and I got to play Trapped
alright. So that bit works. And if CsoundVST (lib) is built as a bundle,
I'd say it would also run OK, although I have not tested it.

As the code is at the moment, if CsoundVST (app) tries to load, as
a python interpreter, bundle containing a copy of the same object
code in CsoundVST (lib), then there will be two different instances
of csound in memory. CsoundVST crashes, but the crash might
be unrelated to that. So if you were to load the csnd module from
a python interpreter in CsoundVST, you might get the same problem.

So it is possible that a CsoundVST without python will work OK. I'm not
sure how you would want to provide this, but to me it seems that if
you want to work with python and algorithmic composition, you
can do it from python by importing csnd and the algorithmic composition
module. That would work on OSX. And if you want VST plugin and frontend,
then you just run CsoundVST.

Victor

At 16:46 27/10/2005, you wrote:

>My plan is for CsoundVST to load and use the csnd module as one Python
>extension, and the Silence algo-comp classes as another Python extension
>module. But, I may simply drop the Python support from CsoundVST and let
>it be purely a VST plugin. I prefer to code and compose in a Python code
>editor anyway.
>
>So, do you think it would work to run Csound from Python importing "csnd"
>on OS X?
>
>Thanks,
>Mike
>
>-----Original Message-----
>From: Victor Lazzarini <[hidden email]>
>Sent: Oct 27, 2005 11:03 AM
>To: [hidden email]
>Subject: [Cs-dev] OSX CsoundVST issues
>
>| At 15:31 27/10/2005, you wrote:
>| What would need to be changed to build and run it on OS X?
>
>I'm still looking at it, but I think the problem lies in the
>way the same dynamic lib is used twice in the program. In other
>systems, this is not a problem as the same type of dynamic lib
>is used as an ordinary shared library and a loadable module.
>
>The problem on OSX lies in the fact that, and that seems to be
>a python-specific issue, python loadable modules are bundles,
>and not shared libs.
>
>There is no way that the same file can be used as a python module
>and a dynamic-link library. So I tried creating two: this  generates two
>separate instances of csound. Then, as Istvan suggested, I tried
>linking the bundle to the dynamic library: python can't find the
>entry point. Finally, moved the wrapper object code to the bundle
>from the dynamic library. Now python can see the entry point,
>but CsoundVST crashes. Gdb tells me that the program
>received EXC_BAD_ACCESS, could not access memory (in
>__dynamic_cast()).
>
>When running python and then loading the module, gdb tells me
>it is stuck in _semaphore_wait_signal_trap().
>
>So I am not sure now, given what gdb is telling me, what problems
>are there. But one thing is sure, the wrapper code works, because
>if I build a bundle with all the csoundVST sources statically in it,
>the module loads OK in Python.
>
>I noticed that System.cpp was throwing a lot of warnings. One of them
>was a linking problem of symbol redefinition (basename(), which is
>found in libpthread and I renamed to basename_(), but that did not
>change anything).
>
>The answer to your question is probably I don't know. However,
>you could consider making a CsoundVST program that is not a
>python module, since there is now a python module in _csnd.so.
>I'm not sure it would be what you want, though.
>
>Victor
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Michael Gogins
In reply to this post by Victor Lazzarini
I'm open to suggestions here. One advantage of CsoundVST for Python work, is that it gives you the ability to stop a running performance by clicking on a button; CsoundVST bails out of the csoundPerformKsmps loop in that case.

I'd be quite interested to hear your thoughts -- or anyone's thoughts! -- on how Csound 5 can, will, or should be used. A wish list would be nice. My interests focus on writing compositions as code including both orchestra and score stuff, but I would like to know what people would like to see: is a GUI front end important, a DSSI plugin important, a VST plugin important, language interfaces, etc.

Thanks,
Mike

-----Original Message-----
From: Victor Lazzarini <[hidden email]>
Sent: Oct 27, 2005 12:14 PM
To: [hidden email]
Subject: Re: [Cs-dev] OSX CsoundVST issues

I have tested python importing csnd on OSX and I could
run a script that Istvan suggested, and I got to play Trapped
alright. So that bit works. And if CsoundVST (lib) is built as a bundle,
I'd say it would also run OK, although I have not tested it.

As the code is at the moment, if CsoundVST (app) tries to load, as
a python interpreter, bundle containing a copy of the same object
code in CsoundVST (lib), then there will be two different instances
of csound in memory. CsoundVST crashes, but the crash might
be unrelated to that. So if you were to load the csnd module from
a python interpreter in CsoundVST, you might get the same problem.

So it is possible that a CsoundVST without python will work OK. I'm not
sure how you would want to provide this, but to me it seems that if
you want to work with python and algorithmic composition, you
can do it from python by importing csnd and the algorithmic composition
module. That would work on OSX. And if you want VST plugin and frontend,
then you just run CsoundVST.

Victor

At 16:46 27/10/2005, you wrote:

>My plan is for CsoundVST to load and use the csnd module as one Python
>extension, and the Silence algo-comp classes as another Python extension
>module. But, I may simply drop the Python support from CsoundVST and let
>it be purely a VST plugin. I prefer to code and compose in a Python code
>editor anyway.
>
>So, do you think it would work to run Csound from Python importing "csnd"
>on OS X?
>
>Thanks,
>Mike
>
>-----Original Message-----
>From: Victor Lazzarini <[hidden email]>
>Sent: Oct 27, 2005 11:03 AM
>To: [hidden email]
>Subject: [Cs-dev] OSX CsoundVST issues
>
>| At 15:31 27/10/2005, you wrote:
>| What would need to be changed to build and run it on OS X?
>
>I'm still looking at it, but I think the problem lies in the
>way the same dynamic lib is used twice in the program. In other
>systems, this is not a problem as the same type of dynamic lib
>is used as an ordinary shared library and a loadable module.
>
>The problem on OSX lies in the fact that, and that seems to be
>a python-specific issue, python loadable modules are bundles,
>and not shared libs.
>
>There is no way that the same file can be used as a python module
>and a dynamic-link library. So I tried creating two: this  generates two
>separate instances of csound. Then, as Istvan suggested, I tried
>linking the bundle to the dynamic library: python can't find the
>entry point. Finally, moved the wrapper object code to the bundle
>from the dynamic library. Now python can see the entry point,
>but CsoundVST crashes. Gdb tells me that the program
>received EXC_BAD_ACCESS, could not access memory (in
>__dynamic_cast()).
>
>When running python and then loading the module, gdb tells me
>it is stuck in _semaphore_wait_signal_trap().
>
>So I am not sure now, given what gdb is telling me, what problems
>are there. But one thing is sure, the wrapper code works, because
>if I build a bundle with all the csoundVST sources statically in it,
>the module loads OK in Python.
>
>I noticed that System.cpp was throwing a lot of warnings. One of them
>was a linking problem of symbol redefinition (basename(), which is
>found in libpthread and I renamed to basename_(), but that did not
>change anything).
>
>The answer to your question is probably I don't know. However,
>you could consider making a CsoundVST program that is not a
>python module, since there is now a python module in _csnd.so.
>I'm not sure it would be what you want, though.
>
>Victor
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel





-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Steven Yi
Hi Michael,

Apologies but don't have too much time at the moment to comment, but I
would like to say that whatever requirements we gather up, it would be
*very* nice to record them and organize them into milestone points
that could lead up to a release.  Having milestones would also help
testing as we could say "these features were implemented for the M3
release; please focus testing on these features".  Besides the
milestones, it might also be worth keeping a queue of features for a
Csound 5.2 or 6.0.

steven

On 10/27/05, Michael Gogins <[hidden email]> wrote:
> I'm open to suggestions here. One advantage of CsoundVST for Python work, is that it gives you the ability to stop a running performance by clicking on a button; CsoundVST bails out of the csoundPerformKsmps loop in that case.
>
> I'd be quite interested to hear your thoughts -- or anyone's thoughts! -- on how Csound 5 can, will, or should be used. A wish list would be nice. My interests focus on writing compositions as code including both orchestra and score stuff, but I would like to know what people would like to see: is a GUI front end important, a DSSI plugin important, a VST plugin important, language interfaces, etc.
>
> Thanks,
> Mike
>


-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

jpff
In reply to this post by Michael Gogins
The main requirement I have is to have a command line program to which
I give an orchestra and score and generates a sound file.  It would be
nice if it could also generate sound in real time, and I could make
simple parameter adjustments to design the sound.

Secondary requirement is to be able to call csound from a C, C++ or
Java program to provide musical services, and those calls to control
parameters.  Actually I only want C, but the students seem to think
Java is useful.

Third requirement is that it should be easy to add new operations in C
to provide experimental operations.

==John ffitch


-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Victor Lazzarini
The Java wrappers that Michael added seem to be building OK. So I reckon
a Java program can be written as a frontend to Csound. It would be nice to
have a Java programmer to take that on board and provide an example frontend.

As far as I can see we can do the following things on a multiplatform basis:

1. libcsound C library
2. csound 'classic' command-line frontend
3. csoundapi~ PD class frontend
4. tclcsound frontend (cstclsh, cswish and  tclcsound extension module)
5. csnd python module
6. csnd.jar java module

and possibly (not tested on OSX): 7. flcsound, lua module

On Linux and Windows only:

8. CsoundVST

So both your primary and secondary requirements seem to be there.
Can you elaborate on your third?

I would add another one to the list: possibility of loading, initialising
and running opcodes from the API only.

Victor

At 08:12 28/10/2005, you wrote:
>Secondary requirement is to be able to call csound from a C, C++ or
>Java program to provide musical services, and those calls to control
>parameters.  Actually I only want C, but the students seem to think
>Java is useful.

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Victor Lazzarini
In reply to this post by Victor Lazzarini
I think that we should provide flexible ways of interacting with
csound. I see that this is already happening. Another important
thing is to be multiplatform, and provide an unified behaviour across
platforms.

In relation to python, is it not possible to write a python script that
will give you a GUI and a button to press, to mimic what CsoundVST
does?

It might be worth, in the long run, to provide CsoundVST as a VST plugin/
standalone frontend only, to ensure it runs on all platforms. That does not
rule out providing a separate windows or linux-only program that does
everything,
but as something separate from the 'canonical' csound 5 distribution.
However I understand you might not want to have to maintain two separate
projects.

Victor

At 18:14 27/10/2005, you wrote:

>I'm open to suggestions here. One advantage of CsoundVST for Python work,
>is that it gives you the ability to stop a running performance by clicking
>on a button; CsoundVST bails out of the csoundPerformKsmps loop in that case.
>
>I'd be quite interested to hear your thoughts -- or anyone's thoughts! --
>on how Csound 5 can, will, or should be used. A wish list would be nice.
>My interests focus on writing compositions as code including both
>orchestra and score stuff, but I would like to know what people would like
>to see: is a GUI front end important, a DSSI plugin important, a VST
>plugin important, language interfaces, etc.
>
>Thanks,
>Mike
>
>-----Original Message-----
>From: Victor Lazzarini <[hidden email]>
>Sent: Oct 27, 2005 12:14 PM
>To: [hidden email]
>Subject: Re: [Cs-dev] OSX CsoundVST issues
>
>I have tested python importing csnd on OSX and I could
>run a script that Istvan suggested, and I got to play Trapped
>alright. So that bit works. And if CsoundVST (lib) is built as a bundle,
>I'd say it would also run OK, although I have not tested it.
>
>As the code is at the moment, if CsoundVST (app) tries to load, as
>a python interpreter, bundle containing a copy of the same object
>code in CsoundVST (lib), then there will be two different instances
>of csound in memory. CsoundVST crashes, but the crash might
>be unrelated to that. So if you were to load the csnd module from
>a python interpreter in CsoundVST, you might get the same problem.
>
>So it is possible that a CsoundVST without python will work OK. I'm not
>sure how you would want to provide this, but to me it seems that if
>you want to work with python and algorithmic composition, you
>can do it from python by importing csnd and the algorithmic composition
>module. That would work on OSX. And if you want VST plugin and frontend,
>then you just run CsoundVST.
>
>Victor
>
>At 16:46 27/10/2005, you wrote:
> >My plan is for CsoundVST to load and use the csnd module as one Python
> >extension, and the Silence algo-comp classes as another Python extension
> >module. But, I may simply drop the Python support from CsoundVST and let
> >it be purely a VST plugin. I prefer to code and compose in a Python code
> >editor anyway.
> >
> >So, do you think it would work to run Csound from Python importing "csnd"
> >on OS X?
> >
> >Thanks,
> >Mike
> >
> >-----Original Message-----
> >From: Victor Lazzarini <[hidden email]>
> >Sent: Oct 27, 2005 11:03 AM
> >To: [hidden email]
> >Subject: [Cs-dev] OSX CsoundVST issues
> >
> >| At 15:31 27/10/2005, you wrote:
> >| What would need to be changed to build and run it on OS X?
> >
> >I'm still looking at it, but I think the problem lies in the
> >way the same dynamic lib is used twice in the program. In other
> >systems, this is not a problem as the same type of dynamic lib
> >is used as an ordinary shared library and a loadable module.
> >
> >The problem on OSX lies in the fact that, and that seems to be
> >a python-specific issue, python loadable modules are bundles,
> >and not shared libs.
> >
> >There is no way that the same file can be used as a python module
> >and a dynamic-link library. So I tried creating two: this  generates two
> >separate instances of csound. Then, as Istvan suggested, I tried
> >linking the bundle to the dynamic library: python can't find the
> >entry point. Finally, moved the wrapper object code to the bundle
> >from the dynamic library. Now python can see the entry point,
> >but CsoundVST crashes. Gdb tells me that the program
> >received EXC_BAD_ACCESS, could not access memory (in
> >__dynamic_cast()).
> >
> >When running python and then loading the module, gdb tells me
> >it is stuck in _semaphore_wait_signal_trap().
> >
> >So I am not sure now, given what gdb is telling me, what problems
> >are there. But one thing is sure, the wrapper code works, because
> >if I build a bundle with all the csoundVST sources statically in it,
> >the module loads OK in Python.
> >
> >I noticed that System.cpp was throwing a lot of warnings. One of them
> >was a linking problem of symbol redefinition (basename(), which is
> >found in libpthread and I renamed to basename_(), but that did not
> >change anything).
> >
> >The answer to your question is probably I don't know. However,
> >you could consider making a CsoundVST program that is not a
> >python module, since there is now a python module in _csnd.so.
> >I'm not sure it would be what you want, though.
> >
> >Victor
> >
> >
> >
> >-------------------------------------------------------
> >This SF.Net email is sponsored by the JBoss Inc.
> >Get Certified Today * Register for a JBoss Training Course
> >Free Certification Exam for All Training Attendees Through End of 2005
> >Visit http://www.jboss.com/services/certification for more information
> >_______________________________________________
> >Csound-devel mailing list
> >[hidden email]
> >https://lists.sourceforge.net/lists/listinfo/csound-devel
> >
> >
> >
> >
> >
> >-------------------------------------------------------
> >This SF.Net email is sponsored by the JBoss Inc.
> >Get Certified Today * Register for a JBoss Training Course
> >Free Certification Exam for All Training Attendees Through End of 2005
> >Visit http://www.jboss.com/services/certification for more information
> >_______________________________________________
> >Csound-devel mailing list
> >[hidden email]
> >https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>Victor Lazzarini
>Music Technology Laboratory
>Music Department
>National University of Ireland, Maynooth
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Istvan Varga
In reply to this post by Victor Lazzarini
Victor Lazzarini wrote:

> I would add another one to the list: possibility of loading, initialising
> and running opcodes from the API only.

This may be tricky as the opcodes depend on a global environment, that is,
various members of the CSOUND structure, and expect the presence of a parent
instrument instance (for run-time parameters like the release flag, MIDI
channel, etc.) and instrument text (for compile time parameters like the
count, name, and type of arguments - a common example is the use of the
p->INOCOUNT and similar macros). So, to run an opcode directly through the
API, a fake environment with an instrument and instrument instance needs
to be set up, as well as several variables in CSOUND should be initialized;
this mode of operation probably cannot be mixed with normal performance on
the same Csound instance.
Nevertheless, some simple opcodes may work without problems (obviously,
there are many that do not make sense, like goto and anything related to
instument control).



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

opcode/instrument interface (was: CsoundVST OSX issues)

Victor Lazzarini
This is starting to sound as if, although tricky, might be
possible. Considering we avoid mixing the normal operation
mode and this one, in other words, have a dedicated instance
for this, would it be possible then to:

1. Create a fake instrument
2. load opcodes and organise the synthesis chain (so-called 'graph')
3. run performKsmps() [with a dummy score of say f0 1000)
4. send in realtime events.

?

Victor

At 09:28 28/10/2005, you wrote:

>This may be tricky as the opcodes depend on a global environment, that is,
>various members of the CSOUND structure, and expect the presence of a parent
>instrument instance (for run-time parameters like the release flag, MIDI
>channel, etc.) and instrument text (for compile time parameters like the
>count, name, and type of arguments - a common example is the use of the
>p->INOCOUNT and similar macros). So, to run an opcode directly through the
>API, a fake environment with an instrument and instrument instance needs
>to be set up, as well as several variables in CSOUND should be initialized;
>this mode of operation probably cannot be mixed with normal performance on
>the same Csound instance.
>Nevertheless, some simple opcodes may work without problems (obviously,
>there are many that do not make sense, like goto and anything related to
>instument control).

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: opcode/instrument interface

Istvan Varga
Victor Lazzarini wrote:

> This is starting to sound as if, although tricky, might be
> possible. Considering we avoid mixing the normal operation
> mode and this one, in other words, have a dedicated instance
> for this, would it be possible then to:
>
> 1. Create a fake instrument
> 2. load opcodes and organise the synthesis chain (so-called 'graph')
> 3. run performKsmps() [with a dummy score of say f0 1000)
> 4. send in realtime events.

This is not exactly what I thought of, I assumed that the new interface
would only allow for calling a single opcode at once to do some signal
processing. A chain of multiple opcodes and sensing real-time events is
a different issue as it basically means building a full instrument and
then performing it with csoundPerformKsmps(); in this case you could
probably just build an actual orchestra file (for example in a high level
language like Python) and perform that.


-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: opcode/instrument interface

Victor Lazzarini
OK. let me reformulate. Loading single opcodes and running
is OK. But also what I had in mind was the possibility of
adding instruments to an already loaded orchestra, even
as csound performs other instruments.

I suppose it doesn't really matter if the instrument code is
actually written in Csound, because you can construct
the text code for such instruments using other languages;
but what I was thinking was to have API functions that could do that, so
that we
could build an instrument only using the API, bypassing
the csound orchestra language.

Victor


At 10:21 28/10/2005, you wrote:

>Victor Lazzarini wrote:
>
>>This is starting to sound as if, although tricky, might be
>>possible. Considering we avoid mixing the normal operation
>>mode and this one, in other words, have a dedicated instance
>>for this, would it be possible then to:
>>1. Create a fake instrument
>>2. load opcodes and organise the synthesis chain (so-called 'graph')
>>3. run performKsmps() [with a dummy score of say f0 1000)
>>4. send in realtime events.
>
>This is not exactly what I thought of, I assumed that the new interface
>would only allow for calling a single opcode at once to do some signal
>processing. A chain of multiple opcodes and sensing real-time events is
>a different issue as it basically means building a full instrument and
>then performing it with csoundPerformKsmps(); in this case you could
>probably just build an actual orchestra file (for example in a high level
>language like Python) and perform that.
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: opcode/instrument interface

Istvan Varga
Victor Lazzarini wrote:

> OK. let me reformulate. Loading single opcodes and running
> is OK. But also what I had in mind was the possibility of
> adding instruments to an already loaded orchestra, even
> as csound performs other instruments.

I do not think this is possible without major changes. Unfortunately
most of the relevant code was designed and written assuming that you
load and compile (in multiple passes) an orchestra before performance,
without allowing for incremental updates later.

> but what I was thinking was to have API functions that could do that, so
> that we could build an instrument only using the API, bypassing
> the csound orchestra language.

But is this really much easier than just building an actual orchestra
file, particularly given that building an instrument at a low level,
bypassing the orchestra language, assumes knowledge of Csound internals
that may possibly even change in future versions ?


-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: opcode/instrument interface

Victor Lazzarini
Probably not.

At 10:54 28/10/2005, you wrote:
>But is this really much easier than just building an actual orchestra
>file, particularly given that building an instrument at a low level,
>bypassing the orchestra language, assumes knowledge of Csound internals
>that may possibly even change in future versions ?

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Steven Yi
In reply to this post by Victor Lazzarini
Hi Victor,

On 10/28/05, Victor Lazzarini <[hidden email]> wrote:
> The Java wrappers that Michael added seem to be building OK. So I reckon
> a Java program can be written as a frontend to Csound. It would be nice to
> have a Java programmer to take that on board and provide an example frontend.

Well, I think I could put one together without too much trouble.
It'll be good for me to try out anyways as I'll eventually be trying
out integrating with blue sometime as a sort of "if csnd.jar exists,
enable these other features" option.  Just need to be able to compile
all this on Windows.  Shall I put something together and put it in the
frontends directory?

steven


-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Victor Lazzarini
Yes please!

Victor
At 13:40 28/10/2005, you wrote:

>Hi Victor,
>
>On 10/28/05, Victor Lazzarini <[hidden email]> wrote:
> > The Java wrappers that Michael added seem to be building OK. So I reckon
> > a Java program can be written as a frontend to Csound. It would be nice to
> > have a Java programmer to take that on board and provide an example
> frontend.
>
>Well, I think I could put one together without too much trouble.
>It'll be good for me to try out anyways as I'll eventually be trying
>out integrating with blue sometime as a sort of "if csnd.jar exists,
>enable these other features" option.  Just need to be able to compile
>all this on Windows.  Shall I put something together and put it in the
>frontends directory?
>
>steven
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Michael Gogins
In reply to this post by Victor Lazzarini
There is (was) a Java frontend in the CsoundVST jar, which I am in process of moving to the interfaces directory. Or it could be a separate thing in the frontends directory.

Regards,
Mike

-----Original Message-----
From: Victor Lazzarini <[hidden email]>
Sent: Oct 28, 2005 4:11 AM
To: [hidden email]
Subject: Re: [Cs-dev] OSX CsoundVST issues

The Java wrappers that Michael added seem to be building OK. So I reckon
a Java program can be written as a frontend to Csound. It would be nice to
have a Java programmer to take that on board and provide an example frontend.

As far as I can see we can do the following things on a multiplatform basis:

1. libcsound C library
2. csound 'classic' command-line frontend
3. csoundapi~ PD class frontend
4. tclcsound frontend (cstclsh, cswish and  tclcsound extension module)
5. csnd python module
6. csnd.jar java module

and possibly (not tested on OSX): 7. flcsound, lua module

On Linux and Windows only:

8. CsoundVST

So both your primary and secondary requirements seem to be there.
Can you elaborate on your third?

I would add another one to the list: possibility of loading, initialising
and running opcodes from the API only.

Victor

At 08:12 28/10/2005, you wrote:
>Secondary requirement is to be able to call csound from a C, C++ or
>Java program to provide musical services, and those calls to control
>parameters.  Actually I only want C, but the students seem to think
>Java is useful.

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel





-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Michael Gogins
In reply to this post by Victor Lazzarini
In general, I think this kind of requirement leaves the scope of Csound and enters the scope of a new software synthesis system.

Recently, James McCartney, the author of SuperCollider, expressed a wish that he had not designed a new language from scratch, but had embedded his synthesis engine into, and made his synthesis engine programmable in, another, more standard, more widely used high-level language.

In my view, this could now be done in a widely used, object-oriented scripting language such as Scheme, Python, or Lua with a DSP graph in C++ and unit generators from the STK, CLAM or some other usable C++ synthesis class library wrapped up in the scripting language. My experience suggests that getting this to work well is about 6 to 12 man-months of work. I currently am working on a new version using the STK in either Python or Lua, but I also plan to take a close look at Scheme. Lua is attractive because it is reasonably fast and the whole language can be embedded in a single small C library. Lua also has a working just-in-time compiler for Intel assembler!

I am not certain whether I will finish this stuff -- I have too much to do in other areas. But it's fun to poke around in.

For a sample of this approach, see the presentation at the recent Lua conference, "Audio Scripting with Lua" (google should find it).

At various times I have coded parts or most of such a new system, but I was never really happy about enough of the details. The basic concepts have to be worked out extremely well. I have a new respect for Barry Vercoe, Paul Lansky, James McCartney, Bill Shottstaedt, Perry Cook, Miller S. Puckette, and the others who have developed usable systems of this sort.

The real problem with doing something new like this is leaving behind all the orcs and scos, the community, the user-defined opcodes, the textbooks, the classes, etc., etc., etc.

Still, by building on experience and using carefully chosen open source foundations, it should be possible to create something that is simpler, more flexible, faster, and easier to build than any of the existing systems.

Regards,
Mike

-----Original Message-----
From: Istvan Varga <[hidden email]>
Sent: Oct 28, 2005 4:28 AM
To: [hidden email]
Subject: Re: [Cs-dev] OSX CsoundVST issues

Victor Lazzarini wrote:

> I would add another one to the list: possibility of loading, initialising
> and running opcodes from the API only.

This may be tricky as the opcodes depend on a global environment, that is,
various members of the CSOUND structure, and expect the presence of a parent
instrument instance (for run-time parameters like the release flag, MIDI
channel, etc.) and instrument text (for compile time parameters like the
count, name, and type of arguments - a common example is the use of the
p->INOCOUNT and similar macros). So, to run an opcode directly through the
API, a fake environment with an instrument and instrument instance needs
to be set up, as well as several variables in CSOUND should be initialized;
this mode of operation probably cannot be mixed with normal performance on
the same Csound instance.
Nevertheless, some simple opcodes may work without problems (obviously,
there are many that do not make sense, like goto and anything related to
instument control).



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel





-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: OSX CsoundVST issues

Michael Gogins
In reply to this post by Victor Lazzarini
It is possible and indeed rather easy to write a GUI for Csound in Python, absolutely, but it is not so easy to get to run as a VST plugin.

I completely agree with you about cross-platform support.

Another possibility is to simply drop the VST plugin support entirely. Is anyone actually using that? When I've played with it, I've really, really liked it. You really can sit in Cubase and score for a running Csound orchestras in regular music notation. Cool!

But unfortunately, I don't compose by scoring, usually.

On the other hand, surely it is possible to get CsoundVST working on OS X.

Regards,
Mike



-----Original Message-----
From: Victor Lazzarini <[hidden email]>
Sent: Oct 28, 2005 4:23 AM
To: [hidden email]
Subject: Re: [Cs-dev] OSX CsoundVST issues

I think that we should provide flexible ways of interacting with
csound. I see that this is already happening. Another important
thing is to be multiplatform, and provide an unified behaviour across
platforms.

In relation to python, is it not possible to write a python script that
will give you a GUI and a button to press, to mimic what CsoundVST
does?

It might be worth, in the long run, to provide CsoundVST as a VST plugin/
standalone frontend only, to ensure it runs on all platforms. That does not
rule out providing a separate windows or linux-only program that does
everything,
but as something separate from the 'canonical' csound 5 distribution.
However I understand you might not want to have to maintain two separate
projects.

Victor

At 18:14 27/10/2005, you wrote:

>I'm open to suggestions here. One advantage of CsoundVST for Python work,
>is that it gives you the ability to stop a running performance by clicking
>on a button; CsoundVST bails out of the csoundPerformKsmps loop in that case.
>
>I'd be quite interested to hear your thoughts -- or anyone's thoughts! --
>on how Csound 5 can, will, or should be used. A wish list would be nice.
>My interests focus on writing compositions as code including both
>orchestra and score stuff, but I would like to know what people would like
>to see: is a GUI front end important, a DSSI plugin important, a VST
>plugin important, language interfaces, etc.
>
>Thanks,
>Mike
>
>-----Original Message-----
>From: Victor Lazzarini <[hidden email]>
>Sent: Oct 27, 2005 12:14 PM
>To: [hidden email]
>Subject: Re: [Cs-dev] OSX CsoundVST issues
>
>I have tested python importing csnd on OSX and I could
>run a script that Istvan suggested, and I got to play Trapped
>alright. So that bit works. And if CsoundVST (lib) is built as a bundle,
>I'd say it would also run OK, although I have not tested it.
>
>As the code is at the moment, if CsoundVST (app) tries to load, as
>a python interpreter, bundle containing a copy of the same object
>code in CsoundVST (lib), then there will be two different instances
>of csound in memory. CsoundVST crashes, but the crash might
>be unrelated to that. So if you were to load the csnd module from
>a python interpreter in CsoundVST, you might get the same problem.
>
>So it is possible that a CsoundVST without python will work OK. I'm not
>sure how you would want to provide this, but to me it seems that if
>you want to work with python and algorithmic composition, you
>can do it from python by importing csnd and the algorithmic composition
>module. That would work on OSX. And if you want VST plugin and frontend,
>then you just run CsoundVST.
>
>Victor
>
>At 16:46 27/10/2005, you wrote:
> >My plan is for CsoundVST to load and use the csnd module as one Python
> >extension, and the Silence algo-comp classes as another Python extension
> >module. But, I may simply drop the Python support from CsoundVST and let
> >it be purely a VST plugin. I prefer to code and compose in a Python code
> >editor anyway.
> >
> >So, do you think it would work to run Csound from Python importing "csnd"
> >on OS X?
> >
> >Thanks,
> >Mike
> >
> >-----Original Message-----
> >From: Victor Lazzarini <[hidden email]>
> >Sent: Oct 27, 2005 11:03 AM
> >To: [hidden email]
> >Subject: [Cs-dev] OSX CsoundVST issues
> >
> >| At 15:31 27/10/2005, you wrote:
> >| What would need to be changed to build and run it on OS X?
> >
> >I'm still looking at it, but I think the problem lies in the
> >way the same dynamic lib is used twice in the program. In other
> >systems, this is not a problem as the same type of dynamic lib
> >is used as an ordinary shared library and a loadable module.
> >
> >The problem on OSX lies in the fact that, and that seems to be
> >a python-specific issue, python loadable modules are bundles,
> >and not shared libs.
> >
> >There is no way that the same file can be used as a python module
> >and a dynamic-link library. So I tried creating two: this  generates two
> >separate instances of csound. Then, as Istvan suggested, I tried
> >linking the bundle to the dynamic library: python can't find the
> >entry point. Finally, moved the wrapper object code to the bundle
> >from the dynamic library. Now python can see the entry point,
> >but CsoundVST crashes. Gdb tells me that the program
> >received EXC_BAD_ACCESS, could not access memory (in
> >__dynamic_cast()).
> >
> >When running python and then loading the module, gdb tells me
> >it is stuck in _semaphore_wait_signal_trap().
> >
> >So I am not sure now, given what gdb is telling me, what problems
> >are there. But one thing is sure, the wrapper code works, because
> >if I build a bundle with all the csoundVST sources statically in it,
> >the module loads OK in Python.
> >
> >I noticed that System.cpp was throwing a lot of warnings. One of them
> >was a linking problem of symbol redefinition (basename(), which is
> >found in libpthread and I renamed to basename_(), but that did not
> >change anything).
> >
> >The answer to your question is probably I don't know. However,
> >you could consider making a CsoundVST program that is not a
> >python module, since there is now a python module in _csnd.so.
> >I'm not sure it would be what you want, though.
> >
> >Victor
> >
> >
> >
> >-------------------------------------------------------
> >This SF.Net email is sponsored by the JBoss Inc.
> >Get Certified Today * Register for a JBoss Training Course
> >Free Certification Exam for All Training Attendees Through End of 2005
> >Visit http://www.jboss.com/services/certification for more information
> >_______________________________________________
> >Csound-devel mailing list
> >[hidden email]
> >https://lists.sourceforge.net/lists/listinfo/csound-devel
> >
> >
> >
> >
> >
> >-------------------------------------------------------
> >This SF.Net email is sponsored by the JBoss Inc.
> >Get Certified Today * Register for a JBoss Training Course
> >Free Certification Exam for All Training Attendees Through End of 2005
> >Visit http://www.jboss.com/services/certification for more information
> >_______________________________________________
> >Csound-devel mailing list
> >[hidden email]
> >https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>Victor Lazzarini
>Music Technology Laboratory
>Music Department
>National University of Ireland, Maynooth
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel





-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
Reply | Threaded
Open this post in threaded view
|

Re: opcode/instrument interface

Michael Gogins
In reply to this post by Istvan Varga
This again would reproduce the situation in SuperCollider of having 3 languages: a composing language (SC3), a signal graph language (SynthDefs), and a unit generator language (C). Instead we would have an orchestra language (e.g. Python), a signal graph language (Csound orc language), and a unit generator language (C).

The fewer languages, the better.

Regards,
Mike



-----Original Message-----
From: Victor Lazzarini <[hidden email]>
Sent: Oct 28, 2005 5:33 AM
To: [hidden email]
Subject: Re: [Cs-dev] opcode/instrument interface

OK. let me reformulate. Loading single opcodes and running
is OK. But also what I had in mind was the possibility of
adding instruments to an already loaded orchestra, even
as csound performs other instruments.

I suppose it doesn't really matter if the instrument code is
actually written in Csound, because you can construct
the text code for such instruments using other languages;
but what I was thinking was to have API functions that could do that, so
that we
could build an instrument only using the API, bypassing
the csound orchestra language.

Victor


At 10:21 28/10/2005, you wrote:

>Victor Lazzarini wrote:
>
>>This is starting to sound as if, although tricky, might be
>>possible. Considering we avoid mixing the normal operation
>>mode and this one, in other words, have a dedicated instance
>>for this, would it be possible then to:
>>1. Create a fake instrument
>>2. load opcodes and organise the synthesis chain (so-called 'graph')
>>3. run performKsmps() [with a dummy score of say f0 1000)
>>4. send in realtime events.
>
>This is not exactly what I thought of, I assumed that the new interface
>would only allow for calling a single opcode at once to do some signal
>processing. A chain of multiple opcodes and sensing real-time events is
>a different issue as it basically means building a full instrument and
>then performing it with csoundPerformKsmps(); in this case you could
>probably just build an actual orchestra file (for example in a high level
>language like Python) and perform that.
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by the JBoss Inc.
>Get Certified Today * Register for a JBoss Training Course
>Free Certification Exam for All Training Attendees Through End of 2005
>Visit http://www.jboss.com/services/certification for more information
>_______________________________________________
>Csound-devel mailing list
>[hidden email]
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel





-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.
Get Certified Today * Register for a JBoss Training Course
Free Certification Exam for All Training Attendees Through End of 2005
Visit http://www.jboss.com/services/certification for more information
_______________________________________________
Csound-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/csound-devel
12