John,

is it possible that your changes to allow for different in and out  
channel numbers have introduced a bug in the calculation of buffersizes?
In rtpa.c now, the values of

pabs->inParm.bufSamp_SW  and pabs->outParm.bufSamp_SW

are not matching anymore, raising an error. I ran 'trapped.csd' with - 
odac -iadc  and checked the values of these and
the input buffer is always 1/2 size of the output. Now because  
nchnls=2, I am wondering if the code is taking input nchnls as 1
by default, when it should take the value of nchnls instead.

This is the reason for the reported errors with portaudio IO in both  
Windows and OSX. Instead of me taking a long time tracing this, I
expect you would know where this is going wrong more promptly.

Victor

------------------------------------------------------------------------------

_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

>>>>> "Victor" == Victor Lazzarini  writes:

 Victor> John,
 Victor> is it possible that your changes to allow for different in and out  
 Victor> channel numbers have introduced a bug in the calculation of buffersizes?
 Victor> In rtpa.c now, the values of

 pabs-> inParm.bufSamp_SW  and pabs->outParm.bufSamp_SW

 Victor> are not matching anymore, raising an error. I ran 'trapped.csd' with - 
 Victor> odac -iadc  and checked the values of these and
 Victor> the input buffer is always 1/2 size of the output. Now because  
 Victor> nchnls=2, I am wondering if the code is taking input nchnls as 1
 Victor> by default, when it should take the value of nchnls instead.

 Victor> This is the reason for the reported errors with portaudio IO in both  
 Victor> Windows and OSX. Instead of me taking a long time tracing this, I
 Victor> expect you would know where this is going wrong more promptly.

Possible, but certainly not the intent.  Will add to the list of
things to do, but marking has the highest priority at present as i am
already late on the deadline

==John ffitch

------------------------------------------------------------------------------

_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

This is fixed now in CVS.

Victor
On 21 May 2010, at 18:42, Victor Lazzarini wrote:

> John,
>
> is it possible that your changes to allow for different in and out
> channel numbers have introduced a bug in the calculation of  
> buffersizes?
> In rtpa.c now, the values of
>
> pabs->inParm.bufSamp_SW  and pabs->outParm.bufSamp_SW
>
> are not matching anymore, raising an error. I ran 'trapped.csd' with -
> odac -iadc  and checked the values of these and
> the input buffer is always 1/2 size of the output. Now because
> nchnls=2, I am wondering if the code is taking input nchnls as 1
> by default, when it should take the value of nchnls instead.
>
> This is the reason for the reported errors with portaudio IO in both
> Windows and OSX. Instead of me taking a long time tracing this, I
> expect you would know where this is going wrong more promptly.
>
> Victor
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------

_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net