Hi Aaron,

I am getting a crash here (MacOS 9) in the sfont library when one of the
soundfont opcode's init-time function is called from the MIDIinsert()
function in Csound.  I cannot tell which one or see anymore details because
I do not have a debug version of sfontlib built.

However, the crash goes away if I move the "sfload" opcode out of instr 1 up
into the header section of the orchestra.  So, give that a try and maybe one
of the other developers can figure out if there is a problem that needs to
be fixed.  (My best guess is that the sfont is being repeatedly loaded by
each note instance until something explodes?  Does it use ldmemfile which
should avoid that?)

Anthony Kozar
mailing-lists-1001 AT anthonykozar DOT net
http://anthonykozar.net/

PS.  I was listening to some of your microtonal music again this week via a
podcast by Prent Rodgers.  Very illuminating -- thanks :)

aaron@akjmusic.com wrote on 1/21/08 5:38 PM:

> I have attempted to play the following .csd file (in version 5.07) and
> after a couple of seconds it always segfaults on me. The same buffer
> and MIDI settings are fine on other instruments I have designed, but
> for some reason, this soundfont instrument gives me trouble.

Hi,

El mar, 22-01-2008 a las 19:45 -0500, Anthony Kozar escribió:

> However, the crash goes away if I move the "sfload" opcode out of instr 1 up
> into the header section of the orchestra.  So, give that a try and maybe one
> of the other developers can figure out if there is a problem that needs to
> be fixed.  (My best guess is that the sfont is being repeatedly loaded by
> each note instance until something explodes?  Does it use ldmemfile which
> should avoid that?)
> 

Is this the solution? Please confirm so I can add a note in the manual.
(Maybe the opcode should also be changed so that it only runs at init
time?)

Cheers,
Andrés

Sorry, I have not had the time to chase this down and suspect no one else
has either.  So, I do not know if it is the solution but I have always seen
sfload used in the header section, not an instr block.

I think the opcode probably does only run at init time, but I was
speculating that each new note would load the soundfont again.  If so, this
probably should be fixed somehow ...

Anthony

Andres Cabrera wrote on 1/28/08 3:16 PM:

> El mar, 22-01-2008 a las 19:45 -0500, Anthony Kozar escribió:
> 
>> However, the crash goes away if I move the "sfload" opcode out of instr 1 up
>> into the header section of the orchestra.  So, give that a try and maybe one
>> of the other developers can figure out if there is a problem that needs to
>> be fixed.  (My best guess is that the sfont is being repeatedly loaded by
>> each note instance until something explodes?  Does it use ldmemfile which
>> should avoid that?)
>> 
> 
> Is this the solution? Please confirm so I can add a note in the manual.
> (Maybe the opcode should also be changed so that it only runs at init
> time?)