Hello everyone.
A while ago, I was working on a modification to the fluid~ object which would allow multiple fluid~ instances to share soundfont images, rather than having to load multiple copies. My personal reason for this was that I have a large piano soundfont, and I was thinking of post-processing each individual note seperately, and one way of doing this would be to employ multiple fluid~'s and have a routing mechanism (like poly) in front of them.
I ran into a roadblock in the fluid API in the course of developing this, but now the fluid API has been enhanced to allow this.
My question is now, is anyone dying urgently for an updated fluid~ that allows this? If so, I'll take some time to send an update to Frank. If not, then when I get to it, I get to it...
Larry Troxler
On 27 Jun 2004, at 03:19, Larry Troxler wrote:
My question is now, is anyone dying urgently for an updated fluid~
that allows this? If so, I'll take some time to send an update to Frank. If not, then when I get to it, I get to it...
I'm not dying (not urgently anyway), but I could use a fluid~ with that functionality...if it's not too much hassle, I would welcome the update.
thanks,
d
Larry Troxler wrote:
A while ago, I was working on a modification to the fluid~ object which would allow multiple fluid~ instances to share soundfont images, rather than having to load multiple copies. My personal reason for this was that I have a large piano soundfont, and I was thinking of post-processing each individual note seperately, and one way of doing this would be to employ multiple fluid~'s and have a routing mechanism (like poly) in front of them.
I ran into a roadblock in the fluid API in the course of developing this, but now the fluid API has been enhanced to allow this.
I hadn't heard of Frank's fluid~ external until now. I ran to the Linux box to download and install it. I'm running pd-0.37.1 freshly compiled from Gentoo portage on an Athlon-XP. PD is running with no other external libraries and in jack mode. I run the fluid~ binary with DSP off...loads fine. I load a 200MB soundfount. Loads fine. I turn DSP on...
segfault...
I install flext from the sourceforge external collection and compile fluid~ myself. I run the freshly compiled version and...
segfault...
I check my fluidsynth version. 1.0.1, looks like there's an update to 1.0.3. I install the update and recompile fluid~. I run fluid~ the same as before and...
segfault...
running out of ideas. I'm a beginning C coder and a friend told me about strace for C debugging. I run strace on pd the same as above. Looks like it kills itself at this system call:
recv(4, "pd dsp 1 ;", 815, 0) = 10 ... GUI timing events send(4, "set pd_whichapi 5\n", 18, 0) = 18 ...GUI timing events send(4, "pdtk_pd_dsp ON\n", 15, 0) = 15 ...many more GUI timing events --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++
Frank and Larry, what system are you running/developing fluid~ on? You work around any problems to get it stable? I'd like to help debug this in any way I can because this would bring my entire live set up together into a PD patch.
Here's my detailed specs for the uber-curious: Athlon-XP 2200 1GB RAM Gentoo Linux (default-x86-1.4, gcc-3.3.2, glibc-2.3.2-r9, 2.4.23-ck1) Pd version 0.37.1 compiled 13:18:01 Jun 27 2004 fluidsynth 1.0.3 flext from pd-externals-20030311 jack-audio-connection-kit-0.98.1
Hallo, lee hat gesagt: // lee wrote:
I hadn't heard of Frank's fluid~ external until now. I ran to the Linux box to download and install it. I'm running pd-0.37.1 freshly compiled from Gentoo portage on an Athlon-XP. PD is running with no other external libraries and in jack mode. I run the fluid~ binary with DSP off...loads fine. I load a 200MB soundfount. Loads fine. I turn DSP on...
segfault...
Did you compile libfluidsynth with LADSPA enabled? Then try disabling LADSPA. There is a unsolved bug inside libfluidsynth's LADSPA code wich is triggerd by fluid~ in Pd.
You also can try building libfluidsynth with debug enabled and run Pd under gdb like "gdb /usr/bin/pd", then run it with "run", after the segfault, print a backtrace with "bt"
Frank Barknecht _ ______footils.org__
On Sunday 27 June 2004 16:22, lee wrote:
I hadn't heard of Frank's fluid~ external until now. I ran to the Linux box to download and install it. I'm running pd-0.37.1 freshly compiled from Gentoo portage on an Athlon-XP. PD is running with no other external libraries and in jack mode. I run the fluid~ binary with DSP off...loads fine. I load a 200MB soundfount. Loads fine. I turn DSP on...
segfault...
I'm running Linux Athlon, pd-0.36-devel, fluidsynth 1.0.3.
Have you tried loading the soundfont into the stand alone fluidsynth? What happens then?
Larry
Larry Troxler wrote:
On Sunday 27 June 2004 16:22, lee wrote:
I hadn't heard of Frank's fluid~ external until now. I ran to the Linux box to download and install it. I'm running pd-0.37.1 freshly compiled from Gentoo portage on an Athlon-XP. PD is running with no other external libraries and in jack mode. I run the fluid~ binary with DSP off...loads fine. I load a 200MB soundfount. Loads fine. I turn DSP on...
segfault...
I'm running Linux Athlon, pd-0.36-devel, fluidsynth 1.0.3.
Have you tried loading the soundfont into the stand alone fluidsynth? What happens then?
It sounds nice. I can play for hours. That's why I why I want to integrate this into the PD patch. I'll try Frank's debugging suggestions. I may have fluidsynth compiled with ladspa support. If not here comes the first venture into gdb...
-lee