Hi there,
I've been working on a patch which takes the output of an fft and "mixes" the real and imaginary parts with the fft data from another sample. It then resynthesizes for the output. It's sort of like a convolution, conceptually.
I have been having enormous amounts of trouble debugging the thing and am presently getting "Consistency Check signal_free 4" errors. I've checked, but can't seem to find any un-connected connections, which is, I believe, what causes that error. I'm using a block size of 8192 in the fft patches, and am both reading from and sending to a whole lot of arrays using tabreceive~ and tabsend~. I've enclosed the patch and all it's associated "stuff."
Can consistency check errors result from tabreceive~ and tabsend~ somehow? Do I have to be careful about where the arrays i'm tabsend~ing to are? Should I put block size objects in the patch that contains the arrays...?
Anyone who could lend some help?
Thanks a million, Jamie A
hi the consistency check errors can also come from env~ or dac~ objects (these objects usually use a vectorsize of 64, but you can specify the size by an argument...) "signal_free 4", i have never seen before. marius.
----- Original Message ----- From: "Jamie Cameron Allen" 0006685A@student.gla.ac.uk To: pd-list@iem.kug.ac.at Sent: Thursday, April 18, 2002 7:56 PM Subject: [PD] FFT Convolution Patch
Hi there,
I've been working on a patch which takes the output of an fft and "mixes" the real and imaginary parts with the fft data from another sample. It then resynthesizes for the output. It's sort of like a convolution, conceptually.
I have been having enormous amounts of trouble debugging the thing and am presently getting "Consistency Check signal_free 4" errors. I've checked, but can't seem to find any un-connected connections, which is, I believe, what causes that error. I'm using a block size of 8192 in the fft patches, and am both reading from and sending to a whole lot of arrays using tabreceive~ and tabsend~. I've enclosed the patch and all it's associated "stuff."
Can consistency check errors result from tabreceive~ and tabsend~ somehow? Do I have to be careful about where the arrays i'm tabsend~ing to are? Should I put block size objects in the patch that contains the arrays...?
Anyone who could lend some help?
Thanks a million, Jamie A
hi Marius
could you explain what you mean by specifying vectorsize? I have thought dac~ has this hardwired. I have also thought env~'s arguments are: the number of samples used for analysis, and the number of samples between successive output points?
It seems as if "signal_free 4" bug meant there is an attempt to free what has been freed already. Could you send me a patch showing this bug, if you have it handy?
Krzysztof
the consistency check errors can also come from env~ or dac~ objects (these objects usually use a vectorsize of 64, but you can specify the size by an argument...) "signal_free 4", i have never seen before.
oh dear, i was already sleeping, when i posted this comment... please just don´t talk about it anymore. actually i don´t know jamie´s error and have no idea how to solve it. marius.
hi Marius
could you explain what you mean by specifying vectorsize? I have thought dac~ has this hardwired. I have also thought env~'s arguments are: the number of samples used for analysis, and the number of samples between successive output points?
It seems as if "signal_free 4" bug meant there is an attempt to free what has been freed already. Could you send me a patch showing this bug, if you have it handy?
Krzysztof
the consistency check errors can also come from env~ or dac~
objects
(these objects usually use a vectorsize of 64, but you can specify
the
size by an argument...) "signal_free 4", i have never seen before.
-
Jamie Cameron Allen -
Krzysztof Czaja -
sme