Hi
I am planning to use PD to make a simple demonstration if IIR and FIR filters. Well, the FIR part is quite easy to figure out, but for the IIR I need to know if there is any way that I can make a delay that will last only ONE sample (before I feed it back).
Any suggestions?
TIA
Soeren
Søren Bovbjerg wrote:
Hi
I am planning to use PD to make a simple demonstration if IIR and FIR filters. Well, the FIR part is quite easy to figure out, but for the IIR I need to know if there is any way that I can make a delay that will last only ONE sample (before I feed it back).
Any suggestions?
yep !
the zexy-external provides functions to convert signals to messages and back again. (pack~ & unpack~) this is kind of annoying, when there should be more delays
=--------- |block~1 |
unfortunately pd cannot reconstruct the signal from block~1 to block~64 so you have to do this manually : pE : via the pack~ & unpack~ objects
i attach a demo patch for both solutions
mfg.sfs.gtr IOhannes
|I am planning to use PD to make a simple demonstration if IIR and FIR filters. Well, the FIR part is quite easy to figure out, but for the IIR I need to know if there is any way that I can make a delay that will last only ONE sample (before I feed it back).
i think z~ from sexy ...
-- > < D V . o 7 g
Yep!
And thank you to all for your help!
(You know who you are)
/Soeren
----- Original Message ----- From: "_-¯-_" jdl@xdv.org To: "Søren Bovbjerg" bovbjerg@musik.auc.dk Cc: "PD-List" pd-list@iem.kug.ac.at Sent: Tuesday, March 20, 2001 7:12 PM Subject: Re: Single sample delay
|I am planning to use PD to make a simple demonstration if IIR and FIR filters. Well, the FIR part is quite easy to figure out, but for the IIR I need to know if there is any way that I can make a delay that will last only ONE sample (before I feed it back).
i think z~ from sexy ...
-- > < D V . o 7 g
-¯- wrote:
i think z~ from sexy ...
the main problem is not to delay a signal for ONE sample, but to do the feedback-loop. pd won't permit simple feedback loops due to its architecture. the best feedback you can get is delayed by at least one block (64 samples) via s~/r~.
mfg.dfsa.sdt IOhannes