[Csnd] pvs fftsize vs winsize questions

Date	2013-03-25 22:22
From	Ben Hackbarth
Subject	[Csnd] pvs fftsize vs winsize questions
	hello, i am a little confused about the pvs initialization parameters. from the manual of pvsanal: fsig pvsanal ain, ifftsize, ioverlap, iwinsize, iwintype [, iformat] [, iinit] it seems to me that, in reading the manual entry, ifftsize is the number of samples taken from the audio stream and iwinsize is the size of the fft. in other words, ifftsize dictates the number of audio samples used while iwinsize affects the frequency resolution of the fft analysis (presumably buffered by zeros if larger than ifftsize). is this a correct reading of the parameters? this nomenclature seems backwards to me -- shouldn't ifftsize affect the bin resolution while iwinsize dictate the number of audio samples used per frame? in any case, i cannot seem to get these parameters to function properly in tadem. with the following code: fsig pvsanal ain, 4096, 1024, 8192, 1 kflag pvsftw fsig, ifna, ifnf ftsavek "/Users/ben/Desktop/tab.txt", kflag, 1, ifnf i get a table with only 2048 frequency values, not 4096 as i would expect from a winsize of 8192 (divided by two). any guidance is much appreciated! -- ben

Date	2013-04-01 12:48
From	Ben Hackbarth
Subject	[Csnd] Re: pvs fftsize vs winsize questions
	bump. can anyone offer any insight? thank you in advance. -- ben On Mon, Mar 25, 2013 at 11:22 PM, Ben Hackbarth <hackbarth@gmail.com> wrote: hello, i am a little confused about the pvs initialization parameters. from the manual of pvsanal: fsig pvsanal ain, ifftsize, ioverlap, iwinsize, iwintype [, iformat] [, iinit] it seems to me that, in reading the manual entry, ifftsize is the number of samples taken from the audio stream and iwinsize is the size of the fft. in other words, ifftsize dictates the number of audio samples used while iwinsize affects the frequency resolution of the fft analysis (presumably buffered by zeros if larger than ifftsize). is this a correct reading of the parameters? this nomenclature seems backwards to me -- shouldn't ifftsize affect the bin resolution while iwinsize dictate the number of audio samples used per frame? in any case, i cannot seem to get these parameters to function properly in tadem. with the following code: fsig pvsanal ain, 4096, 1024, 8192, 1 kflag pvsftw fsig, ifna, ifnf ftsavek "/Users/ben/Desktop/tab.txt", kflag, 1, ifnf i get a table with only 2048 frequency values, not 4096 as i would expect from a winsize of 8192 (divided by two). any guidance is much appreciated! -- ben

Date	2013-04-01 12:51
From	peiman khosravi
Subject	Re: [Csnd] Re: pvs fftsize vs winsize questions
	I think I asked a similar question here a while ago. See if you can find the thread. Otherwise let me know and I'll dig it out (on my way out now). P On 1 April 2013 12:48, Ben Hackbarth <hackbarth@gmail.com> wrote: bump. can anyone offer any insight? thank you in advance. -- ben On Mon, Mar 25, 2013 at 11:22 PM, Ben Hackbarth <hackbarth@gmail.com> wrote: hello, i am a little confused about the pvs initialization parameters. from the manual of pvsanal: fsig pvsanal ain, ifftsize, ioverlap, iwinsize, iwintype [, iformat] [, iinit] it seems to me that, in reading the manual entry, ifftsize is the number of samples taken from the audio stream and iwinsize is the size of the fft. in other words, ifftsize dictates the number of audio samples used while iwinsize affects the frequency resolution of the fft analysis (presumably buffered by zeros if larger than ifftsize). is this a correct reading of the parameters? this nomenclature seems backwards to me -- shouldn't ifftsize affect the bin resolution while iwinsize dictate the number of audio samples used per frame? in any case, i cannot seem to get these parameters to function properly in tadem. with the following code: fsig pvsanal ain, 4096, 1024, 8192, 1 kflag pvsftw fsig, ifna, ifnf ftsavek "/Users/ben/Desktop/tab.txt", kflag, 1, ifnf i get a table with only 2048 frequency values, not 4096 as i would expect from a winsize of 8192 (divided by two). any guidance is much appreciated! -- ben