[chuck-users] fft issues

eduard aylon eduard.aylon at gmail.com
Wed Nov 14 16:44:27 EST 2007


Hello list,

some time ago I reported the following as a possible bug on the wiki:

* Assigining window to ifft seems to yield "poorer" results than not  
assigning it. Try it in /examples/win.ck and you'll notice a rougher  
(faster beating) sound. (eduard)

However, looking again at the code of win.ck I realised the size of  
the window was set to the half of fft_size. Unless zero-padding is  
performed, shouldn't window_size and fft_size be the same? Anyway,  
setting the window size to fft_size (both for forward fft and iverse  
fft) yields correct results in examples/win.ck. This applies to some  
of my previous emails where brick wall filters and vocoder were  
discussed. That is, window_size should be set to fft_size and  
resulting spectrum to fft_size/2. But now I am also concerned about  
the size of the spectrum. In chuck this is FFT_SIZE/2, but shouldn't  
it be FFT_SIZE/2+1 or does the first value in spectrum hold both DC  
and Nyquist components?

By the way, could chuck perform zero-padding when window_size <  
fft_size?

eduard


More information about the chuck-users mailing list