erila107
06-02-2007, 04:20 PM
Hi,
I’m trying to use realFT() to convert my 2048 samples of data. The problem is not the FFT it’s the re-packing of the result from the FFT that the problem.
I start by run a FFT on a simple sin(2*pi*freq*n/2048 ), n= [1:2048]. The result of this is, what I think, a correct FFT. I can se to “peaks” , one at bin 201 and one at bin 825 (N-n).
When I try to repack the data I always end up width a lot of junk peaks at N-n (825).
I started to look in to the math a bit more, did some MatLab-ing. I can now prove (for my self) eq. 12.3.3 and 12.3.4 But I can’t prove 12.3.5. I always end up with this junk-peaks, both in MatLab and in C-code.
Please, can someone help my….I already a bit bald, and ripping of more hair because of the is not good for my hair style… ;-)
I’m trying to use realFT() to convert my 2048 samples of data. The problem is not the FFT it’s the re-packing of the result from the FFT that the problem.
I start by run a FFT on a simple sin(2*pi*freq*n/2048 ), n= [1:2048]. The result of this is, what I think, a correct FFT. I can se to “peaks” , one at bin 201 and one at bin 825 (N-n).
When I try to repack the data I always end up width a lot of junk peaks at N-n (825).
I started to look in to the math a bit more, did some MatLab-ing. I can now prove (for my self) eq. 12.3.3 and 12.3.4 But I can’t prove 12.3.5. I always end up with this junk-peaks, both in MatLab and in C-code.
Please, can someone help my….I already a bit bald, and ripping of more hair because of the is not good for my hair style… ;-)