Reply by chirag jagadish February 19, 20092009-02-19
i had problems with TI fft library before
i felt that easier than fixing that is to build your own code to do these transforms etc
--- On Tue, 17/2/09, namrita_shenoy wrote:
From: namrita_shenoy
Subject: [c28x] Real FFT using sprc081
To: c...
Date: Tuesday, 17 February, 2009, 6:26 AM

Hi,

I have integrated sprc081 with my project and was able to get some
results. The fft.mag gave 0 for all 64 points (I used 128 point real
FFT). The document mentions that for fft.calc "The N1 scaling of FFT is
distributed across the stages". Probably this might be the reason.

So I scaled up my input with some trial and error to get the magnitude
close enough with the one calculated using microsoft excel. I found
that scaling factor of approx 5000 gave me a close enough result. But
this does not have any relation with 1/N factor. Can anyone please tell
me what should be the factor?

Appreciate any help in this regard,
thanks,
Nams

_____________________________________
Reply by namrita_shenoy February 17, 20092009-02-17
Hi,

I have integrated sprc081 with my project and was able to get some
results. The fft.mag gave 0 for all 64 points (I used 128 point real
FFT). The document mentions that for fft.calc "The N1 scaling of FFT is
distributed across the stages". Probably this might be the reason.

So I scaled up my input with some trial and error to get the magnitude
close enough with the one calculated using microsoft excel. I found
that scaling factor of approx 5000 gave me a close enough result. But
this does not have any relation with 1/N factor. Can anyone please tell
me what should be the factor?

Appreciate any help in this regard,
thanks,
Nams

_____________________________________