Reply by George Kadziolka ●July 28, 20062006-07-28
You may want to check the data type you are plotting. It sounds like the
default char might have been selected. You want int.

Cheers,

George

-----Original Message-----
From: a... [mailto:a...] On Behalf Of
r...@inspiredsystems.net
Sent: Thursday, July 27, 2006 3:42 PM
To: a...
Subject: [adsp] rfft/cfft outputs not as expected

I've recently dug out an old 21061 Eval board and started 'dusting off' some
brain cells by running the demos with EZSharc and VisualDSP. I'm trying the
fft.c routines, and it appears to work (sort of!) in the EZSharc GUI, but
when using VisualDSP to modify/build/display the results, I don't see what I
expect for output. When I execute the code and plot the data, I see the
expected waveforms for the input data and the windowed data, but not for
outputs. The real data seems to 'oscillate' back and forth between two
random (?) numbers and the imaginary data is always 0. If my memory serves
me, it seems like I should see mostly 0's (or small numbers) except in the
index that represents the input frequency. I'm sampling at 48kHz and
executing a 256 point fft, so it seems like I should see larger numbers at
the 5th or 6th index in one or both outputs. Any suggestions or experience?

Rick
Reply by rick...@inspiredsystems.net ●July 27, 20062006-07-27
I've recently dug out an old 21061 Eval board and started 'dusting off' some brain cells by running the demos with EZSharc and VisualDSP. I'm trying the fft.c routines, and it appears to work (sort of!) in the EZSharc GUI, but when using VisualDSP to modify/build/display the results, I don't see what I expect for output. When I execute the code and plot the data, I see the expected waveforms for the input data and the windowed data, but not for outputs. The real data seems to 'oscillate' back and forth between two random (?) numbers and the imaginary data is always 0. If my memory serves me, it seems like I should see mostly 0's (or small numbers) except in the index that represents the input frequency. I'm sampling at 48kHz and executing a 256 point fft, so it seems like I should see larger numbers at the 5th or 6th index in one or both outputs. Any suggestions or experience?

Rick