windowing DFT vs FFT (newbie)

Hello,

I wrote a program in Java that does a DFT on raw 8bit samples stored in
memory from a RF ADC (Post processing). This allows me to easily adjust the
span (zoom) when I'm viewing the spectrum. Works great, but slow as hell. So
I'm now trying an FFT with a Chirp Z-Transform so I can zoom in on the
desired frequency and look for subtle modulation characteristics. I borrowed
this code from the net (see below) and seems to work as expected. The
spectrum looks like my DFT spectrum without any windowing applied. So I then
tried to window (Hamming) my data before applying the CZT to reduce the
tails on my recovered carrier, however this only smoothes the tails out and
does not reduce there apparent power like my DFT Hamming window does. Am I
doing something fundamentally wrong? How do I reduce/remove the tails from
my spectrum when I do the CZT? Here is the site where I borrowed the code:
http://www.nauticom.net/www/jdtaft/JavaCZT.htm

OK I solved it myself. I threw away the CZT and went with the more
traditional ZoomFFT. Where you down convert to baseband and run a standard
FFT. In fact, I didn't even have to downconvert since I was radically
subsampling anyways. I just used one of my negative images.

For some reason I don't feel like a newbie anymore.

"Thomas Magma" <kholmes@beer.com> wrote in message
news:%Qt3b.858062$ro6.17064603@news2.calgary.shaw.ca...
> Hello,
>
> I wrote a program in Java that does a DFT on raw 8bit samples stored in
> memory from a RF ADC (Post processing). This allows me to easily adjust
the
> span (zoom) when I'm viewing the spectrum. Works great, but slow as hell.
So
> I'm now trying an FFT with a Chirp Z-Transform so I can zoom in on the
> desired frequency and look for subtle modulation characteristics. I
borrowed
> this code from the net (see below) and seems to work as expected. The
> spectrum looks like my DFT spectrum without any windowing applied. So I
then
> tried to window (Hamming) my data before applying the CZT to reduce the
> tails on my recovered carrier, however this only smoothes the tails out
and
> does not reduce there apparent power like my DFT Hamming window does. Am I
> doing something fundamentally wrong? How do I reduce/remove the tails from
> my spectrum when I do the CZT? Here is the site where I borrowed the code:
> http://www.nauticom.net/www/jdtaft/JavaCZT.htm
>
>