On Sun, 16 Jun 2013 11:29:02 -0500, manishp wrote:

> Sirs,
> 
> Generally, when fft is imimplemented in hw, due to prohibitive silicon
> area, the intermediate stages are allowed only 2 bit (for eg) growth. If
> I simply see this from a mathematical perspective,  we do lose a very
> significant amount of bits due to this. How come this is acceptable and
> widely used in many hw fft implementation s?
> 
> Thanks, manish

No answers yet?!?!

Try again, but look at it from a signal to noise perspective, 
quantization taken as noise.  See if only adding two bits per stage looks 
like it makes more sense.

-- 
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

Sirs,

Generally, when fft is imimplemented in hw, due to prohibitive silicon
area, the intermediate stages are allowed only 2 bit (for eg) growth. If I
simply see this from a mathematical perspective,  we do lose a very
significant amount of bits due to this. How come this is acceptable and
widely used in many hw fft implementation s?

Thanks, manish