Floating and fixed point

Dear all,

What always read about floating and fixed point in DSP. What is that
mean?
What is the different between floating and fixed points in DSP hardware
and DSP software design ?

Thank you very much.


Best wishes,
estdev

estdev wrote:

> What always read about floating and fixed point in DSP. What is that
> mean?
> What is the different between floating and fixed points in DSP
> hardware and DSP software design ?
>

Floating point means the sort of arithmetic you get on a standard
calculator.

Fixed point means, essentially, integer arithmetic scaled to a given
range.

When using floating point, issues like underflow (numbers becoming too
small) or overflow (numbers becoming too large) are less of an issue
than when using fixed point.

Ciao,

Peter K.

Go through Proakis's book on DSP. Even Oppenheim and Schafer has a good
deal on the difference between fixed and floating point math and
particularly why fixed point math is commonly used for implementing
most DSP algorithms.

Then you can advance to some really good tutorials by Randy Yates.
Google for 'Fixed point arithmetic and Randy Yates' and it should come
up.

T.


estdev wrote:
> Dear all,
>
> What always read about floating and fixed point in DSP. What is that
> mean?
> What is the different between floating and fixed points in DSP hardware
> and DSP software design ?
> 
> Thank you very much.
> 
> 
> Best wishes,
> estdev