> > But anyway couldn't DSP be used to surpress/attenuate noise? Digital
> > bandpass filters exist so therefore isn't this DSP?
>
> Again, only if the noise has some characteristic that
> distinguishes it from
> the real signal.  Suppose the noise is a 1KHz sine wave.  How is the DSP
> supposed to know that the radio station isn't sending a signal
> that happens
> to contain a 1KHz sine wave?


As Olin is usually correct, I will not dispute him.

However, there is a process of pattern matching used in DSP, that will
remove most noise from a system.

Read the input, wait for the audio to drop below a set point for an average
period of time. Record and look for a pattern of noise, being pops and
clicks, and hiss. Then when one of these patterns are present or detected
within the normal audio, the DSP will use a negative signal to phase cancel
the noise.  Its tricky, and may remove some characterization of the original
audio but with out the pops and clicks and hiss. This is a process used in
some high end automobiles to remove road noise from the passenger
compartment.  Some recording studios and radio stations use it to remove
hiss and feed back, from the microphones.

Gordon Varney

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.