FPGA Central

I am not a student anymore... and my math is falling apart.
I was hoping that somebody could point me to some algorithms that might
remove or reduce static from a PCM array.

Basically, I am trying to perform some speech recognition on a police
scanner, but the static is killing me.
Any help would be greatly appreciated.

Thanks!

"Steve" <[email protected]> wrote in message
news:c42tf.976$JT.773@fed1read06...
> I am not a student anymore... and my math is falling apart.
> I was hoping that somebody could point me to some algorithms that might
> remove or reduce static from a PCM array.
>
> Basically, I am trying to perform some speech recognition on a police
> scanner, but the static is killing me.
> Any help would be greatly appreciated.
>
> Thanks!
>
>
Define Static - it's characteristics etc. Do you mean FM thresholding?

Tam

I guess you want to remove the DC offset, right?

Calculate the mean value of all PCM samples, e.g. Mean([1,4,3,2]) =
(1+4+3+2)/4 = 2.5

Subtract this value from each sample in the array.

Uli

There is a crackle when a person first begins to speak..
Also.. there is a constant static buzz...
Filters should help this, no?


"Cor blimey mate" <[email protected]> wrote in message
news:lUftf.11379$[email protected]..
>
> "Steve" <[email protected]> wrote in message
> news:c42tf.976$JT.773@fed1read06...
> > I am not a student anymore... and my math is falling apart.
> > I was hoping that somebody could point me to some algorithms that might
> > remove or reduce static from a PCM array.
> >
> > Basically, I am trying to perform some speech recognition on a police
> > scanner, but the static is killing me.
> > Any help would be greatly appreciated.
> >
> > Thanks!
> >
> >
> Define Static - it's characteristics etc. Do you mean FM thresholding?
>
> Tam
>
>

Thanks, but that didn't quite do what I was looking for.

I really need something to clear up lound crackles... or static bursts.
When they speak into the scanner microphone, there is often a burst of
static... a hiss I guess it the correct word.

Any other ideas?


<[email protected]> wrote in message
news:[email protected] oups.com...
> I guess you want to remove the DC offset, right?
>
> Calculate the mean value of all PCM samples, e.g. Mean([1,4,3,2]) =
> (1+4+3+2)/4 = 2.5
>
> Subtract this value from each sample in the array.
>
> Uli
>

Steve wrote:
> Thanks, but that didn't quite do what I was looking for.
>
> I really need something to clear up lound crackles... or static bursts.
> When they speak into the scanner microphone, there is often a burst of
> static... a hiss I guess it the correct word.
>
> Any other ideas?

Bandpass the audio [center on say 1Khz] and then just pre/post mute
when you hit a threshold [e.g. samples >90%].

A simple 2-step IIR could probably do enough of the work for ya.

Not perfect but will help isolate the noise to the part of the spectrum
at least so you have a fighting chance.

Tom

"Steve" <[email protected]> wrote in message
news:I6ltf.1129$JT.327@fed1read06...
> Thanks, but that didn't quite do what I was looking for.
>
> I really need something to clear up lound crackles... or static bursts.
> When they speak into the scanner microphone, there is often a burst of
> static... a hiss I guess it the correct word.
>
> Any other ideas?
>
>

Yes I think this is thresholding caused by poor SNR in narrowband FM. It's
right across the audio spectrum and there is no known cure. Muting works
well when there is no signal of course.The problem is casued by Rician
spikes which are random in nature. Since they are spike like they are like
impulses.

Tam