On Sep 3, 8:15 am, Jerry Avins <j...@ieee.org> wrote:
> pnalelia wrote:
> > Hi people,
> >            I am trying to selectively mask a speech signal with white
> > noise.what do i mean by selectively msaking? I mean i would like the no=
ise
> > to mask when a word is spoken and to stop masking when there is a perio=
d of
> > silence. Any ideas or code on how to do this?
>
> The simple way uses a speech detector and turns on a noise source loud
> enough to blast the listener out of his seat whenever speech is
> detected. Since you equat speech to not silence, your speech detector
> should be very easy to design. I'm puzzled, though. If you blot out
> everything but silence, why not just turn the signal off altogether and
> run your noise generator with a random timer?
>
> Jerry
> --
> Engineering is the art of making what you want from things you can get.
> =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF

May be he is on his first step to build a karaoke system.

-
hurry.

glen herrmannsfeldt wrote:
> Jerry Avins wrote:
> 
> (snip)
> 
>> The simple way uses a speech detector and turns on a noise source loud 
>> enough to blast the listener out of his seat whenever speech is 
>> detected. Since you equat speech to not silence, your speech detector 
>> should be very easy to design. I'm puzzled, though. If you blot out 
>> everything but silence, why not just turn the signal off altogether 
>> and run your noise generator with a random timer?
> 
> This reminds me of making "random word generator" computer programs.
> 
> Generating letters with the right frequency doesn't tend to make
> readable words.
> 
> I will guess that one can tell the difference between random noise and
> masked speech unless one is very good with the random numbers.

True, but does it matter?

jerry
-- 
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Jerry Avins wrote:

(snip)

> The simple way uses a speech detector and turns on a noise source loud 
> enough to blast the listener out of his seat whenever speech is 
> detected. Since you equat speech to not silence, your speech detector 
> should be very easy to design. I'm puzzled, though. If you blot out 
> everything but silence, why not just turn the signal off altogether and 
> run your noise generator with a random timer?

This reminds me of making "random word generator" computer programs.

Generating letters with the right frequency doesn't tend to make
readable words.

I will guess that one can tell the difference between random noise and
masked speech unless one is very good with the random numbers.

-- glen

pnalelia wrote:
> Hi people,
>            I am trying to selectively mask a speech signal with white
> noise.what do i mean by selectively msaking? I mean i would like the noise
> to mask when a word is spoken and to stop masking when there is a period of
> silence. Any ideas or code on how to do this?

The simple way uses a speech detector and turns on a noise source loud 
enough to blast the listener out of his seat whenever speech is 
detected. Since you equat speech to not silence, your speech detector 
should be very easy to design. I'm puzzled, though. If you blot out 
everything but silence, why not just turn the signal off altogether and 
run your noise generator with a random timer?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

>On Jul 10, 5:10 am, "pnalelia" <pnale...@yahoo.com> wrote:
>> Hi people,
>>            I am trying to selectively mask a speech signal with white
>> noise.what do i mean by selectively msaking? I mean i would like the
noise
>> to mask when a word is spoken and to stop masking when there is a
period of
>> silence. Any ideas or code on how to do this?
>> Best regards
>> G.N
>
>If your SNR is high enough, you could use a simple lowpass filter and
>squelch circuit. Design the filter so that it will track the envelope
>of the speech waveform, and compare its output to a threshold. If
>you're above the threshold, you can assume that speech is present, and
>you can insert the noise. The aspect of your filter that you want to
>be careful about is its delay, however. If the filter is too long,
>there may be a discernible time during which the filter output is
>ramping up to the threshold, resulting in unmasked speech.
>
>As I said, this is only good if you have a large enough SNR that you
>can reliably discriminate between noise and signal + noise.
>
>Jason
>
Jason, I realised I posted the question in the general DSP forum, I was
meant to post it under the Matlab forum because I have to implement it in
matlab..

>On Jul 10, 5:10 am, "pnalelia" <pnale...@yahoo.com> wrote:
>> Hi people,
>>            I am trying to selectively mask a speech signal with white
>> noise.what do i mean by selectively msaking? I mean i would like the
noise
>> to mask when a word is spoken and to stop masking when there is a
period of
>> silence. Any ideas or code on how to do this?
>> Best regards
>> G.N
>
>If your SNR is high enough, you could use a simple lowpass filter and
>squelch circuit. Design the filter so that it will track the envelope
>of the speech waveform, and compare its output to a threshold. If
>you're above the threshold, you can assume that speech is present, and
>you can insert the noise. The aspect of your filter that you want to
>be careful about is its delay, however. If the filter is too long,
>there may be a discernible time during which the filter output is
>ramping up to the threshold, resulting in unmasked speech.
>
>As I said, this is only good if you have a large enough SNR that you
>can reliably discriminate between noise and signal + noise.
>
>Jason
>
Thanks Jason, I will try to do it the way you proposed.Incase of any
problems Can I get back at you?
Regards

On Jul 10, 5:10 am, "pnalelia" <pnale...@yahoo.com> wrote:
> Hi people,
>            I am trying to selectively mask a speech signal with white
> noise.what do i mean by selectively msaking? I mean i would like the noise
> to mask when a word is spoken and to stop masking when there is a period of
> silence. Any ideas or code on how to do this?
> Best regards
> G.N

Why white noise?

Dirk

On Jul 10, 5:10 am, "pnalelia" <pnale...@yahoo.com> wrote:
> Hi people,
>            I am trying to selectively mask a speech signal with white
> noise.what do i mean by selectively msaking? I mean i would like the noise
> to mask when a word is spoken and to stop masking when there is a period of
> silence. Any ideas or code on how to do this?
> Best regards
> G.N

If your SNR is high enough, you could use a simple lowpass filter and
squelch circuit. Design the filter so that it will track the envelope
of the speech waveform, and compare its output to a threshold. If
you're above the threshold, you can assume that speech is present, and
you can insert the noise. The aspect of your filter that you want to
be careful about is its delay, however. If the filter is too long,
there may be a discernible time during which the filter output is
ramping up to the threshold, resulting in unmasked speech.

As I said, this is only good if you have a large enough SNR that you
can reliably discriminate between noise and signal + noise.

Jason

Hi people,
           I am trying to selectively mask a speech signal with white
noise.what do i mean by selectively msaking? I mean i would like the noise
to mask when a word is spoken and to stop masking when there is a period of
silence. Any ideas or code on how to do this?
Best regards
G.N