h a l f b a k e r yLike a magnifying lens, only with rocks.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
The worker hears a yelled warning over the noise of the jackhammer. The bouncer makes his way towards angry voices across a crowded nightclub floor. The infantryman listens for voices upstairs through the tumult of a helicopter attack outside.
Noise cancellation is nothing new. These devices generate
matched frequencies to cancel those from the environment, thus rendering the near side of the headphone quieter. Noice cancelling microphones are also available - these devices get rid of noise not in the frequency of the voice, and allow the microphone to be more effectively used in loud environments.
The NCVA hybridizes these two with some stochastic resonance thrown in to boot. The NCVA device allows people to operate in loud environments with retained perception for human voices. Sound with wavelengths outside the voice are cancelled, but voice wavelengths are allowed through.
One would not want these voice wavelengths amplified, for fear of deafening the user by amplifiying noise components (eg: cymbals) which happened to be in that wavelength as well. The NCVA improves the signal of voice sounds by adding a low background of white noise at these same voice frequencies. This will not make voices louder, but will make them sharper.
I imagine the NCVA as like night vision for noise: hear the important things and not the rest.
These use a similar type concept based on volume
http://www.elvex.com/Com-550.htm
[jhomrighaus, Mar 12 2007]
a number of other similar items
http://www.westerns...or/peltorpage3.html
[jhomrighaus, Mar 12 2007]
Stochastic resonance
http://en.wikipedia...tochastic_resonance
It is all about signal to noise ratio [bungston, Mar 13 2007]
Noise cancelling microphone
http://www.adaptive...com/nc_mic.php?mm=4
and how it works. [bungston, Mar 13 2007]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
I don't think I understand. Is this a
noise-canceller that cancels only the
noise outside of the human voice
range? |
|
| |
Also, a query: what is the spectral
width of the human voice (think of the
range between a vowel like "ohh" and
the hiss of a "th"), and to what extent
would it be helpful to cancel out only
the noise which was outside of this
range? |
|
| |
Finally, I'm not clear about the white-
noise benefit. Stochastic resonance
helps (sometimes) when the signal is
just very very weak, but I'm not sure it
helps when the signal is strong but is
swamped with noise. |
|
| |
But maybe I'm missing the point. |
|
| |
/a noise-canceller that cancels only the noise outside of the human voice range?/
Yes. Like a noise-cancelling microphone. |
|
| |
/Stochastic resonance helps (sometimes) when the signal is just very very weak/ |
|
| |
Yes. After cancellation you will have the remaining frequencies in which there might be a voice. I am not sure if the added noise has to be similar in frequency to the desired signal, but I bet it does. This is a way to improve the perception of voices without simply making them louder. It is really the only new invention in this idea. I lashed it to the noise canceller because only in a situation that was already very noisy would you want to avoid simply amplifying ambient sounds to hear voices better. |
|
| |
OK, but it still doesn't convince me that
it'll work. (I have no idea, but it sounds
dodgy). What evidence do you have to
show: |
|
| |
(a) that cancelling only the noise
outside of the voice range will really
help, given that the voice range is
probably quite wide, and given that the
most "interfering" noise is likely to be
the noise *within* the vocal range? and |
|
| |
(b) that white noise superimposed on
top of "voice plus vocal-range-noise" is
going to help in any way? Stochastic
resonance is mean to help by pushing
near-threshold signals over the edge
into above-threshold; I don't think it
helps with signal-to-noise problems. |
|
| |