Forums

Use of DSP Technology to filter ambient noise

Started by parallel54 June 9, 2009
Hi there,

I am interested in using DSP technology to filter ambient noise. That is
not just headphone noise cancellation but to reduce noise in an external
area. So for example to use a technology to "block" / "reduce" the noise of
an aircraft/train in the garden area of a restaurant. 

So the device would project the inverse of the background noise and so
reduce it within the given area. Does anyone know of any systems that
currently do this?? Very interested in the group's thoughts on this.

Thanks

TIM



parallel54 wrote:
> Hi there, > > So the device would project the inverse of the background noise and so > reduce it within the given area. Does anyone know of any systems that > currently do this?? Very interested in the group's thoughts on this.
There is no problem to reduce the noise in a given small area with the size of the fraction of the audio wavelength, but it is practically impossible to cover the larger area. Active noise reduction headphones are common, noise cancellation in the cars had been experimented with. In the later case, the target area only includes the locations of the heads of the driver and the passengers. Somewhat 10dB of the noise reduction can be achieved; however I don't know of any commercial mass deployed systems which use this concept. Fixing the source of the noise is a lot simpler, cheaper and more efficient. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
>Yeah but in the case of planes & trains you can't prevent the noise. My
aim is not to reduce it to zero but simply to reduce it. I would know the direction of the noise if that helps. TIM
> >parallel54 wrote: >> Hi there, >> >> So the device would project the inverse of the background noise and so >> reduce it within the given area. Does anyone know of any systems that >> currently do this?? Very interested in the group's thoughts on this. > >There is no problem to reduce the noise in a given small area with the >size of the fraction of the audio wavelength, but it is practically >impossible to cover the larger area. Active noise reduction headphones >are common, noise cancellation in the cars had been experimented with. >In the later case, the target area only includes the locations of the >heads of the driver and the passengers. Somewhat 10dB of the noise >reduction can be achieved; however I don't know of any commercial mass >deployed systems which use this concept. Fixing the source of the noise >is a lot simpler, cheaper and more efficient. > > >Vladimir Vassilevsky >DSP and Mixed Signal Design Consultant >http://www.abvolt.com > >
On Jun 9, 9:01&#2013266080;am, "parallel54" <t...@newmediaintelligence.com> wrote:
> >Yeah but in the case of planes & trains you can't prevent the noise. My > > aim is not to reduce it to zero but simply to reduce it. I would know the > direction of the noise if that helps. > > TIM > > > > > > >parallel54 wrote: > >> Hi there, > > >> So the device would project the inverse of the background noise and so > >> reduce it within the given area. Does anyone know of any systems that > >> currently do this?? Very interested in the group's thoughts on this. > > >There is no problem to reduce the noise in a given small area with the > >size of the fraction of the audio wavelength, but it is practically > >impossible to cover the larger area. Active noise reduction headphones > >are common, noise cancellation in the cars had been experimented with. > >In the later case, the target area only includes the locations of the > >heads of the driver and the passengers. Somewhat 10dB of the noise > >reduction can be achieved; however I don't know of any commercial mass > >deployed systems which use this concept. Fixing the source of the noise > >is a lot simpler, cheaper and more efficient. > > >Vladimir Vassilevsky > >DSP and Mixed Signal Design Consultant > >http://www.abvolt.com- Hide quoted text - > > - Show quoted text -
You know that adding noise to noise in order to reduce it is very problematic. The adding of energy doesn't destroy any energy; instead it increases the total energy. Thus all you are pretty much doing is redistributing the energy. This can be great if you can remove the noise from the listener, but I hate to be where the energy has been moved to. Sure in some cases one can use a noise cancellor like a passive absorber. This is like learning to move yours hands with the ball when you catch a ball in order to reduce its likelyhood of reflecting out of your hands. Or in this case have a speaker diaphragm move with an incoming wave to partially absorb it. But as Vladimir said the active noise reduction is not great. With planes, pilots often use noise cancelling headphones, but there the area that needs to have the noise reduced is quite small and a good bit of the noise reduction is via absorption in the headphone itself. And simple active noise cancellation is done since the physical distance between the inside and the outside of the headphone is quite small in contrast to the wavelengths of "most" sounds. By "most" I'm referring to sounds that have most of their energy in the low frequency part of the spectrum. But in answer to your question about common implementations of active noise reduction. I know of the noise cancelling headphones. I recall some work with using a speaker in an automotive exhaust system, but I don't think that got too far. A big muffler is simpler and cheaper than an active system in that case. Clay
W dniu 2009-06-09 15:59, clay@claysturner.com pisze:
(...)
> But in answer to your question about common implementations of active > noise reduction. I know of the noise cancelling headphones. I recall > some work with using a speaker in an automotive exhaust system, but I > don't think that got too far. A big muffler is simpler and cheaper > than an active system in that case.
http://www.nokia.com/press/bh-905 regards Roman Rumian
clay@claysturner.com wrote:
 
< You know that adding noise to noise in order to reduce it is very
< problematic. The adding of energy doesn't destroy any energy; instead
< it increases the total energy. Thus all you are pretty much doing is
< redistributing the energy. This can be great if you can remove the
< noise from the listener, but I hate to be where the energy has been
< moved to. Sure in some cases one can use a noise cancellor like a
< passive absorber. This is like learning to move yours hands with the
< ball when you catch a ball in order to reduce its likelyhood of
< reflecting out of your hands. Or in this case have a speaker diaphragm
< move with an incoming wave to partially absorb it. But as Vladimir
< said the active noise reduction is not great.

It seems that there are two things both called "noise shaping" that
apply here.  One is very applicable to DSP, as seen in the wikipedia
entry for "noise shaping".  That allows one to reduce the effect
(but not the amount) of quantization noise by moving it to frequencies
that are less audible to humans.  That is in addition (or in
combination) with dither, which could also be looked at as adding
noise to reduce noise.

The other is in quantum mechanics, where one can reduce quantum
noise and the corresponding uncertianty by moving the noise around.
That is, decrease the uncertainty where it is most important, while
increasing it somewhere else.
 
< With planes, pilots often use noise cancelling headphones, but there
< the area that needs to have the noise reduced is quite small and a
< good bit of the noise reduction is via absorption in the headphone
< itself. And simple active noise cancellation is done since the
< physical distance between the inside and the outside of the headphone
< is quite small in contrast to the wavelengths of "most" sounds. By
< "most" I'm referring to sounds that have most of their energy in the
< low frequency part of the spectrum.

I only tried them once.  It seems that they do a good job of
reducing low frequency noise, but not high frequency.  Most of
the time that works pretty well, though.

-- glen
On Tue, 09 Jun 2009 07:13:05 -0500, Vladimir Vassilevsky wrote:

> parallel54 wrote: >> Hi there, >> >> So the device would project the inverse of the background noise and so >> reduce it within the given area. Does anyone know of any systems that >> currently do this?? Very interested in the group's thoughts on this. > > There is no problem to reduce the noise in a given small area with the > size of the fraction of the audio wavelength, but it is practically > impossible to cover the larger area. Active noise reduction headphones > are common, noise cancellation in the cars had been experimented with. > In the later case, the target area only includes the locations of the > heads of the driver and the passengers. Somewhat 10dB of the noise > reduction can be achieved; however I don't know of any commercial mass > deployed systems which use this concept. Fixing the source of the noise > is a lot simpler, cheaper and more efficient.
It would only even be theoretically possible if you surrounded the area with independently-driven loudspeakers, and ringed it with microphones. You'd probably need several times as many loudspeakers as there were square wavelengths of the highest frequency you wanted to damp, and at least as many microphones as loudspeakers. Paying the railroad to move the tracks would probably be cheaper than the speakers and microphones, much less the processing power needed to figure out how to drive the loudspeakers and the present value of the power to drive the processor, amortized over the (probably short) lifetime of the system. -- www.wescottdesign.com
On 6/9/2009 4:38 AM, parallel54 wrote:
> Hi there, > > I am interested in using DSP technology to filter ambient noise. That is > not just headphone noise cancellation but to reduce noise in an external > area. So for example to use a technology to "block" / "reduce" the noise of > an aircraft/train in the garden area of a restaurant. > > So the device would project the inverse of the background noise and so > reduce it within the given area. Does anyone know of any systems that > currently do this?? Very interested in the group's thoughts on this. > > Thanks > > TIM >
As others have mentioned, it's tough to noise cancel over a large area. You might look into shrubbery/vegetation around the garden (tall stuff) that can reflect the noise away from the garden. Some well-placed heavy reflectors (I see them curved a lot) may work well, too, depending on how localized the noise source(s) are. That sort of thing combined with low level music/noise/sound generators can often do wonders.
On Jun 10, 12:50&#2013266080;am, Eric Jacobsen <eric.jacob...@ieee.org> wrote:
> > As others have mentioned, it's tough to noise cancel over a large area. > > You might look into shrubbery .. around the garden ...
any suggestions of where to find a shrubber? my previous shrubber, Roger, ain't in the biz anymore. Ni, r b-j
On 6/10/2009 10:57 AM, robert bristow-johnson wrote:
> On Jun 10, 12:50 am, Eric Jacobsen<eric.jacob...@ieee.org> wrote: >> As others have mentioned, it's tough to noise cancel over a large area. >> >> You might look into shrubbery .. around the garden ... > > any suggestions of where to find a shrubber? my previous shrubber, > Roger, ain't in the biz anymore. > > > Ni, > > r b-j
A shrubbery! Ni!