wb1j+@andrew.cmu.edu (William M. Bumgarner) (11/15/88)
Forwarded by b.bum because the computers at Harvard are being mean to David. Forwarded mail follows: ---- From: fry%zariski@harvard.harvard.edu (David Fry) I was wondering if anyone has thought of using the NeXT computer for an anti-noise demonstration. I think it would be a nice little trick. For those who don't know, systems are being developed that can do a Fourier analysis of repeated sounds to find their component sound waves, and then emit a sound composed of the the mirror image of these waves, resulting in destructive interference. Your ear then hears nothing. This could be used in an airplane, for instance, so the passengers would hear no engine noise. With its built-in digitizer, its sound generation capabilities, and its powerful DSP chip for Fourier analysis, it seems that it would be pretty simple for the NeXT to do this in real time. Maybe the people at NeXT have already done this? David Fry fry@huma1.harvard.EDU Department of Mathematics fry@huma1.bitnet Harvard University ...!harvard!huma1!fry Cambridge, MA 02138
relph@presto.ig.com (John M. Relph) (11/16/88)
In article <YXTslXy00UgXQd6HAX@andrew.cmu.edu> fry%zariski@harvard.harvard.edu (David Fry) writes: >For those who don't know, systems are being developed that >can do a Fourier analysis of repeated sounds to find their >component sound waves, and then emit a sound composed of the >the mirror image of these waves, resulting in destructive >interference. Your ear then hears nothing. This could be >used in an airplane, for instance, so the passengers would >hear no engine noise. > >With its built-in digitizer, its sound generation >capabilities, and its powerful DSP chip for Fourier analysis, >it seems that it would be pretty simple for the NeXT to do >this in real time. I don't think the NeXT machine is up to doing this. In the physics demo where voice was being analysed real-time and frequency plots being displayed, it seemed to be able to do a display every second (perhaps slightly faster or slower). Perhaps, with repeated sounds, this is fast enough to keep the generated "mirror-image" sounds in sync, but I don't think so. It would seem that, for example, if the engine was slowing, the "mirror-image" sound would always be a few Hz higher than the actual engine noise, resulting in a non-zero interference sound. This could be real nasty. In some situations this might actually result in louder noise. Ouch. But then again, perhaps not. -- John
sandell@batcomputer.tn.cornell.edu (Gregory Sandell) (11/17/88)
In article <YXTslXy00UgXQd6HAX@andrew.cmu.edu> wb1j+@andrew.cmu.edu (William M. Bumgarner) writes: >---- >From: fry%zariski@harvard.harvard.edu (David Fry) > >I was wondering if anyone has thought of using the NeXT >computer for an anti-noise demonstration. I think it would be >a nice little trick. >For those who don't know, systems are being developed that >can do a Fourier analysis of repeated sounds to find their >component sound waves, and then emit a sound composed of the >the mirror image of these waves, resulting in destructive >interference. Your ear then hears nothing. This could be >used in an airplane, for instance, so the passengers would >hear no engine noise. > >it seems that it would be pretty simple for the NeXT to do >this in real time. > >Maybe the people at NeXT have already done this? > >David Fry fry@huma1.harvard.EDU Well, since you're a mathematician you may have an idea of how complicated a task this is. Most people in DSP work I have talked to about this idea says, "yeah, sure...show me it happening." Noise by definition is aperiodic, so your FFT is going to have a hell of a time finding some REGULAR waveform to destructively interfere. If that can't be done, then were talking about analyzing the signal in real time, and for each sample inputted, outputting a destructive signal for that sample. So they say that the Motorolla could squeeze in 600 instructions between every sample at 44.1kHz. Maybe the calculation could be done...but you also have to be recording samples and running both D/A and A/D at the same time. Could it really be done? Greg Sandell mostly aperiodic it will be VERY difficult for a fourier analysis to find any kind of regularity in the
jacquemin-michel@CS.YALE.EDU (Michel Jacquemin) (11/17/88)
>From: fry%zariski@harvard.harvard.edu (David Fry) > >I was wondering if anyone has thought of using the NeXT >computer for an anti-noise demonstration. I think it would be >a nice little trick. >For those who don't know, systems are being developed that >can do a Fourier analysis of repeated sounds to find their >component sound waves, and then emit a sound composed of the >the mirror image of these waves, resulting in destructive >interference. Your ear then hears nothing. This could be >used in an airplane, for instance, so the passengers would >hear no engine noise. Even better in a personal workstation: you will be able to work in silence wherever you are (remark: somebody on this group said that the disk drive was noisy; don't worry, you can anihilate this noise) ... Or, if somebody is disturbing you by talking all the time, you could turn your nihilizer on, and no sound would come out of his mouth! >it seems that it would be pretty simple for the NeXT to do >this in real time. Michel Jacquemin "L'ouie de l'oie de Louis a oui ce que toute oie oie" Raymond Devos
srw%pws1@Sun.COM (Scott Wiesner) (11/17/88)
>>For those who don't know, systems are being developed that >>can do a Fourier analysis of repeated sounds to find their >>component sound waves, and then emit a sound composed of the >>the mirror image of these waves, resulting in destructive >>interference. Your ear then hears nothing. This could be >>used in an airplane, for instance, so the passengers would >>hear no engine noise. >> > Well, since you're a mathematician you may have an idea of how complicated > a task this is. Most people in DSP work I have talked to about this > idea says, "yeah, sure...show me it happening." I'm currently reading the Rutan/Yeager book about their trip around the world in the Voyager airplane. One of their problems was the noise they had to deal with. Since excess weight meant death to them, there was no soundproofing. They tell of using a system by Bose that does exactly what is described above. They said it worked great.
mayer@hplabsz.HPL.HP.COM (Niels Mayer) (11/17/88)
In article <6843@batcomputer.tn.cornell.edu> sandell@tcgould.tn.cornell.edu (Gregory Sandell) writes: >So they say that the Motorolla could squeeze in 600 instructions between >every sample at 44.1kHz. Maybe the calculation could be done...but you also >have to be recording samples and running both D/A and A/D at the same time. >Could it really be done? Do you really need HI-FI anti-noise at 44.1 kilosamples/second? Since people really have trouble hearing stuff over 15khz (especially NOISE), you could probably scale back the time-domain resolution to under 30 kilosamples/second. You probably don't need CD quality signal domain resolution either (CD's have about 32 bits/sample for both channels). Hey! I got a GREAT idea. You could have it SAMPLE its own FAN and DISK drive noize and CANCEL this out using the DSP and an expensive A/D/D/A subsystem -- the NeXT box as a zen computational Rube Goldgberg device. I once built such "cone of silence" for Joe Isuzu's CAR, as well as for my WIFE's (Morgan Fairchild) battery operated sex toys, and now my life is so TWISTED and STRANGE that you have to put it in REVERSE to go FORWARD, and get it UP to go DOWN (respectively)... Yeah, that's the ticket.... -- Niels "maybe we can apply DSP noise cancellation to comp.sys.next" Mayer.
cantrell@Alliant.COM (Paul Cantrell) (11/18/88)
In article <77955@sun.uucp> srw%pws1@Sun.COM (Scott Wiesner) writes: >I'm currently reading the Rutan/Yeager book about their trip >around the world in the Voyager airplane. One of their problems >was the noise they had to deal with. Since excess weight meant >death to them, there was no soundproofing. They tell of using >a system by Bose that does exactly what is described above. They >said it worked great. Yes indeed, Bose does make a set of anti-noise headphones. They are now supposed to be standard equipment in Enstrom helicopters. There was an article in one of the aviation rags, possibly AOPA-Pilot, where they reviewed them. They work better at some frequency than others, but are supposed to be quite effective. The big gain here is that they can get a large drop in noise without having to clamp down on your head as hard as some of the competition. Anyone who has had to fly for 6-8 hours with a good set of noise reducing headphones knows how the clamp pressure can induce a nice headache. I'm sure the Bose set does no analysis of the noise - they are far too simple for that. They have to be simply reproducing what is recorded with a microphone with reverse phase. I know of one system currently in use that works by analysis of the noise waveform - MDHC Helicopters in Mesa, Arizona was doing some work called Higher Harmonic something-or-other in which accelerometers are placed at various places in the helicopter, and (I believe an FFT) analysis tries to determine the basic components of the vibration. It then actually manipulates the controls of the helicopter in such a way that the vibration is decreased or eliminated. Since vibrations induce fatigue in both humans and helicopters, this is fairly exciting. As far as I know, the system worked quite well. Perhaps the NeXT box can use this to get rid of the fan noise? :) PC
rjfrey@kepler1.UUCP (Robert J Frey) (11/18/88)
Regarding the notion of using an FFT to analyze white noise then cancelling it out to increase signal-to-noise, trying to net out one white noise signal with another will probably just mean more noise. I have seen systems that do what's being described here. There's a two-part mike, one part is directed outward so that it picks up ambient noise, the other is directed inward so that it picks up voice plus ambient noise. A simple circuit then nets out the two signals. If you wanted to get a little fancy you could perform a spectral analysis on the two inputs separately, then try and analyze the difference. The important thing to remember is that white noise is by definition unpredictable. If you want to perform this sort of enhancement you need *simultaneous* samples. Of course, there are lots of signal processing techniques that work on a different basis: that of assuming there is something fundamentally different between the character of the signal and the "noise". For example, if the signal-to-noise ratio is fairly low, then filtering out frequencies below some threshold (amplitude or frequency) works okay. These techniques might have trouble filtering out alot of aperiodic noise which was "well mixed" with the signal, and none of them attempt any real prediction of the noise per se. Some remarkable enhancement is possible, however. If the noise were not "white", then an appropriate analysis might have a chance at filtering it out. For example, certain over-the-horizon radars employed by the USSR can disrupt shortwave broadcasts by overlaying an 8 Hz "clicking". This type of interference might be amendable to a predictive filtering. ============================================================================== |Dr. Robert J. Frey | {icus, spl1, dasys1}!acsm!kepler1!rjfrey | |Kepler Financial Management, Ltd.|------------------------------------------| |100 North Country Rd., Bldg. B | The views expressed are wholly my own and| |Setauket, NY 11766 | and do not reflect those of the Indepen- | |(516) 689-6300 x.16 | dent Republic of Latvia. | ==============================================================================
spasss@tekigm2.TEK.COM (Spudboy) (11/18/88)
In a recent article David Fry wrote: *I was wondering if anyone has thought of using the NeXT *computer for an anti-noise demonstration. I think it would be *a nice little trick. [and explains how a sound can be generated to cancel another sound] *Maybe the people at NeXT have already done this? let me quote another recent article for you by Adam Feigin: *The beastie has the noisiest disk drive that I have ever heard. On every *access, you can hear ka-chunk, ka-chunk, ka-chunk as information is read off *the disk. If you get one, I hope you have a loud air-condo system to drown *out the noise...and people complain about Mac fan noise ! (they aint seen *nothin yet) nope, I don't think the people at NeXT have... spass -- spass stoiantschewsky (206)695-4691 after 9pm | It has been raining a lot lately, glug, glug... (206)253-5632 after 9am |
flip@pixar.UUCP (Flip Phillips) (11/18/88)
In article <6840@ig.ig.com> relph@PRESTO.IG.COM.UUCP (John M. Relph) writes: >In article <YXTslXy00UgXQd6HAX@andrew.cmu.edu> fry%zariski@harvard.harvard.edu (David Fry) writes: [...] >I don't think the NeXT machine is up to doing this. In the physics >demo where voice was being analysed real-time and frequency plots >being displayed, it seemed to be able to do a display every second >(perhaps slightly faster or slower). [...] As I recall, the freq transform was being done by the 68030, not the DSP. Remember, be weary of demos, people like me write them! -- Flip Phillips {sun | ucbvax}!pixar!flip Pixar - Marin County, California
simon@macondo.sw.mcc.com (Simon John Gibbs) (11/19/88)
In article <6843@batcomputer.tn.cornell.edu>, sandell@batcomputer.tn.cornell.edu (Gregory Sandell) writes: : In article <YXTslXy00UgXQd6HAX@andrew.cmu.edu> wb1j+@andrew.cmu.edu (William M. Bumgarner) writes: : >---- : >From: fry%zariski@harvard.harvard.edu (David Fry) : > : >I was wondering if anyone has thought of using the NeXT : >computer for an anti-noise demonstration. I think it would be : >a nice little trick... : : Well, since you're a mathematician you may have an idea of how complicated : a task this is. Most people in DSP work I have talked to about this : idea says, "yeah, sure...show me it happening." ... This is not a new idea, and it is possible when the noise source is sufficiently predicatable. There's a company named Noise Cancellation Technology Inc. who've been active in this area for a few years (they were mentioned in Omni about a year ago). I believe they have a product for factory environments and are working in a car muffler. Simon Gibbs (simon@mcc.COM)
unniks@notrees.ACA.MCC.COM (C. Unnikrishnan) (11/21/88)
From article <1608@macondo.sw.mcc.com>, by simon@macondo.sw.mcc.com (Simon John Gibbs): > > This is not a new idea, and it is possible when the noise > source is sufficiently predicatable. There's a company named Noise > Cancellation Technology Inc. who've been active in this area for > a few years (they were mentioned in Omni about a year ago). I > believe they have a product for factory environments and are > working in a car muffler. > > Simon Gibbs (simon@mcc.COM) any info on how it is supposed to work? this reminds me of one of arthur c. clarke's short novels which i read many years back. basically some folks are trying to sabotage a concert by using a sound cancellation device. they tap into the main amplifier output and feed it in to produce a waveform which completely cancels out the original sound wave before it reaches the audience's ears. dont remember the name of the book or the whole story now. unni
boyle@altair.mcs.anl.gov (11/22/88)
Hmmm..... Let's see. It seems to me that sound cancellation, especially for a distributed source, must be more complicated than this discussion suggests. Isn't it the case that if the source of the cancelling cound is not coincident with the source of the sound, cancellation would be perfect only if a few places, not everywhere? Guess I'd have to ask for a seat in the "No Sounding" section of that airplane!
jr@bbn.com (John Robinson) (11/22/88)
In article <4000001@altair>, boyle@altair writes: > >Hmmm..... Let's see. It seems to me that sound cancellation, >especially for a distributed source, must be more complicated than >this discussion suggests. Isn't it the case that if the source of the >cancelling cound is not coincident with the source of the sound, >cancellation would be perfect only if a few places, not everywhere? Can you say sound hologram? If you are trying to gets the peaks and valleys of air pressure to go away everywhere, you'd have to produce the perfect inverse signal in 3-space somehow. If you get to put your jamming signal through the same point source as the original, this might be possible, but once it has gotten into the room (or whatever, as long as it isn't anechoic) the sound will be impossible to cancel. Note that the systems that have been described (e.g., for helicopters) use headphones, hence interpose a source pretty close to the path of all sound getting to your ear. Doing it in a room, though... The only thing I am aware of that works something like this is the conjugate mirror stuff (see a recent SciAm article) for coherent light. Maybe if they would tell us about how the stealth planes work too. Sound reinforcement systems for big outdoor concerts (Watkins Glen was like this) where they put speakers a ways out from the stage have to be careful to introduce the right amount of delay to avoid muddying the signal. Some musicianns use out-of-phase mike pairs to cancel most of the ambient noise pickup (they sing into just one). And, if you have ever hooked up stereo speakers out of phase you probably know that that can kill some frequencies pretty effectively. So noise can be canceled some, but things'll have to be a lot more fancy before: >Guess I'd have to ask for a seat in the "No Sounding" section of that >airplane! -- /jr jr@bbn.com or bbn!jr
c60a-2di@e260-2d.berkeley.edu (The Cybermat Rider) (12/07/88)
In article <32595@bbn.COM> jr@bbn.com (John Robinson) writes: >In article <4000001@altair>, boyle@altair writes: [misc. stuff about sound cancellation deleted] >The only thing I am aware of that works something like this is the >conjugate mirror stuff (see a recent SciAm article) for coherent >light. The device (if my memory serves me correctly) is called a "phase conjugator". Constructing a similar device for sound would be pretty difficult, coz you still have to worry about the delay between the original and conjugated waves. > Maybe if they would tell us about how the stealth planes work >too. The way I see it, Stealth merely ABSORBS or DEFLECTS radar waves to ensure that they don't return to the receiver and betray its position. Sound cancellation equipment probably isn't built-in, since it's designed to fly SLOW and (given its SIZE) HIGH, so unless you're thinking of a ground-based equivalent of "Big Ears", such expensive (and probably cumbersome) equipment probably isn't necessary. >-- >/jr >jr@bbn.com or bbn!jr ---------------------------------------------------------------------------- Adrian Ho a.k.a. The Cybermat Rider University of California, Berkeley c60a-2di@web.berkeley.edu Disclaimer: Nobody takes me seriously, so is it really necessary?