 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Natural Science Forum / Physics / Acoustics / March 2004Noise-canceling headset microphones
My company, Soliloquy Learning (http://soliloquylearning.com/), has developed an educational software product that uses speech recognition to help children increase their reading fluency. It is intended primarily for use in schools, and frequently the environment is pretty noisy; in some cases, we have 20+ kids all in one room, each with a laptop and each reading out loud into our product. We've always required the use of a headset, and now my concern is in details of headset mike performance, including noise canceling (NC) performance. From my survey of the headset market (computer headsets, not cellphone headsets), it appears that most use an NC mike element that's 2-ported, one port toward the mouth, and another facing away. I lack basic knowledge of how NC mikes work, and I haven't been able to find much on the web. Can anyone point me to any material on how NC mikes work, their directional, near/far-field properties, etc.? Andrea Electronics also has a proprietary "Active NC" technology, apparently with 2 microphones back-to-back; can anyone suggest how and whether that's superior to the more common (and less expensive) NC headsets? As I see it, we have at least 2 kinds of acoustic environmental problems (in addition to the possibility of non-speech environmental noise) -- general room noise stemming from many talkers distributed over space (presumably a diffuse field being reflected from all structural surfaces) and also speech from nearby students, some of whom can be rather loud. So one of my concerns is whether the NC is more effective in some directions compared to others, and whether I can use such knowledge to advise users how best to position their students when space is limited. Thanks for any assistance. cheers, jerry wolf soliloquy learning, inc. > From my survey of the headset market (computer headsets, not cellphone > headsets), it appears that most use an NC mike element that's > 2-ported, one port toward the mouth, and another facing away. I lack Yes. That's correct. This arrangement is called a "gradient microphone". That is, the two microphones are connected in series and out-of-phase. > basic knowledge of how NC mikes work, and I haven't been able to find > much on the web. Can anyone point me to any material on how NC mikes > work, their directional, near/far-field properties, etc.? Andrea In the gradient microphone, you might think that the net sum signal is "nothing", but like many things in life, that's not always true. It is chiefly true only for sounds originating at great distances, so that the sound arring from same is essentailly a plane wave, and aabout the same strength from one inch to the next. But when the sound source is very close to the pair of microphones, the sound level in the microphone nearer to the source, e.g. the mouth of a talker for instance, is stronger than the sound level at the more distant microphone more distant. The latter sound filed around the microphone is said to have a "gradient". Since noise usually arises at some distance to the microphone and talker, such remote sound is discriminated against in comparison tothe talker sound near the microphone. > Electronics also has a proprietary "Active NC" technology, apparently > with 2 microphones back-to-back; can anyone suggest how and whether > that's superior to the more common (and less expensive) NC headsets? Sounds like a play on words. Tech ically the microphone is an active device. But what Andrea describes (if I undestand their words) is not ANC. > As I see it, we have at least 2 kinds of acoustic environmental > problems (in addition to the possibility of non-speech environmental [quoted text clipped - 5 lines] > can use such knowledge to advise users how best to position their > students when space is limited. It will help in that there will be less babble in what they hear if they use noise cancelling mics, since this will reduce the amount of babble picked up along with their own voice and transmitted to their ears. Another factor, not explicitly noted by you, is that additional improvement can come from the use of ANC headsets, since that reduced the amount of babble that reaches their ears by bypassing the headset earcup altogether. To that extent, Andrea can be correct, provided that they also provide genuine noise cancellation circuitry to the headset speakers. Angelo Campanella  Signature--------- www.CampanellaAcoustics.com --------- "I have simply studied carefully whatever I've undertaken, and tried to hold a reserve that would carry me through." - Charles A. Lindbergh. "As for background noise level; 35 dBA is a good classroom; 45 dBA is a sound masking system!" - Anthony K. Hoover > Yes. That's correct. This arrangement is called a "gradient > microphone". That is, the two microphones are connected in series and > out-of-phase. Is there anything more to the ANC mics than just a cardiod pattern or have they figured out how to effect a second order pattern? Bob Signature"Things should be described as simply as possible, but no simpler." A. Einstein > Is there anything more to the ANC mics than just a cardiod > pattern or have they figured out how to effect a second > order pattern? First of all, Bob, I'd like to thank you for the kind remarks about me that you posted in rec.audio.pro. I'd like to take this opportunity to reciprocate by stating that your gross ignorance of the fundamental concepts never ceases to amaze me, particularly in view of your interest in microphone technology. As for the first part of your question, noise cancelling pressure gradient microphones have a figure of eight directional pattern, not a cardioid directional pattern. This is so because at +/-90 degrees the front and the rear of the diaphragm of a pressure gradient microphone see essentially the same pressure. The pressure difference is therefore theoretically zero. As for the second part of your question,"they" have indeed figured out how to effect higher order directional patterns. In this regard, perhaps you should buy some books and read them. A good start would be the AES anthology on Microphones. If you have it collecting dust on your bookshelf, you might want take it down, dust it off and brouse the article by Beavers & Brown entitled "Third Order Gradient Microphone for Speech Reception." The original article was published in the early 70s, and is not representative of current technology. But, you clearly have a lot of catching up to do. Angelo, I don't know a thing about electronics. (Very) Long ago I worked for a safety company, that sold such headsets for industrial purposes (at the time the producer was called Optac and is was build into the best type Bilsom ear mufflers). It was something as described here, but they called it an "Electret compensation microphone" (whatever that means). It was extremely small. Anyhow the effect was amazing: We used it with walky talkies in the industry, and we could talk with one another very easily in environmental noise exceeding 90 dB(A). We used it standard, in very loud circumstances, when we did measurements with several people simultaniously. We didn't need to raise our voice, and even, just for fun, could lower it to a level lower than needed for standard conversations. The mic was indeed very close to the mouth but really small. But I'm the wrong guy to explain how it worked. I always thought (maybe wrongly) that it had only one membrane accessible on both sides. Eric > > From my survey of the headset market (computer headsets, not cellphone > > headsets), it appears that most use an NC mike element that's [quoted text clipped - 49 lines] > > Angelo Campanella > But I'm the wrong guy to explain how it worked. I always thought (maybe > wrongly) that it had only one membrane accessible on both sides. You are correct in principal. Such a membrane microphone is in effect a gradient microphone, bur could also be termed a velocity microphone. Clearly, in a long wave (low frequency) field the membrane sees the same pressure on both sides, and hence does little or nothing in response. For a wave coming normal to the plane of the membrane, the acoustic velocity of the sound wave will cause a small response. For sound coming along the plane of the membrane, the pressure is the same on both sides of the membrane, so that no response at all occurs. When a sound source is very close to the membrane, closer than the width of the memrane holder, the pressure on the source side is greater than on the side away from the source, so that a great response occurs due to this "gradient" situation. Such a membrane microphone can also be termed a "velocity" microphone since it can be driven by the acoustic velocity of acoustic wave it enconters. The "ribbon microphone" works on this principle, though the output voltage produced arises from motion in a magnetic field rather than from the deflection of a diaphragm. I have not seen any derivation of the near vs far field of the single membrane electret, though I believe it it be a worthy cause to pursue same. Angelo Campanella > I have not seen any derivation of the near vs far field of the > single membrane electret, though I believe it it be a worthy cause to > pursue same. See: "Understanding the Transfer Functions of Directional Condenser Microphones in Response to Different Sound Sources" by Guy Torio, available from the AES and: Unique Directional Properties of Dual-Diaphragm Microphones" by Guy Torio and Jeff Segota, also available from the AES or at www.shure.com I've tried to locate Guy Torio to see if I could get permission to host these papers at the micbuilders Yahoo mailing list site but he seems to have dropped off the face of the earth. Bob Signature"Things should be described as simply as possible, but no simpler." A. Einstein > > From my survey of the headset market (computer headsets, not cellphone > > headsets), it appears that most use an NC mike element that's [quoted text clipped - 19 lines] > microphone and talker, such remote sound is discriminated against in > comparison tothe talker sound near the microphone. It's worth noting that the smaller the mic capsule and it's housing, the greater the possible theoretical cancellation however, the mic placement has to be correspondingly close to the source for a reasonable frequency response* and effective cancellation. The spacing from the source becomes quite critical with very small electret capsules in small housings and spacing changes of a few millimetres can markedly affect frequency response and sensitivity. This may be a consideration in the choice of headsets to be used by students. A slightly larger capsule such as the dynamic unit used on Peltor high-noise headsets (no connection with me!) might be a good compromise in all but the noisiest environments - I have seen it in use by pit crews at Formula One motor races. *the simple pressure gradient mic has a free-field response having a first-order high pass characteristic with a turnover frequency determined principally by its dimensions. Countering that is a low-pass characteristic determined by the proximity of the source to the microphone. At optimum source spacing the response will be extended downward to the lowest frequency of interest - say 200 Hz. At greater-than-optimum spacing the low-pass frequency shifts upwards resulting in a net high-pass effect making the speech sound 'thin' as well as significantly quieter. Regards, Neil Adams Angelo -- thanks for the basic information on gradient mikes. Do you have any information (or pointers/references) on my questions (below) about directionality of the noise-canceling effect -- is it more effective in some directions compared to others? cheers, jerry > As I see it, we have at least 2 kinds of acoustic environmental > problems (in addition to the possibility of non-speech environmental [quoted text clipped - 5 lines] > can use such knowledge to advise users how best to position their > students when space is limited. The far-field (noise cancelling) directionality for a first-order gradient microphone is figure of eight. The normalized response is given by R=cos(theta) where R is the normalized response at angle theta, and theta is the angle between the direction of the impinging sound wave and the major axis of the microphone. Accordingly, the normalized response is unity for sound incident at 0 and 180 degrees and is zero for sound incident at +/- 90 degrees. > Angelo -- thanks for the basic information on gradient mikes. Do you > have any information (or pointers/references) on my questions (below) [quoted text clipped - 13 lines] >> can use such knowledge to advise users how best to position their >> students when space is limited. > Angelo -- thanks for the basic information on gradient mikes. Do you > have any information (or pointers/references) on my questions (below) > about directionality of the noise-canceling effect -- is it more > effective in some directions compared to others? I think "The Ghost" has adequately answered the directionality question. >>As I see it, we have at least 2 kinds of acoustic environmental >>problems (in addition to the possibility of non-speech environmental [quoted text clipped - 5 lines] >>can use such knowledge to advise users how best to position their >>students when space is limited. I'm not certain whether the figure-8 aspect, or the gradient aspect, are more prominent for a given "NC" microphone. Clearly, if the gradient aspect dominates for a given classroom set-up it would be the best of all. This means it would be both immune to the reverberant field part of the noise (coming from all directions simultaneously) as well as for local noise sources situated in one direction only. If only the figure-8 aspect of a given mic dominates, then one could expect reasonable performance in reverberant fields, but possibly problematic performance with nearby noise sources such as a specific competing talker or a noisy HVAC vent in front of or behind the mic user. So, even with allegedly "noise canceling" mics, careful classroom environmental testing specific to each classroom style by you is still necessary. Teachers vary in their classroom orientations. Gone is the row-and-column arrangement I grew up with. Now, it could also be around a table, in a living-room settee style, or random sit-on-floor and stand about style, or sitting alone in the corridor (favored for pragmatic satisfaction of this "no child left behind" era; where "Special Education" teachers ride the circuit in a school building, plucking out for a while each individual student known to be lagging and in need of spot tutoring). Who knows? When the headsets come out of the box who knows how they are going to be used? It's a cruel world out there. Hopefully, you will have the time and foresight to sort this NC/orinetation aspect out before you leave the scenario... Angelo Campanella --------- www.CampanellaAcoustics.com --------- "I have simply studied carefully whatever I've undertaken, and tried to hold a reserve that would carry me through." - Charles A. Lindbergh. "As for background noise level; 35 dBA is a good classroom; 45 dBA is a sound masking system!" - Anthony K. Hoover > I'm not certain whether the figure-8 aspect, or the gradient aspect, are > more prominent for a given "NC" microphone. Clearly, if the gradient [quoted text clipped - 7 lines] > problematic performance with nearby noise sources such as a specific > competing talker or a noisy HVAC vent in front of or behind the mic user. In the reverberant field both the gradient aspect and the directivity contribute to the overall noise cancelling performance. Which one of the two is more prominent depends on frequency. This is so because on-axis noise cancelling ability deterioriates with increasing frequency at a rate of 20dB per decade above the low-frequency cancellation assymtote. At frequencies below 200Hz-300H, where 23-25dB on-axis cancellation is possible, the contribution to overall noise cancelling performance due to figure-8 directivity is negligible. On the other hand, at 2KHz-3KHz where only 3dB-5dB on-axis cancellaation exists, the figure-8 directivity contribution is proportionally significant, albeit inconsequential. . > I'm not certain whether the figure-8 aspect, or the gradient aspect, are > more prominent for a given "NC" microphone. Clearly, if the gradient [quoted text clipped - 7 lines] > problematic performance with nearby noise sources such as a specific > competing talker or a noisy HVAC vent in front of or behind the mic user. There is another slight complication in their use for close-talking, such as with NC headsets. The front-back symmetry of a bi-directional mic is considerably disturbed by the close proximity of the face of the talker, as a result of which the far-field cancellation is reduced. To counter this, most mfrs mechanically attenuate the acoustic entry into the far side of the capsule. It's a compromise and, again, very dependent on the exact spacing of the mic from the mouth, adding to the need for exact placement to obtain best performance. Possibly, in a classroom environment, especially if the students are around a table and are not actually so close they are rubbing shoulders, a hypercardioid mic for each one at a distance of 12" ~ 18" would offer sufficient discrimination and background attenuation (do they have to wear headsets?). Regards, Neil adams |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2009 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.