 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Natural Science Forum / Physics / Acoustics / December 2003Detect "breath pops" in headset-microphone signal?
Many speech recognition system users use a headset microphone, which in addition to the advantages it offers, also presents some challenges, one of which is the importance of positioning the microphone to the side of the mouth, out of the breath stream. Failure to do so results in "breath pops", particularly during front consonants, which I assume are due to turbulence due to the breath stream at the microphone. Most headsets come with a foam "wind sock" on the microphone, but this does not eliminate the problem. Has anyone done any work on detecting these breath pops in a microphone signal? My motivation is to alert the user to the fact that the microphone is positioned incorrectly (and to induce him to correct it so we can have a better signal going to the recognizer). jerry wolf Jerry, > Has anyone done any work on detecting these breath pops < I've used a multi-band compressor to eliminate the effects of pops quite well. The idea is to hard limit only frequencies below about 80 Hz. If you happen to have the UltraFunk plug-ins pack I can send you a pop removal preset I created for the multi-band compressor plug-in. As for detecting pops with circuitry or DSP code, you need to identify sudden bursts of low frequency energy. I designed a pop eliminator many years ago and wrote an article about it for the August 1981 issue of R-e/p magazine. That article, with schematics, is on my web site: www.ethanwiner.com/articles.html Look for "Audio filters - theory and practice," about a third of the way down the page. --Ethan > As for detecting pops with circuitry or DSP code, you need to identify > sudden bursts of low frequency energy. I designed a pop eliminator many [quoted text clipped - 5 lines] > Look for "Audio filters - theory and practice," about a third of the way > down the page. If breath pops were only very low frequency then they wouldn't cause so much trouble because they would be filtered out at some stage between the microphone and the final listener. Problem is, they produce a very high level signal so at some point they cause overload. This causes a wideband output that you can't completely get rid of by filtering. Overload often happens very early in the chain, could even be the microphone diaphragm. If you have access to the amplifier that is overloading you could detect when the signal level is getting too near to the supply rail. Bear in mind that the signal will be very asymmetrical. Tony Woolf Tony, > they produce a very high level signal so at some point they cause overload. < Yes, you are absolutely correct. But Jerry asked how to detect pops, and this is how you do it. Also, in practice, what I described works pretty well. Obviously if the mike capsule or preamp overloads and "farts out" there's no way to clean it up and make it sound perfect. But dropping the bass level as described in my article, or using the multi-band compressor preset I mentioned, does do a decent job after the fact. --Ethan > Tony, > > they produce a very high level signal so at some point they cause > overload. < > Yes, you are absolutely correct. But Jerry asked how to detect pops, and > this is how you do it. Also, in practice, what I described works pretty > well. Obviously if the mike capsule or preamp overloads and "farts out" > there's no way to clean it up and make it sound perfect. But dropping the > bass level as described in my article, or using the multi-band compressor > preset I mentioned, does do a decent job after the fact. Fix single click in CE/Audition can solve many and very varied issues, including gospel singers hitting a mic stand with a tambouring, members of the audience kicking the mic stand, something falling off of a chair on stage (x), clicks and pops in phonograph records, FM multipath crackle from an airplane waiting to land at lane 12-30 in Kastrup, the guy at 3' row who claps louder than everybody else and quit probably also blast noises, overloaded or not. (x) partly, the natural reverb of it in the hall can not be removed .... but it can be made a lot less blatantly obvious. > --Ethan Kind regards Peter Larsen  Signature************************************************************* * My site is at: http://www.muyiovatki.dk * ************************************************************* |
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.