 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Natural Science Forum / Physics / Acoustics / September 2005Listening bats, something missing?
Bats make ultrasonic sounds. Different species of bat can be identified by their sounds. There are different types of bat sound detectors. Heterodyne seems to be most common. Time division and time expansion are other methods used in detectors. All detectors convert ultrasound to normal hearing range. Listening to bats is thrilling. I suspect, that something of the real sounds of bats is missing after the conversion from ultrasound to "normal" sound. I have not been able to find, what is missing, or is anything... If a bat makes sounds in the region of 16 kHz to 22 kHz, the sounds include frequency shift, pulses with intrapulse modulation and there are different types of pulses and pulse frequencies. With a heterodyne detector the system subtracts eg 15kHz signal from the ultrasound and the result is somewhere between 1kHz and 7kHz. The question is, what do we lose in this procedure? If the bat makes a pulse with 10 wavelengths, when the frequency drops eg from 20kHz to 5kHz, how many wavelenghts are there within the pulse? Does the conversion distort something? Time division detectors divide the frequency by eg 8 or 16 sometimes with a cheap digital counter. Somethig is lost again in the process. Time expansion takes a sample, eg 10 seconds and plays it slower, perhaps in 100 seconds. What are we losing here? If you think of music and cd-players, can you invent a system that compresses a song in tenth part of space without loss? Dropping 44kHz sample rate to 4kHz? My real question is, how can I catch and analyse bat sounds so that no information is lost in the process? Can I use current detector techniques or must I create something new? > Bats make ultrasonic sounds. Different species of bat can be identified by > their sounds. There are different types of bat sound detectors. Heterodyne [quoted text clipped - 30 lines] > information is lost in the process? Can I use current detector techniques or > must I create something new? I built an ultrasonic converter based upon a divider chip. What makes you think something is being lost in the conversion? And unless you're another bat, why would a small loss-- if any-- matter to you so much? Ron "Ron Hubbard" <notat@hotmail.com> kirjoitti viestissä:11hngic611pkaa@corp.supernews.com... >> Bats make ultrasonic sounds. Different species of bat can be identified >> by [quoted text clipped - 50 lines] > > Ron To be able to identify different species, the smallest differencies in the sounds are very important. The importance for me is "only" academic, but there is something more important behind here. Some of the bats are classified as endangered species here. If someone wants to build a house near a possible habitant of some rare animals or plants, the living area of those animals and plants must be mapped. If an endangered bat is found somewhere, no alteration of the habitat is allowed. Some people just ignore the bats, some want to know exactly their species and home areas. I am trying to record bat sounds to my pc. There are so many different possibilities. I am just trying to find out the best one. The main requirement is to make the recording as pure as possible without no loss of information. > "Ron Hubbard" <notat@hotmail.com> kirjoitti > viestissä:11hngic611pkaa@corp.supernews.com... [quoted text clipped - 67 lines] > requirement is to make the recording as pure as possible without no loss of > information. Okay. Here's a suggestion that may work try taping the bat sounds on a good tape recorder. I *think* there are microphones designed for ultrasonic uses but I'm not sure. Once you have it on tape, use a program like Cool Edit to study and edit the sound recordings. While you won't be able to hear the ultrasonics, the sounds will be stored in the computer along with the audible sounds. You can get really good quality recordings if you use a high sampling rate like those used for CD quality recordings (44 kHz). While it's probably not a "perfect" solution to your problem, it will give you better quality recordings than just feeding the sounds into your PC. Ron >> "Ron Hubbard" <notat@hotmail.com> kirjoitti >> viestissä:11hngic611pkaa@corp.supernews.com... [quoted text clipped - 97 lines] > > Ron By simple search I found that bats hearing may cover the frequency range from about 9 kHz to 160 kHz or more. So for listening you would have to divide the frequency or speed by at least 8.... Even a normal "good" tape recorder will not record much over 15-20 kHz, but there are of course special ones. (But Ron, why the tape recorder if you can record.the same quality directly on the harddisk?) Cool Edit is now called Adobe Audition, and is an excellent sound recording, editing, converting and analyzing tool and supports sampling rates up to 192 kHz. Sampling rate must be more than twice the highest frequency of interest, so the highest recordable frequency with that will be almost 96 kHz and you will still loose something. And you will also need a special soundcard and a special microphone system to cover this sampling and recording frequency range, and a very fast PC.... Asbjørn > >> "Ron Hubbard" <notat@hotmail.com> kirjoitti > >> viestissä:11hngic611pkaa@corp.supernews.com... [quoted text clipped - 105 lines] > (But Ron, why the tape recorder if you can record.the same quality directly > on the harddisk?) From personal experience with Cool Edit I find that recoeding fro a tape recorder has a couple of advantages like being able to adjust the incoming signal strength by just changing the volume control on the recorder. If the signal is too low, CE won't be able to recognize it but if the volume is too high, clipping and amplitude distortion might occur. Also, since you have the sounds on tape, if the ediiting process goes wrong somehow, you can do it over and over agian. If somehow you feed the sounds directly to the sound card, your dependent upon nothing going wrong otherwise the sounds would get lost. But like I said, it was only a suggestion. :-) However the OP may want to look into the ANABAT system: http://srmwww.gov.bc.ca/risc/pubs/tebiodiv/bats/batsml20-05.htm He may find what he needs. Ron > Cool Edit is now called Adobe Audition, and is an excellent sound recording, > editing, converting and analyzing tool and supports sampling rates up to 192 [quoted text clipped - 5 lines] > > Asbjørn Snip >> > Okay. Here's a suggestion that may work try taping the >> > bat sounds on a good tape recorder. I *think* there are [quoted text clipped - 47 lines] > > Ron OK, especially to that CE will not record if the signal starts too low, but besides that I can see no more difficulty in adjusting signal level to CE than to a tape recorder, and you can always keep a copy of the original recording with CE as well. Asbjørn > OK, especially to that CE will not record if the signal starts too low, > but besides that I can see no more difficulty in adjusting signal level to > CE than to a tape recorder, > and you can always keep a copy of the original recording with CE as well. I use CEP considerably and it will record whatever happens to be at the inputs when you arm a track for recording and punch start. It doesn't even look at the values on the inputs other than to put them in a file. Bob  Signature"Things should be described as simply as possible, but no simpler." A. Einstein >> OK, especially to that CE will not record if the signal starts too low, >> but besides that I can see no more difficulty in adjusting signal level [quoted text clipped - 6 lines] > > Bob Well, I have experienced quite a few times when multitrack recording with the newest version (Adobe Audition 1.5) that the track with very low signal stops recording after a few seconds, but the other tracks go on. Tried to reproduce that flaw now on another PC, but here it works fine. I never had that flaw with Cool Edit Pro 1.2. With version 2 I'm not sure. I thought this was what Ron mentioned but may have misunderstood.. Cool Edit / Adobe Audition is a very fine recorder and toolbox, and the wav files can also be handled by other tools like MatLab. Asbjørn > >> OK, especially to that CE will not record if the signal starts too low, > >> but besides that I can see no more difficulty in adjusting signal level [quoted text clipped - 17 lines] > Cool Edit / Adobe Audition is a very fine recorder and toolbox, and the wav > files can also be handled by other tools like MatLab. Hi, Asbjorn; I have been using the first, freeware version of Cool Edit, which isn't capable of recording a signal if the input is a bit low. I've tried later versions, but I personally prefer the original despite a couple of very minor flaws. I don't know if Adobe fixed the input level problem, but it's a kind of minor thing, and you're right, it is a good program to have around. Ron > Well, I have experienced quite a few times when multitrack recording with > the newest version (Adobe Audition 1.5) that the track with very low signal [quoted text clipped - 5 lines] > Cool Edit / Adobe Audition is a very fine recorder and toolbox, and the wav > files can also be handled by other tools like MatLab. I think you are encountering a problem with your device rather than with CEP itself. With 1.2, 1.5 and 2.1 I've done a fair bit of recording of resistive shorted inputs of various things in order to measure the noise floor, less than -90 dB FS, without seeing such a problem. The inability to reliably record silence would render the program pretty useless for all kinds of studio recording and would be much discussed on their forum, which it hasn't been. Bob Signature"Things should be described as simply as possible, but no simpler." A. Einstein > > Well, I have experienced quite a few times when multitrack recording with > > the newest version (Adobe Audition 1.5) that the track with very low signal [quoted text clipped - 15 lines] > program pretty useless for all kinds of studio recording and > would be much discussed on their forum, which it hasn't been. Bob, it's such a minor problem, I don't worry about it. Ron . > If a bat makes sounds in the region of 16 kHz to 22 kHz, the sounds include > frequency shift, pulses with intrapulse modulation and there are different > types of pulses and pulse frequencies. OK.. That's a good start. The small wavelength rage, about 1.5:1, implies radar functioning. > With a heterodyne detector the system subtracts eg 15kHz signal from > the ultrasound and the result is somewhere between 1kHz and 7kHz. > The question is, what do we lose in this procedure? I think that heterodyning is not a good choice, because these are not speech sounds. This is a pulse echo system, and time sequencing is most important. Therefore time-strecthing is better. And we do not nee a big ratio. 3 to one (5 to 7 kHz, or for old ears, about 6:1 will do fine. > If the bat makes a pulse with 10 wavelengths, when the frequency drops > eg from 20kHz to > > 5kHz, how many wavelenghts are there within the pulse? That's the problem with heterodyning sonar signals... Does the > conversion distort something? A mess! > Time expansion takes a sample, eg 10 seconds and plays it slower, perhaps in > 100 seconds. What are we losing here? > > If you think of music and cd-players, can you invent a system that > compresses a song in tenth part of space without loss? Dropping 44kHz sample > rate to 4kHz? You are making it too complicated. Just take a 6:1 slowed down playback, listen, and think of spaces six time as big, moths with a 6" wingspan and let your imagination do the walking while you listen! Angelo Campanella "Angelo Campanella" <a.campanella@att.net> kirjoitti viestissä:%26Te.13604$qY1.556@bgtnsc04-news.ops.worldnet.att.net... >> If a bat makes sounds in the region of 16 kHz to 22 kHz, the sounds >> include frequency shift, pulses with intrapulse modulation and there are [quoted text clipped - 3 lines] > as big, moths with a 6" wingspan and let your imagination do the walking > while you listen! The problem with slow playback is in the recording. 96kHz sound card is not enough. Propably 192kHz sampling rate is near enough (for 100kHz signals), but I think I need 16bits at that rate. 1bit is not enough. A portable pc would be enough, but the sound card for pcmcia with 192kHz capability is more expensive than my car... I think that I have to build an A/D-converter myself. An ultrasonic microphone with a preamplifier and a converter can be built and connected via usb to a pc. The next problem would be calibrating the analog section (sensitivity vs freq). Thanks for everybody for ideas. Now I am convinced that a heterodyne or frequency divider is not the perfect solution. > I think that I have to build an A/D-converter myself. An ultrasonic > microphone with a preamplifier and a converter can be built and connected > via usb to a pc. The next problem would be calibrating the analog section > (sensitivity vs freq). I build and sell an ultrasonic reference sound source (URSS). It is calibrated to 100 kHz. Typically, a 1/4" condenser microphone (common instrumentation microphone, grid cap removed for use). I'm looking at options for 100-300 kHz.. Angelo Campanella www.campanellaacoustics.com Forgetting the specifics of the recording process for the moment, there is indeed something missing The ear is limited to a 20 kHz bandwidth at its absolute best. You should really count on 10 kHz. The bat vocalisations occupy a bandwidth considerably larger so unless you expand time and listen to the calls at a slower rate you are missing something. The time bandwidth product is a constant and something has to give. Heterodyning to baseband is fine but you will only hear the bottom 10 kHz. You cannot compress the bandwidth and retain the same observation time without losing information. I ascertained this in discussions with a professor of underwater acoustics when discussing dolphin "chirp" vocalisations. On the specifics of the recording process the bandwidth of the initial tape recorder is important/critical and if the purpose of the tape recorder is to manage dynamic range why not utilise one of the 24 bit 96 kHz sound cards now available. They have a much larger dynamic range. > Bats make ultrasonic sounds. Different species of bat can be identified by > their sounds. There are different types of bat sound detectors. Heterodyne [quoted text clipped - 30 lines] > information is lost in the process? Can I use current detector techniques or > must I create something new? |
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.