Can You Hear That Dog?
July 16, 2005 3:41 PM Subscribe
How can I generate an audio frequency only dogs can hear?
This link says "Dogs can hear frequencies up to about 35,000 Hz (35 kHz)".
I hope to use my computer and home monitors (BostonT830's) although i'm not sure if they are capable of broadcasting in that frequency range.
Any home methods for doing this? I have several audio apps to work in but i'm not sure how to accomplish this. Or is there a wav I can download? Thanks!
This link says "Dogs can hear frequencies up to about 35,000 Hz (35 kHz)".
I hope to use my computer and home monitors (BostonT830's) although i'm not sure if they are capable of broadcasting in that frequency range.
Any home methods for doing this? I have several audio apps to work in but i'm not sure how to accomplish this. Or is there a wav I can download? Thanks!
I've always thought that most speakers would be OK -- there's always the dog whistle at the end of "A Day in the Life" on Sgt. Pepper, because John thought it'd be funny if someone was just listening to the record and their dog started going crazy; I remember a long time ago, forgetting about this and listening to the album with a dog in the room, but it could be a manufactured memory and/or the dog could have seen something and decided to bark -- so does anyone have any actually better-remembered experiences with this album? Because that might go to answer this particular question, at least in the instance of whether or not you need any particularly special monitors to play that frequency of sound.
posted by Rev. Syung Myung Me at 3:59 PM on July 16, 2005
posted by Rev. Syung Myung Me at 3:59 PM on July 16, 2005
I doubt home speaker systems operate much outside the range of human hearing (for signal processing reasons that I won't go in to right now), but it's not inconcievable that your speaker can emit frequencies just outside human hearing - around 25-30 kHz. Check the labels on your speakers, or if you still have the manual, it should print the frequency response function. If I wanted to do this, I would use a program like csound, which can generate simple sine waves at any frequency, but that program has a pretty steep learning curve, and there must be better options.
posted by muddgirl at 4:39 PM on July 16, 2005
posted by muddgirl at 4:39 PM on July 16, 2005
You won't be able to use a computer unless you have a 96kHz or 192kHz DAC. This is because of the Nyquist frequency (well, I didn't read that link, so I'm not 100% sure it is a good explanation).
A home theatre speaker should be able to produce the sound, the volume might be attenuated because you are working out of the design range of the speaker, but dogs hearing is very sensitive. Which brings up a major problem, because you don't want to hurt said pooch.
The amplifier needs to be able to do it too, but most amplifiers probably can.
posted by Chuckles at 4:41 PM on July 16, 2005
A home theatre speaker should be able to produce the sound, the volume might be attenuated because you are working out of the design range of the speaker, but dogs hearing is very sensitive. Which brings up a major problem, because you don't want to hurt said pooch.
The amplifier needs to be able to do it too, but most amplifiers probably can.
posted by Chuckles at 4:41 PM on July 16, 2005
This study indicates that humans have a hearing range of approximately 64 - 23000 Hz, while dogs have a hearing range of approximately 67 - 45000 Hz. So you'd need to produce a tone with a frequency somewhere between 23000 Hz and 45000 Hz.
The problem with producing a tone like this digitally is that most common digital sound equipment (like CD players, or most standard sound cards) have a maximum sampling frequency of 44100 Hz. This sounds like it'll be high enough, but thanks to the Nyquist-Shannon sampling theorem, you can't accurately reproduce a sound unless it has a frequency less than half the maximum sampling frequency of your device. So you'll only be able to accurately produce sounds with a maximum frequency of 22050 Hz.
You can hear the effects of this on the Sgt. Pepper album. On the CD version, there is a perfectly audible sound at the end of the track, which is the infamous dog whistle. However, on an analogue original (like the vinyl version), you can't hear anything at that point. The original signal is still there, but it's at too high a frequency for you to hear (or, possibly, for your speakers to reproduce loud enough). On the CD version, the sound has been aliased down to a lower frequency sound thanks to the limited maximimum sampling frequency of the CD format.
Erm, so yeah, I think a dog whistle is your best bet.
posted by chrismear at 4:44 PM on July 16, 2005
The problem with producing a tone like this digitally is that most common digital sound equipment (like CD players, or most standard sound cards) have a maximum sampling frequency of 44100 Hz. This sounds like it'll be high enough, but thanks to the Nyquist-Shannon sampling theorem, you can't accurately reproduce a sound unless it has a frequency less than half the maximum sampling frequency of your device. So you'll only be able to accurately produce sounds with a maximum frequency of 22050 Hz.
You can hear the effects of this on the Sgt. Pepper album. On the CD version, there is a perfectly audible sound at the end of the track, which is the infamous dog whistle. However, on an analogue original (like the vinyl version), you can't hear anything at that point. The original signal is still there, but it's at too high a frequency for you to hear (or, possibly, for your speakers to reproduce loud enough). On the CD version, the sound has been aliased down to a lower frequency sound thanks to the limited maximimum sampling frequency of the CD format.
Erm, so yeah, I think a dog whistle is your best bet.
posted by chrismear at 4:44 PM on July 16, 2005
Oh ya, you probably could rig something up using a serial port or something. All you really need is a signal that turns on and off fast enough and a low pass filter to make it into a sign wave. So if you forget about the sound card, and sound software, you can probably finesse something.
Argh, three posts in a row, argh!
posted by Chuckles at 4:45 PM on July 16, 2005
Argh, three posts in a row, argh!
posted by Chuckles at 4:45 PM on July 16, 2005
Response by poster: Hmmm....there are some ideas here. Yeah I know I can go the old "analog" way and just buy a damn whistle but it seems possible enough to do it on my setup.
As I mentioned my speakers are Boston T830's and my amp is JVC RX-554V, i've got an m-audio quattro that i use with cubase as well as other audio apps...
I guess I'll check the range of my monitor's before I move forward. thanks guys!
posted by freudianslipper at 5:21 PM on July 16, 2005
As I mentioned my speakers are Boston T830's and my amp is JVC RX-554V, i've got an m-audio quattro that i use with cubase as well as other audio apps...
I guess I'll check the range of my monitor's before I move forward. thanks guys!
posted by freudianslipper at 5:21 PM on July 16, 2005
Err, for DAC read Sound Card...
For some reason, this made me laugh very loudly...
posted by muddgirl at 5:26 PM on July 16, 2005
For some reason, this made me laugh very loudly...
posted by muddgirl at 5:26 PM on July 16, 2005
Well, it appears that the m-audio quattro supports 96kHz sampling (most m-audio products probably do, so if that isn't the right one...). Test Tone Generator should therefore be able to generate a ~30kHz signal.
The specifications for your speakers may not give you much useful information. They probably don't specify the out of band characteristic very carefully, it is out of band after all. The speakers will not simply stop working at a certain frequency, they just won't work as well.
Finally, always be careful when driving speakers with a test tone. Speakers and amplifiers are designed to deal with music, which has much less power than a test tone at the same level. Keep the 'beeps' brief, and leave time in between.
muddgirl, I don't get it...
posted by Chuckles at 6:16 PM on July 16, 2005
The specifications for your speakers may not give you much useful information. They probably don't specify the out of band characteristic very carefully, it is out of band after all. The speakers will not simply stop working at a certain frequency, they just won't work as well.
Finally, always be careful when driving speakers with a test tone. Speakers and amplifiers are designed to deal with music, which has much less power than a test tone at the same level. Keep the 'beeps' brief, and leave time in between.
muddgirl, I don't get it...
posted by Chuckles at 6:16 PM on July 16, 2005
Small piezoelectric transducers should be able to go that high. Trouble is, if you try to make a "speaker" out of one by attaching a plate or a cone, the extra mass will slow it down. Tunable analog electronics can generate the frequency if your sound card can't.
On preview: Oooh look, some OEM solutions.
posted by fatllama at 7:03 PM on July 16, 2005
On preview: Oooh look, some OEM solutions.
posted by fatllama at 7:03 PM on July 16, 2005
and a low pass filter to make it into a sign wave.
1) To some extent any reasonable consumer-grade speaker will do this automatically.
2) If one generates a square wave at the edge of the canine hearing spectrum, the other harmonics that make up the wave won't be heard and the sound will be indistinguishable from a pure sine wave at the fundamental frequency.
and for the poster:
If the radio frequency spectrum, one can buy frequency doublers that have rectifying elements (diodes) inside; conversion of one frequency into twice that frequency is about 10% efficient. Maybe a similar solution exists in the audio frequency domain, likely a PIN diode. Anyone?
posted by fatllama at 7:08 PM on July 16, 2005
1) To some extent any reasonable consumer-grade speaker will do this automatically.
2) If one generates a square wave at the edge of the canine hearing spectrum, the other harmonics that make up the wave won't be heard and the sound will be indistinguishable from a pure sine wave at the fundamental frequency.
and for the poster:
If the radio frequency spectrum, one can buy frequency doublers that have rectifying elements (diodes) inside; conversion of one frequency into twice that frequency is about 10% efficient. Maybe a similar solution exists in the audio frequency domain, likely a PIN diode. Anyone?
posted by fatllama at 7:08 PM on July 16, 2005
Your sound card's output is filtered at 22 kHz, so that won't work.
fatllama: You can rectify an audio signal with a diode bridge. In that case it's best to start with a tiangle wave instead of a sine, as the rectified triangle is still a triangle, with little high harmonic contents. Such a circuit requires two op-amps and one or two diodes and probably isn't worth the trouble.
A better solution is probably to build a square wave oscillator, as fatllama said. The simplest circuit I can think of consists of a Schmitt trigger inverter, an capacitor and a potentiometer (variable resistor). A schematic is available here. Attach a piezo speaker to this. They are cheap and efficient at high frequencies. If you just want to check the effect out briefly, I second the suggestions above about dog whistles.
posted by springload at 4:39 AM on July 17, 2005
fatllama: You can rectify an audio signal with a diode bridge. In that case it's best to start with a tiangle wave instead of a sine, as the rectified triangle is still a triangle, with little high harmonic contents. Such a circuit requires two op-amps and one or two diodes and probably isn't worth the trouble.
A better solution is probably to build a square wave oscillator, as fatllama said. The simplest circuit I can think of consists of a Schmitt trigger inverter, an capacitor and a potentiometer (variable resistor). A schematic is available here. Attach a piezo speaker to this. They are cheap and efficient at high frequencies. If you just want to check the effect out briefly, I second the suggestions above about dog whistles.
posted by springload at 4:39 AM on July 17, 2005
Another non-gear way (well, I don't consider a dog whistle to be "gear" in this case) to produce high frequency sounds is to purse your lips tightly and sort of inwards and blow hard. Your lips and surrounding area should puff out a little so you look like a chimpanzee. As air escapes, it may sound like it does when you let the air out of a balloon very slowly: a squealy, squeaky sound. In between the sounds you can hear are sounds that your dog or small rodent pet will perk their ears up at.
posted by redteam at 5:05 AM on July 17, 2005
posted by redteam at 5:05 AM on July 17, 2005
fatllama, great point about the natural low pass filtering characteristic of speakers.
springload, freudianslipper's sound card is 96kHz capable, I suspect that means that he can output signals well into the octave above 20kHz. (I don't know this for certain, after all none of us could hear it even if it was there, but there wouldn't be any point supporting 96kHz if you can't output it...)
Anyway, your diode idea is pretty good, but it wouldn't do any good to make it a triangle wave... The harmonics of the triangle wave will have to pass the 22kHz reconstruction filter anyway. So, whatever it looks like on your computer, if the signal's fundamental is above 10kHz or so it will look like a sinusoid on the output.
But the important question is: freudianslipper, have you tried it out yet? What did the poor pooch do? My Mom's dog freaks out when you disturb the crystals in a chandelier, it is hilarious!
posted by Chuckles at 6:13 AM on July 18, 2005
springload, freudianslipper's sound card is 96kHz capable, I suspect that means that he can output signals well into the octave above 20kHz. (I don't know this for certain, after all none of us could hear it even if it was there, but there wouldn't be any point supporting 96kHz if you can't output it...)
Anyway, your diode idea is pretty good, but it wouldn't do any good to make it a triangle wave... The harmonics of the triangle wave will have to pass the 22kHz reconstruction filter anyway. So, whatever it looks like on your computer, if the signal's fundamental is above 10kHz or so it will look like a sinusoid on the output.
But the important question is: freudianslipper, have you tried it out yet? What did the poor pooch do? My Mom's dog freaks out when you disturb the crystals in a chandelier, it is hilarious!
posted by Chuckles at 6:13 AM on July 18, 2005
Chuckles: The spec says "frequency response: 22Hz-22kHz", so I assume the output cuts off at 22kHz. There is indeed little point in sampling at 96kHz, except that the extra data can be used for signal processing, and gives some fault margin for imperfections in the analog circuits. Good point about the triangle though. It will end up almost as a sine at high frequency, which gives a lot of harmonic content if it's rectified.
posted by springload at 3:42 AM on July 19, 2005
posted by springload at 3:42 AM on July 19, 2005
This thread is closed to new comments.
posted by vacapinta at 3:52 PM on July 16, 2005