Poll: Can you hear sound over 20kHz? - Page 28

Nope

does it really matter what headphones you use? I am asking because i cant hear 17khz on turtle beach x12 but i can hear 17khz fine on my OEM galaxy s2 earphones...

Quote:

Originally Posted by jake120 

does it really matter what headphones you use?

It does at least to some extent, there can be large differences in the frequency response of headphones in the top octave of the audio band. Although human hearing does drop off rather steeply at the upper end of its range.

There have been studies on human perception of music and none of them have indicated that frequencies above the range of human hearing add anything at all to music. In fact, the frequencies above 10kHz do very little.

> music is all about frequencies interacting with each other

Frequencies don't "interact" with each other. Okay, so I understand what you're saying. It may be intuitive to think that by "combining" the ultrasonic frequency with the other frequencies, it would make a difference in sound, since the signal is different. However, you'll get the same result as if you were listening to that ultrasonic frequency on its own.

Firstly, any naturally-occurring signal can be exactly represented by a linear combination of periodic complex exponential signals (technically harmonically-related). You don't have to completely understand what that means -- just that a naturally-occurring signal is made up of fundamental signals that have their own frequency. Kind of like the 22 kHz tone. Our ears process signals (sound waves) as if they were dealing with those fundamental frequencies independently. A 22 kHz frequency component, whether it is on its own or superimposed with other frequencies, is treated the same. Indeed, the way we perceive sound is the same as how a filter works.

A 22kHz frequency component is processed by not being heard at all.

Frequencies overlapping other frequencies don't change pitch. But frequencies at higher octaves can mask ones at octaves below.

Yes, two sine waves with the same frequency but out of phase can cancel each other, but that has nothing to do with how we hear two sine waves with different frequencies.

The hair cells in our ears are tuned to narrow frequency bands. If there are no hair cells for a certain frequency (lets say 21 kHz) you cannot hear the sound.

Yes, ultrasonics can make a difference at playback because of intermodulation distortion (IMD). IMD is bad, very bad.

Quote:

Originally Posted by Venasa 
Thank you, thats rather neat to know and im pretty sure I get what your saying. Is it possible however for frequencies to sort of modulate other frequencies. So say I was listening to 9kHz and then added 13kHz, would the combined signal really still be just 9 & 13kHz or something closer to 8 & 14? These things come to mind because I remember from my SAE course that you can can cancel out a pure sine wave by copying it and phase shifting it, then playing them together, establishing at least that there is some interesting interactions between frequencies

If the sine wave was out of phase by 180 degrees, then yes, they would cancel each other out, because they would be additive inverses of each other. sin(x)+[-sin(x)]=0

If you combined 9 kHz and 13 kHz, the output signal would have energy at those frequencies and only those frequencies. The time domain of a signal has a linear relationship with the frequency domain. That is, ax(t)+by(t)+...=aX(iw)+bY(iw)+... where the left-hand side represents the linear combination of signals, and the right-hand side represents the linear combination of the corresponding Fourier transform of those signals, with w being the angular frequency and i=sqrt(-1). With the sine wave example, the Fourier transforms are also additive inverses of each other, so the spectral coefficients cancel each other out (yes, a signal can have negative spectral coefficients, because they represent the amplitude and the phase).

xnor makes a point about IMD. In practice, due to distortion, the above doesn't necessarily hold.

Edit: I've revised this post because I didn't like how I was using the term "frequency component". A frequency component of a signal is simply one of the frequencies. I should have said "spectral coefficient", which is what the Fourier transform calculates. The magnitude of a spectral coefficient, squared, |X(iw)|^2, is the energy. The frequency spectrum graphs you usually see are for energy vs frequency. The notation I used for the linear relationship is also misleading (it's not an equality), but I don't want to make this post any more confusing!

I like answers that are so complex it makes people not want to ask questions any more.

Quote:

Originally Posted by bigshot 

I like answers that are so complex it makes people not want to ask questions any more.

True, I should have summarised what I was saying in that second paragraph. Basically, if you combine two or more signals that have frequencies X, Y and so on, the output signal will have energy at those frequencies and only those frequencies. Frequencies don't "interact" with each other. I hope that makes sense.

Now THAT I can understand! I need a copy of "Physics for Dummies".

As a physicist, I would like to continue my earlier comment by stating that uncontrolled tests with equipment of unknown characteristics may be leading many people to think they have hearing that extends into the inaudible.  This is bad science.  There is the burden of proof using real lab gear, and that, as far as I can tell, has not been demonstrated on this forum.  The physics of hearing and sound can't be fooled.  Regarding my one student who claimed hearing just over 20K, age and gender may allow it, but I did not have a chance to verify the physics of what was happening to my satisfaction.  There was some research way back about defeating the low pass filter of the middle ear by using bone conduction, however, I was never able to test this with my students.  Put a 19KHz ultrasonic transducer on the skull - see if there is a difference in perception.  There was a limit to what I was willing to send through students without signed releases.  Oddly, the lab had an x-ray tube from around 1960, and the included experiment was to allow students to view hand bones using the included fluoroscope.  How times change...soon all actual experiments will be eliminated as being too scary, and far fewer kids will ever develop an interest in science.  Kids need to touch the real thing.  I remember my first time alone with an electron microscope - didn't come out for days, changed my life.

Anyway, using uncalibrated computer tone generators through low-bidder (even what we consider high-end) sound cards into unknown output electronics into transducers of dubious pedigree needs to be replaced with lab-quality, calibrated function generators and transducers.  For those who have never used such equipment, I can assure you that there is a difference.  Calibration is a key to getting this verified.  Also, it is very easy to verify what was said above about energies and frequencies - just get two generators, a scope, and minimal components.  Mix, watch.  Oh, as long as you have two frequency sources, put one to X and one to Y on the scope and watch the lissajous patterns for hours.  Mix in other waveforms (triangle, square) for hypnotic patterns.  Hours of fun for those of us who dislike everything about computer games.  I will post a few pics if I get a chance to set it all up on a bench.  For anyone who wants to see a video of a sound experiment lead by one of the coolest physicists ever, google the TED talk by Cliff Stoll.  You'll enjoy his talk, and you can do it at home.  Experiment - enjoy!