Beat Frequencies: Newsletter Challenge (01/22/08)

I agree entirely. Beat frequencies which are audible are the result of air vibrations. If they are audible in the concert hall, they will be audible in a faithful recording with mics at ear positions*. This is easily proved by recording audible beat frequencies. Beat frequencies are not something the only exist in one's mind, neurons, or that require processing beyond the eardrum. If there are beat frequencies to be heard, then they can be recorded.

Simply said: If you can hear it, you can pick it up (and record it) given a good microphone.

If a recording "sounds different" with ultrasonic (by definition unhearable) components added back in, then we are simply redefining the limits of hearing aren't we? If the recording "feels" different (tactilely or emotionally) then we are postulating a means other than "hearing" for responding to vibrations beyond 20,000 Hz.

.sorell :er noitavresbo ruoy deyojne I.

*This sidesteps the whole non sequitur discussion re the mechanism required to make a beat frequency audible, if neither frequency contributing to the beat is audible.

The professor is correct in that beat frequencies will occur, but there is a more important effect occurring at the speaker that affects the sound. Suppose we have a large low-frequency component, say 55 Hz (key of A) with a smaller high-frequency component superimposed on top of it. As the speaker moves toward the listener, the doppler shift caused by the 55 Hz tone will cause a rise in frequency of the high-frequency components and as the cone recedes, it will cause a drop in frequency. The modulation of the high-frequency tone produces a spread-spectrum distortion caused by frequency modulation. Since the speaker is the only item with large (say, 0.5") excursions which the ear does not have, it should be easy to calculate the maximum speed of the speaker cone, compare it to the speed of sound, and get a maximum frequency excursion.

Some music synthesizers (starting with the Yamaha DX-7) use frequency modulation synthesis to provide approximations of the sound of other instruments, and this happens to a lesser extent here.

This effect is reduced by using separate speakers for different frequencies.

Everyone seems to be caught up in the recodability of beat frequencies and reproduction thereof. Simply put: if you were to take a wall of microphones of varying sizes and threasholds, and wire each one to a similar sized wall of speakers of similar qualities to the microphone they are wired to; the result would still be an audibly poor product. The reason is regardless how many places that the sound is recorded in a given space, interactions from other frequencies in 3 dimentions and many directions shape that sound as it is percieved by our ear, but a microphone still records, and the speaker still reproduces this sound as a mostly single-direction waveform. Much like a "lens" for the sound waves.

Recordings have been able to reproduce beat frequencies in headphones by using "binaural beats" which are interpreted by the brain only-i.e. changing the timing and/or volume of tracks that are shared by both the R and L ear. This method is prohibitively expensive, and loses its effect when many instruments of varying volumes/waveshapes are being recorded at the same time. Another method used 2 microphones a "head-width" apart reproducing the shape of the ear, but again they were ineffective unless the speaker could be placed in the same area of the ear that it was recorded in. The other problem with speakers, is even if you were able to record at absolute realism, the speakers would create a certain ammount of "crosstalk" that is impossible in a concert hall.

The solution: Other than the fact that the student is going to fail because he is showing up the prof....Ambiophonic technology is the answer. Two single points in space that represent L and R ear ("headwidth" apart). Each point is fitted with 1)A single pressure microphone, whose output is the instantaneous value of all compressions at that point without direction. 2)The second is a velocity microphone pointed straight ahead, basically a direction-to-amplitude encoder. Any wave directly in front would record at "1", anything at 90degrees to the ear as "0". 3)The third microphone is the same as #2, except directionally orientated L or R, depending on which side of the "head" they are on. It would pick up the sounds originating from the side of the head. As for reproduction: It gets a little complicated because it uses 6 channels, 6 sets of speakers, and the speakers have to be set up so that the delivery is over a 90degree quadrant from the front, and directional side speakers focused to a point= x-(headwidth/2) to minimize crosstalk.

It may be unreasonable, but it's the only way I'll listen to the Ramones today.