More Ear Tricks

So far, we've seen that the human ear can figure out where a sound is coming from, that it has a built-in automatic level control with multiple time constants, and that at the basilar membrane it converts the sound wave into 30,000 or so neural signals, each representing a single frequency (this is roughly akin to doing a so-called Fourier Transform, a rather daunting computational task usually handled by Very Expensive Computers -- that's VEC to you propeller-heads). But that's not all, as the Steak Knife salesman tells us. The ear also does some remarkable things with time (and direction) that are far beyond the capacity of any microphone.

The primary time-based trick that the ear does is called the Precedence (of Haas) Effect. In order to keep us from being hopelessly confused by early reflections of sounds (the aural equivalent of a fun-house mirror-room), the ear integrates such early reflections (for up to 50 milliseconds after the original sound arrives) with the original sound, so that they are not heard as reflections or echoes but as part of the timbre of the original sound. The microphone, on the other hand, simply sums the original sound and all of its reflections over time, yielding a kind of interference pattern called comb-filtering that imposes yet another timbre on the sound, a timbre generated by the room. So when we hear a sound in a room, we generally don't consciously perceive the interference patterns caused by room reflections but only a richer and more satisfying version of the sound itself (due to information carried by all the reflections). This is why we don't like to listen to or play music outside nearly as much as indoors. Those reflections and the richness aren't there. On the other hand, for recording, the lack of reflections generally helps with clarity, which is why recording studios often have extremely absorptive acoustic treatments.

Another facet of this time-based complexity is how our ears treat high and low frequencies differently. As we mentioned above, high frequency interference patterns at each ear help us to determine what direction a sound is coming from. These high frequencies happen too quickly (the nervous system doesn't operate fast enough) for information about relative phase at each ear to be sent to the brain. So, we localize where a sound is in space by use of the high frequencies present at each ear. Low frequencies (long wavelengths) take a comparatively long time to happen. Their phase state is sent to the brain where it is compared with data about the same frequency from the other ear, and from this compared phase information our brain learns about the room or environment. We particularly enjoy low frequency reflections coming from the side walls, and we use low frequencies to learn about the space in which we are listening.

Another oddity to mention is the way the auditory nerve itself functions. The auditory nerve is the bundle of nerves carrying the 30,000 or so nerve endings from the basilar membrane to the brain. But unlike a bundle of audio cables, the individual nerves aren't insulated from each other, which means that they interact as nerve impulses travel from the ear to the brain, processing the neural information as it travels to the brain. The result of this is that the information received at the brain is a lot different from the information sent from the basilar membrane. There is a substantial refinement of pitch and timbre that occurs as a function of this process (think of a sound reinforcement system where much of the mixing and equalization occurs in the snake en route from the microphones on stage to the mixing console!). Further, additional nerves in the auditory nerve group send information back to the basilar membrane, and this information causes very low-level pitches to be generated on the basilar membrane to help us extract pitch information out of complex, noisy sounds. (Yup, you guessed it. The ear is also a low-amplitude bank of sine-wave oscillators! Not quite your basic microphone!!) This was the stuff that got suppressed when I was busy tripping for the sake of medical science.

You might also be interested to know that this hearing process all takes some time to occur, so that there is a significant delay (about 6 milliseconds -- as part of my aforementioned medical misadventure, I also got to see the delay time between a sonic impulse at my ear drum and the resulting brain response which occurred about 7 milliseconds later on a dual-trace oscilloscope -- I've always known I was a little slow!) between when a sound wave reaches our outer ears and when we actually consciously perceive it! This means it should be impossible for musicians to play together. In fact, much of what we do to play together is by visual cue (which is one of the reasons overdubbing is so hard), and there also is some sort of masking process that actually blocks out what is really happening so we can concentrate on what we'd like to think is happening, keeping us from noticing the time difference between what we see and what we hear, and (even more confusingly and paradoxically) between what comes into our ears (er, what we hear) and what we hear! Yikes!

One final attribute of the ear to consider is the nature of its frequency response. That response isn't the same for all frequencies. We tend to hear low frequencies less well than mid-range frequencies. And we have troubles with extremely high frequencies as well (incidentally, men seem to hear extreme highs less well than women). But, more interestingly, that response behavior changes with overall loudness, which is to say that the bass and treble controls in our mind vary automatically and dramatically with loudness. As levels get softer, low and extreme high frequencies get softer at about twice the rate as mid-range frequencies. This is what the "loudness" button on your stereo receiver is supposed to compensate for. When you are listening at low levels, you push the button and it boosts bass (and sometimes extreme treble) to compensate for this effect (the person mixing the recording probably didn't mix it at very low levels and almost certainly didn't mix it to be listened to at low levels). The nature of these changes is expressed by a set of so-called Equal Loudness Contours (sometimes known as the "Fletcher-Munson Curves"). As Casey Stengal used to say about the NY Mets, "Amazin'!"

COMMENTS

Post a Comment