Bill Sommerwerck: The more intelligent you are, the less you know. I don't mean just the obvious use of "know," where one rigidly holds certain beliefs as absolute. I'm also suggesting that the properly disciplined mind replaces a lot of specific facts with a few generalities.
The inductive process (reasoning from general observations to specific principles) has obvious application to math and science and is just as applicable to such studies as economics or history. And once one understands enough basic rules, deductive reasoning allows one to predict the outcome of a particular set of conditions from a general principle. The predictive quality of scientific and engineering principles gives them their power.
Deductive reasoning can also be used to test the plausibility of observations. For example, if listener L claims that turning the channel selector on her TV to an in-between position improves the sound of her stereo, even with the TV off, listener S may object to that observation as wrong-headed because he cannot think of any mechanism that would change the sound of a stereo coincident with a change in TV channel-selector position.
Most people intuitively understand in ductive and deductive reasoning without knowing their names. They apply these principles to their work, and, as time goes on, their systematic understanding increases. There are exceptions, however, the most obvious one in hi-fi being Ivor Tiefenbrun of Linn Products.
IT (as he is known in the UK) has a reputation in some quarters for citing observations (right or wrong) and following them up with a lot of apparently ill-drawn conclusions. I want to set one of his most serious errors straight, and then show why it is not always a good idea for equipment reviewers to publicly parade their devotion to someone of doubtful authority.
The issue is this business of "single-speaker" listening and demonstration, which has become fashionable in the UK.
The premise: bringing a second pair of loudspeakers into your auditioning room upsets the sound of the pair you're listening so badly that the first speaker's ability to correctly reproduce the timbre of musical instruments is destroyed. This observation is almost surely correct. But IT and his coworkers are so devoid of any real understanding of mathematics or acoustics that they draw invalid conclusions. In fact, they miss the real significance of their observations.
There is no question that adding another speaker does alter the sound. But what conclusion are we to draw from this observation? The one drawn by IT, et al, is that no additional speakers, of any kind, are allowed. This includes any tympanic surface, including microphones, and the transmitters and receivers in telephones!
I cannot recreate IT's line of reasoning, but, given the conclusion, I can make a stab. If another speaker upsets the sound, what distinguishes the speaker from any other object? Easy; it has a vibrating surface. Ergo, all vibrating surfaces are banished from the room. (As Linn puts it, "all transducers, no matter how small".) Does this include phono pickups, the motor in a Polaroid SX-70 camera, and electric lights?)
As most readers will have already seen, this reasoning is incomplete because it ignores a broader principle: how does anything in the room, speaker or otherwise, affect the sound? The sound arriving at our ears is a combination of the direct sound from the speaker, plus sound reflected from and diffracted by other surfaces. Generally speaking, the larger an object and the closer it is to the speakers, the more it affects what we hear. One cannot single out the effects of a speaker's driver diaphragm, while conveniently ignoring everything else in the room!
Imagine a totally absorptive room, containing only a pair of speakers and the listener. Now position a wood box, about the size of a shoe box, a few feet in front of the listener, above his head and to the right. What effects will it have?
At low frequencies (those whose wavelengths are much longer than the dimensions of the box) nothing will happen; sound simply passes around the box as if it weren't there. At high frequencies (wavelengths much shorter than the box), the box both reflects and diffracts sound. (Diffraction occurs when wave motion en counters an obstacle that is comparable in dimension to its wavelength. The wave re-radiates from the edge of the obstacle as if the edge itself were a source of sound.) At intermediate frequencies, there is some combination of reflection, diffraction, and passing-around.
These sounds, of course, reach the listener after the direct sound from the speaker. The delay introduces varying phase shifts between the direct and reflected sound, increasing the amplitude where they are in-phase, and reducing it where they are anti-phase. This upsets the frequency bal ance and changes the phase relationships among the harmonics of musical sounds.
These effects are audible! The change in spectral balance introduces colorations. In fact, our hypothetical wooden box should be audible as an obstruction in the sound field. With music playing, I'm certain that any of our readers could walk blindfolded into this imaginary room and describe the approximate size and position of the box.
Of course, I've picked a deliberately exaggerated example. How about something simpler, such as putting the box in the plane of the listener's ears, and a few feet to the side? That, too, would certainly be audible as a slight change in balance and timbre, and a minor degradation of the image.
In short, it is possible to predict all the effects heard by Ivor Tiefenbrun et al, not by introducing a second speaker, but merely a second speaker box. Is there any significant difference if the box contains a driver? I think not. Some of the incident sound will reflect from the driver just as if it were the surface of the box.
The driver will also be set into motion, re-radiating some of the sound. The spectral balance and phase of the re-radiation will be altered by the mechanical characteristics of the driver. However, most drivers are small compared to the wavelengths they reproduce, so they make a poor impedance match to the air. This means that the incident sound will move the diaphragm little, and that even less energy will be returned to the air. In short, the driver has far less effect on the sound field it is immersed in than the box which holds it.
Besides: Who listens to a single speaker?! Almost everyone has stereo. If bringing a third speaker into the room degrades the sound, then why doesn't the second speaker of a stereo pair foul up the sound produced by the first speaker?
In fact, the speakers in a stereo pair do mutually interfere. Visiting your friendly neighborhood audio saloon you probably see three to five pairs of speakers arranged in a single row. If you play a pair with whose sound you're familiar, you should notice that it is suddenly afflicted with severe boxiness and noticeable colorations. Why?
It has been known for 40 years that the edge of a loudspeaker enclosure does a great job of diffracting the sound produced by the drivers. One way to minimize this diffraction is to make the front panel of the speaker as narrow as possible, which tends to keep diffraction effects out of the passband of the driver. When a speaker system is set flush with a half-dozen other systems, the front panel suddenly becomes very wide, and the diffraction effects move down into the midrange. Getting the pair as far as possible away from any other surface produces the best sound.
Although the effects are not as pronounced, exactly the same thing happens between...the two speakers of your stereo pair. The larger the speakers, the worse the interaction.
The side-effects of this interaction can be minimized by covering the speaker with acoustically absorbent material. Several firms, notably Infinity and Acoustic Research, have layered their front panels with felt or polyfoam, with the intent of reducing diffraction and re-radiation effects. It may have not occurred to them that this would also reduce coloration caused by sounds emanating from the other speaker.
I said earlier that we would examine what happens when reviewers endorse the principles of dubious authorities. I'm thinking specifically of the article in the March 1984 issue of Hi-Fi News & Record Review, which shows the listening set-ups of their reviewers. One in particular says ". . . this room is a 'single speaker' room—television sets have been banished elsewhere, and the telephone is unplugged and removed when I am listening."
In a sketch of the room in the article, you can note the "low table" in front of the listening position. Egad! The speaker's output will reflect from the table, combining with the direct sound to chop holes in the response, particularly in the midrange! How can one take this reviewer's opinions seriously when he doesn't recognize the fidelity-destroying character of his own setup? His obeisance to the principle of "single-speaker" listening, while allowing a coffee table to corrupt the sound of his system, is a classic example of straining at gnats and swallowing camels.—Bill Sommerwerck
Footnote 1: Nevertheless, if listener L somehow establishes that sonic changes do take place with a change in channel-selector position—we'll leave to another discussion how L might do so—it simply indicates that S is ignorant of the mechanism at work. The history of audio is filled with examples of audible results from impossible or non-existent mechanisms.—Larry Archibald
Footnote 1: Nevertheless, if listener L somehow establishes that sonic changes do take place with a change in channel-selector position—we'll leave to another discussion how L might do so—it simply indicates that S is ignorant of the mechanism at work. The history of audio is filled with examples of audible results from impossible or non-existent mechanisms.—Larry Archibald
