|Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting
Departments | Features | ADs | Equipment | Music/Recordings | History
By BOB BERKOVITZ [Acoustic Research, Cambridge, Mass.]
Of many puzzling questions encountered by the music listener learning about sound reproduction, one must gnaw persistently. If the design of high-fidelity loudspeaker systems is aimed at the single clear objective of accuracy, as is the case for amplifiers, cartridges, and tuners, why should there be so many different kinds of speaker systems available? Surely, the listener thinks, engineers, men of science, those who speak the language of mathematics, will agree upon meter readings or the location of marks upon a paper chart. Then, one method of design, one kind of speaker system will show itself to be the best in such measurements. Yet speaker systems which sound very different from each other are offered for sale, each on the basis that it alone stands on the frontier of acoustic knowledge. Each is advertised and the listener is urged to move the frontier into his living room. And the listener, having attempted to do so, is not unlikely to discover that the frontier, in the meantime, has moved on.
Of course, scientists do not really disagree about the ultimate objective of speaker design--it is a device of such high impedance as to consume practically no energy, in order that it be most efficient; it is so small and thin as to verge on invisibility, making it ideal for installation in any room; it is intelligent, and therefore able to analyze any signal fed to it, deducing and radiating in the spatial pattern correct for each instrument in the ensemble, while cancelling out the acoustic properties of the room in which it is being played. Its frequency response is assuredly fiat when heard from any direction, indoors or outside.
And it can be stamped out of sheet plastic, like an electroacoustic cookie, or, to satisfy the most glassy-eyed dream of a certified accountant, sliced from lumps of trash now being discarded as waste by another member of the corporate conglomerate. If that is a correct summary of the device which will, by its appearance, signal the end of the road in speaker-system development, we certainly have a long way to go.
On the other hand, it is fair to believe that the appearance of a speaker system like that described above will not end the aspirations of those who make and sell competitive systems. They will think of a feature that is missing, a function that is still unfulfilled. Just as many designs have been hailed in the past as having achieved the ultimate in speaker system design, each has been supplemented by another approach, offered by its maker to fill lacunae unobserved or unattended by his predecessor. The past editions of mail-order catalogs of electronic equipment are filled with such afterthoughts about someone else's afterthoughts.
Didn't they know that they were barking up the wrong tree, that their inventions, measured according to the most elementary acoustic criteria, could be shown to be inaccurate? The answer, at least in part, is that accuracy has not always seemed equally important to all speaker system designers, and even many listeners have, at least for a while, been persuaded that satisfaction of their personal taste in sound is what they are shopping for at the audio salon. It is an interesting conception, offering the gratification that comes only with the knowledge that the listener is going to establish his status and discover his identity in one transaction. The idea is attractive, until it is set in less appealing language: "What kind of distortion do you like best?" It is like asking a prospective camera buyer not to purchase the sharpest lens with the best correction, attested by data, but to buy instead an instrument that gives a little barrel distortion to the scene, for "presence," and a subtle red tint to color shots to give "warmth" to outdoor portraits and landscapes. Just as the optical designer producing such a camera would claim to have risen above conventional testing methods, some speaker system designers have claimed to have discovered higher truths. Such claims are often associated with little-understood but extraordinary discoveries about hearing or music. To know why such speaker systems fade from sight, while older amplifiers and tuners still go on, we will have to consider some actual cases.
There is an entire class of speaker systems, all but a few in the blessed past, which are based not upon science but upon metaphor.
A good example is the speaker system with f-holes, like those on a violin or cello. The literature describing this product, seen a number of years ago by one of the authors, showed an enclosure of a size large enough to make a bass-reflex design of good performance. However, instead of the classical rectangular port, or the somewhat more modernistic duct, this one had f-holes cut in the front panel, and carried an explanation which was both clear and thought-provoking. The finest instruments used to play music derived their great sound quality from the vibration of their enclosures and the perfection of design of such details as the f-holes of the string instruments. So, the system in question was made of wood which would vibrate just right. It was natural for a reader to joke, "But how does a tuba solo sound through it?" Actually, the system may have sounded rather good, despite the appeal to analogy by its advertiser; we have no way of knowing. But we do know that if the enclosure did vibrate like a violin, with its own set of resonant modes, these would be a result of the dimensions and materials used to make it, and would bear no necessary relation to the music being reproduced by the system. Not only tuba music, but violin music would have added to it the "music" of the speaker system.
... And Artistic Enclosures
An artist, a genuinely wise man and a highly gifted one, expressed to the author a fascinating conviction about the home-made speaker system owned by the artist. It sounded rather good, in fact, having been made by mounting a JBL 12-inch loudspeaker on a flat baffle, and placing the flat baffle across a corner of the room. The circular cut-out for the JBL speaker was covered with gray silk.
The flat baffle itself, which rose from the floor to about 6 feet in height, was white. It was made of marble, as thick as a man's hand, cut by a maker of monuments. The artist smiled as he told visitors that the brilliance of "the highs", caused by the use of the white marble, surpassed all other speaker systems.
Lest readers laugh at such a judgment as naive, it must be pointed out that the possibility that the marble-facade speaker system is exceptionally accurate is in no way compromised by the owner's possible error as to the reason for its excellence.
We are concerned with the disparity between subjective judgments and probable realities. On the other hand, just looking at a thick, polished white marble slab might make some listeners hear singing "highs." Moreover, there is a precedent, of sorts.
In a unit described in Wood's Acoustics as "the H.M.V. loudspeaker," the voice coil was attached to a flat sheet of aluminum alloy about thirty inches in diameter, tightly stretched and bolted to metal rings at its edge.
The 'Cats Door' System
Another area of unceasing activity has been that in which breakthroughs of various sorts are uncovered. The "perfect baffle," last seen by the author about fifteen years ago, was such a breakthrough. To all intents and purposes, it was an ordinary speaker system, at least from the front. Cone speaker. Grill cloth. Rectangular box. Then you turned it around and saw the little door, and your eyes popped. Every time the system hit a really loud bass note, the door opened for an instant on a thin leather hinge, and closed again a moment later. It was sold under the name, "perfect baffle," and the little swinging door was said to relieve pressure built up in an ordinary enclosure which limited cone motion. Its manufacturer pointed out that the "perfect baffle" was based on an "ultimate scientific principle." Possibly this was to forestall any attempts at post-ultimate developments; if so, the ploy was ineffective.
The design of speaker systems is, or can be a rational process, and engineers do not really differ much on the best current techniques for the achievement of each design goal. Nor is there much disagreement regarding techniques of measurement. The differences of opinion among speaker-system designers probably center about the criteria used to establish which design goals are most important, and which are less relevant to the application for which the system is intended. To a certain extent, these differences are understandable.
A designer who has spent his life, or the better part of it, designing speaker systems to be used for sound reinforcement in large halls or theaters will have naturally come to place a higher degree of importance on efficiency. Since large regions in an assembly hall or theater are unoccupied by listeners, the theater system designer will tend to design directional systems, rather than flood all of space with sound where there are no listeners to hear it. The designer primarily concerned with and experienced in the engineering of systems for home music reproduction, on the other hand, must recognize that the location of the speaker systems by the home listener can be controlled only in a very general way, and that listening locations in the room will vary greatly in suitability, unless the speaker system is omnidirectional. Efficiency will be less of a problem to the designer for home use, since concert levels are easily matched in typical living rooms. Moreover, bandwidth, small size, and freedom from peaks in response will be preferred by most listeners; unfortunately, theater-type speakers must often sacrifice excellence in these respects in order to achieve maximum efficiency.
It is remarkable that such divergences do not lead to gross variations in speaker systems, since the different viewpoints alluded to above are often expressed by their proponents with great vigor. However, the important fact is that engineers, as has been said, do not disagree fundamentally, and few, if any, would say that rough frequency response is good, or that efficiency is a bad or worthless design goal. As a result, loudspeakers that are well-engineered sound good, regardless of the design philosophy of the firm or person responsible for them. Engineers can meet and discuss the merits of their designs, be applauded or congratulated by their colleagues, and exchange ideas about speaker evaluation with little of the recrimination displayed by amateurs toward each other when arguing about their favorite loudspeakers.
When a speaker system is designed in a fit of passion, or in a state of technical or intellectual isolation, different factors emerge, however. The absence of proper test facilities can give supreme confidence to a designer who never needs to subject his brainchild to rigorous examination. As a result, concepts which have become almost stock jokes among engineers reappear periodically on the market as novel "breakthroughs." The "Sweet Sixteen Array" The array of many small speakers is a good example. In a rather thorough paper published years ago in Audio, James Novak and Philip Williams of Jensen established rather conclusively that the benefits typically claimed by designers of "build it-yourself" arrays, such as that called one "Sweet Sixteen," are non-existent, or so meager that they are easily surpassed by a conventional speaker system both in economy and performance. The gist of the analysis by the two Jensen engineers was that the "Sweet Sixteen"-which used sixteen inexpensive Jensen radio replacement speakers-sounded like sixteen inexpensive speakers. The directionality of such an array, the prominent peaks in response, and the cumbersome physical object which results, all worked against its acceptability as a music reproducer, yet a surprising number of home hobbyists sent away for the loudspeakers prescribed by the designer. The main idea appears to have been that if the excursion requirements of the loudspeaker are reduced to one-sixteenth of what they would normally be, very low frequencies can be reproduced without distortion.
Although the statement has an element of truth (as is often the case with such designs ), achieving the excursion needed for good bass response has not been a real problem for a long time. The design appears to have been intended to solve a problem which simply did not exist; Novak and Williams were able to show that much better bass response with less distortion was obtainable from a single bookshelf system! At a very well-attended meeting of audio technicians and engineers in Chicago, some years ago, the author produced a "Sweet Sixteen" he had built himself, and compared it directly to a Jensen TF-3, with results which had no ambiguity at all. The array was demolished.
The solution of problems which do not really exist, at least in the sense that at least a few systems are already available which offer a clean, definite solution to the problem, seems to be a recurrent theme in "eccentric" speaker systems.
Typically, a statement is made by the designer to the effect that others have sought for years to solve such and such a problem. This may be true, but the author overlooks the fact that a solution is already at hand, and goes on to propose a considerably more complex, often questionable solution.
It is not a matter of engineers vs. amateurs, or of rational beings vs. irrational ones. It is the earnest unrestrained quest for a "feature," for a "breakthrough" that the independent inventor emerges from his study with an eccentric speaker system design. More often than not, the result is interesting. On occasion, it is even a challenge to analyze the acoustic behavior of the invention.
But the basic separation between speaker systems is not into horns, sealed enclosures, and bass reflex; nor is it a matter of large systems warring against small ones.
At least not in scientific terms. It is the quest for accuracy, which most designers try to pursue, but which is difficult enough to discourage and sidetrack others.
(adapted from Audio magazine, Apr. 1970)
= = = =
Prev. | Next