July 13, 2022
J. Gordon Holt | May 12, 2022 | First Published: Feb 1, 1968
The idea of a loudspeaker system whose frequency response could be tailored to suit room acoustics and/or personal taste is one that has always appealed to the high-fidelity perfectionist. Ideally, such a loudspeaker would allow you to raise or lower the level of any part of the audio spectrum to correct for, say, a sharp 370Hz room resonance or a mild absorptive condition that weakens, say, the 800Hz–3kHz range. Obviously, though, this kind of flexibility would require an infinite number of bandpass networks, each with its own volume control—which is an obvious impracticality.
In the past, the closest approach to this “ideal” that has been reached was in multi-way speaker systems with level-control pots on each upper-range speaker, and in bi-amplified systems using electronic crossover networks that allowed for some variation in crossover points. The KLH Twelve combines some aspects of both approaches and provides more flexibility than either one alone.
An Adjustable Crossover
The Model Twelve is a three-way, four-speaker system with a 12″ acoustic-suspension-type woofer, two midrange cone speakers (connected in parallel), and a single cone-type tweeter. Thus, there are two crossover frequencies involved—between the woofer and the midrange speakers, and between the midrange speakers and the tweeter.
In a conventional speaker system, this would call for a crossover network having four reactive elements: a low-pass (treble cut) filter for the woofer, a high-pass filter for the bottom of the midrange speaker’s range, another low-pass filter of higher frequency to set the upper limit of the midrange speaker, and another high-pass filter to feed only the very highest frequencies to the tweeter.
Assuming the loudspeakers to be uniform in response and efficiency, the smoothest overall frequency response would be obtained when the crossover filters between two speakers are tuned to the same frequency. If, however, the woofer’s low-pass filter is tuned to 500Hz and the midrange’s high-pass filter is tuned to 1kHz, the woofer will start to die out before the midrange speaker starts to take over fully, and the result will be a dip in their combined output between 500 and 1000Hz. On the other hand, if the woofer is allowed to function out to 1kHz and the midrange is allowed to extend down to 500Hz, the speakers will tend to reinforce one another through the overlapping crossover range, and there will be a rise in response from 500 to 1000Hz.
If we adjust both crossovers upward by an octave, moving the overlap range to between 1000 and 2000Hz, we can produce an equivalent response hump at a higher frequency. Or, we can produce a dip at this higher frequency by crossing the woofer at 1kHz and the tweeter at 2kHz. In this way, we can use the crossovers between two speakers to varying the overall frequency response over two different parts of the audio spectrum—above and below the “center” crossover point. And with two “center” crossovers to work with (the woofer-to-midrange one and the midrange-to-tweeter one), we can provide some measure of control over the system’s response over four different frequency ranges.
In the KLH Twelve, the crossover network has been so designed that the four controllable ranges are of equal subjective width. (Each is about 1/3 octave wide.) As usual, there is no level adjustment for the bottom part of the woofer’s range. The four contour controls cover the ranges of 300–800Hz, 800Hz–2.5kHz, 2.5–7kHz, and 7–20kHz, and instead of continuously variable pots, three-position rotary switches are used. Each switch selects flat response or 2.5dB of boost or cut—not a wide range of adjustment, but quite immediately audible to a critical listener, and quite adequate for most purposes. As a matter of fact, KLH informs us that the adjustment range was limited intentionally to prevent tin-eared users from fouling up the sound of the system by grossly misadjusting the controls.
These contour controls, by the way, could be an education in themselves for anyone wishing to train his evaluative ear, for they make it easy to learn the subjective effects of a mild response hump or dip in various different frequency ranges. As a matter of fact, we would like to urge anyone with preconceived notions about the inaudibility of small response deviations to spend an evening or two fooling with the Model Twelve’s controls and, perhaps for the first time, listening.
The Model Twelve’s crossover components and the four contour controls are contained in an attractive, walnut box of about the same size, shape and weight as a desk dictionary. It is connected to the speaker system via a four-wire cable (with color-coded spade lugs), and two cable lengths are supplied. A short cable is for use when the control center is to be located at the speaker system itself, while a 40′ cable allows the control center to be placed within easy reach of the listening location in the room.
An ingenious innovation here is the use of Velcro hook-and-pile fabric swatches to enable the contour control center to be fastened to the back of the speaker enclosure, out of sight but with all controls still accessible. Velcro fabric, by the way, is the same stuff that’s used to fasten the arm band on those barometer-like devices that doctors use for checking blood pressure. The fabric comes in two mating varieties, one with a surface of fine, fuzzy closed loops of stiff thread, the other with regular rows of tiny plastic hooks. You press one against the other, the hooks get caught in the loops, and the two surfaces adhere tenaciously. If you want to separate them you just pull, and the hooks straighten out and let go with a horrendous tearing sound. The nice thing about it is that neither fabric is damaged when they’re pulled apart, so they may be joined and separated countless times without losing their grip.
In the Model Twelve, the looped fabric strips are on the bottom of the contour control box, where they also prevent the box from scratching finished surfaces, while the hooked strips are at the back of the speaker cabinet.
We received a pair of Model Twelves for testing, but the initial tests were made on a single unit, located in a spot in the listening room that has proven suitable for most other speakers of equivalent size. All contour controls were initially set for Flat response.
Sweeping an audio oscillator through its range, we found the Model Twelve to be extremely smooth in subjective response from well beyond audibility (appreciably above 16kHz) down to about 70Hz. At 70Hz, there was a sharp drop in output, a leveling-off at around 60Hz, and then the response continued flat but at significantly reduced level down to a healthy, floor-shaking 35Hz. Fundamental output was still faintly audible at 25Hz, and when the signal finally disappeared, the speaker went silent. There was no audible trace of distortion or flutter until much greater amounts of power were fed to it at 20Hz. (At 1kHz, ½ watt of RMS power input sounded very loud; it took 5W at 20Hz to produce flutter.)
Treble dispersion was very wide (about 90°) but somewhat irregular, evidently because of interference effects between the two fairly widely spaced midrange speakers (which carry the ranges up to around 5kHz.) Moving from one side of the speaker to the other, we detected two distinct areas (about 20° on each side of axis) where the mid-highs were slightly weakened. The upper high-end range, coming from the single tweeter, was unaffected.
Overall efficiency was similar to that of most other acoustic suspension systems; we estimated it to be about 1.5%.
Predictably, the single Model 12 sounded well balanced, with no boominess nor marked thinness, but it was felt to be noticeably lacking in deep foundation. The deeper notes were there, but significantly reduced in level; usable bass extended little below about 60Hz.
Moving the speaker into a corner of the room (where bass output is at maximum) brought up the entire bottom range to some extent, but the overall sound was still felt to be a shade bottom-shy. We did find, though, that because of the speaker’s power-handling ability at low frequencies and the nature of its bass characteristic, it was easy to provide almost perfect correction via a small amount of bass boost from the main system’s tone controls, and this extended the usable low end to a deep, firm 35Hz.
Apart from the shelf in the lower-bass range, the Model Twelve’s low-frequency performance was superb—deep, tight, exceedingly well-defined, and without a trace of heaviness or boominess, even with the addition of tone control correction.
Our reaction to the system’s overall sound, though, was oddly ambiguous, even after several weeks of listening on a wide variety of program material. Our first reaction was that the speaker was a wee bit shrill, and we are still aware of a certain subtle hardness from the speaker. But we have also been forced to the conclusion that much of the hardness we hear is directly attributable to the recordings themselves.
Early LP discs (before the record manufacturers started getting self-conscious about hi-fi), 78-rpm discs, and tapes that we have made ourselves have none of this hardness. As a matter of fact, the Model Twelve has shown itself to be one of the most natural-sounding speaker systems we have heard. It neither favors nor disfavors woodwinds, brasses, strings or percussion instruments, and sonic details are there in abundance without being exaggerated in the manner of some speaker systems with tipped-up or peaky highs.
The Model Twelve does very well reproducing voice, too—either speaking or singing. Male voice has none of the boominess or heaviness that is all too common in modern “hi-fi” loudspeaker systems, and some of the original tapes we have of singers whose “live” voices we have heard frequently reproduce with startling similarity to the real thing.
And yet, paradoxically, for all of its naturalness of timbre, there is something missing from the Model Twelve. Perhaps we have just been spoiled by the extraordinary transparency of electrostatics, but the impression we get from the Model Twelve is that of a faint haze or curtain over the sound. There is just not the feeling of aliveness—of the lack of an intervening loudspeaker—that we have heard from some systems that are often more colored than is the Model Twelve.
The Janszen Z-600, for example, does not sound as smooth in the middle-high-end range, it does not quite match the Twelve’s superb bass detail, and it is not quite as convincing a reproducer of the human voice and the larger brass and woodwind instruments. But its ability to convey an illusion of actuality—a feeling that the instruments are right there behind a clear “window”—is unmistakably better than that of the KLH Twelve.
We can’t explain why this might be the case. Perhaps it is a result of the very thing that makes the Model Twelve so flexible: the relatively complex crossover network, which may tend to decouple the speakers (electrically) from the amplifier, thus sacrificing some degree of feedback control over their cone movement. And then again, perhaps not. We can only say that, for whatever the reason, this was the way the Model Twelve sounded to us.
How is the Twelve as a stereo speaker? In view of the wide dispersion and the uniform response throughout the audio range, is it not surprising that a pair of Twelves should yield an immensely broad and cohesive stereo field, with excellent localization and a superb sense of depth and perspective. Stereo pairing also helps to strengthen the deep bass response, too, although not quite enough to obviate the need for at least a small amount of tone control correction.
In terms of the competition, the Model Twelve is an extraordinarily good loudspeaker, and is better in many respects than the great bulk of systems costing well over $400. It is easy to live with, and provides a degree of flexibility that is not to be found in any other available speaker system. If you’re shopping for speakers in or about this price range, we’d put this at the top of the list of candidates for your consideration. You may find you are less aware of the slightly hard, yet veiled quality than we were, and frankly, you’ll hunt for a long time before you find another one that sounds as natural.
KLH Model Twelve loudspeaker Manufacturer’s Comment
Any “perfectionist” loudspeaker is a difficult thing to evaluate, and we think that The Stereophile has gone at it very well on balance. But a few very important factors are worth talking about.
There is only one subjective disagreement we have: on the elusive subject of clarity” or “transparency.” We certainly agree that, in a direct comparison between the Model Twelve and the Model Nine electrostatic, the former sounds a bit “veiled.” We think it sounds less so, however, than any other moving-coil loudspeaker made, with the possible exception of our own Model Five. We would happily pit the Model Twelve against any other speaker in a three-cornered comparison (direct or indirect) with the Model Nine.
The Model Nine, of course, has severe limitations as a practical device. It takes more driving power than any other loudspeaker we know of, and it is not really suited for use at very high levels in a big room. The whole idea of the Model Twelve is to approach as closely as possible the Model Nine’s clarity and accuracy, but with far greater efficiency and power capacity. In return for a small loss of transparency, the Twelve allows you the soul-satisfying opportunity of listening at really loud levels once in a while.
The second factor is a more objective one, and one that we think needs real discussion in the future. It has to do with The Stereophile‘s impression of “hardness” in the sound of the Twelve. The issue is very simple: The amount of high-frequency response needed to do really full justice to the best sound sources these days will produce a less-than-fully comfortable quality with the great majority of lesser-quality recordings.
Most of the discomfort will probably seem to center in the “low” high-frequency region, where most of the “souping-up” is done on modern recordings. It is possible to reduce the annoyance value of this by turning down the 2.5–7kHz switch on the Twelve’s contour control, but this won’t entirely eliminate it. It is, simply, a function of the speaker having been designed for the most natural sound and optimal dispersion from the best program material.
As a matter of interest, we long ago developed a very small tweeter which was virtually ideal in theoretical performance, and magnificent in “live-versus-recorded” comparisons, but was absolutely unlistenable with any commercial record or tape. One of the challenges for us in designing the Model Twelve was to see just how much of the ideal proportion of high-frequency energy we could design into the Twelve without impairing its ability to reproduce commercial recordings satisfactorily.
With regard to the Model Twelve’s bass performance, it should be noted that the speaker was designed primarily as a stereo reproducer, in which application both speakers will reinforce one another’s bass output. If the system had been designed for “full” bass from a single speaker, a pair of them would have produced the overly heavy, tubby bass that The Stereophile has criticized in some other stereo pairs of speakers. In most listening rooms, and again assuming reasonably good program material, it should be possible to vary the bass of a pair of Twelves from slightly heavy to slightly sparse, according to taste merely by judicious placement.
We think it is important, particularly if we and other manufacturers are going to attempt genuinely “perfectionist” products once in a while, to clarify in a publication like The Stereophile the problems of attempting to duplicate live sound from commercial recordings In a home environment. “Pleasant” sound is easy to obtain, and is all too often the sole design objective of manufacturers that pay lip service to “ideal high fidelity.” But if the public is ever to get anything better than mediocre-fi, it is necessary for at least a few manufacturers to aim for ideal performance, even at the risk of making poor recordings sound as poor as they really are.
Our observations about the sound of the Model Twelve were based on a consensus of several listeners, who heard the speakers on a number of different occasions, in varying room locations and on a variety of program sources including some “live,” un-souped-up tapes. We plan, though, to continue our tests for some weeks in the future, and if prolonged listening prompts us to revise our opinions of the Twelve (in stereo) we’ll have a follow-up report in the next issue.
Afterthoughts on the KLH Twelve, from July 1968 (Vol.2 No.6)
Those of you who read our review of the KLH Model Twelve speaker system will recall that, although we were enthusiastic about certain aspects of its sound, we had reservations about its deep-bass response and what struck us as a certain hardness throughout the middle and upper ranges. Well, since we wrote that report we have done considerably more listening to the Twelves under a variety of different conditions, and must at this point conclude that we were wrong and KLH was right.
The “hardness” that we attributed to the Twelve has turned out to be almost entirely a function of the program material, aided to a small extent by a similar tendency in most preamp/control units, and abetted by our own preference in musical sound.
To eliminate the influence of the preamp unit, we did most of our later listening via direct connection between tape machines and the power amps. Program material consisted of commercially recorded 4-track tapes and a number of original 2-track masters, including some we had made ourselves using professional capacitor mikes. With but a few exceptions, the commercial tapes were still judged to be quite hard and brittle, particularly in massed violin tone. Tapes we had made ourselves were less so, but still seemed to yield more roughness from the strings than we are accustomed to hearing from other top-quality systems or from our favorite concert-hall seats.
Subsequently, some of us were afforded the opportunity to hear several different orchestras from other, less preferred (to us) seats, and we observed much the same kind of roughness from the live violin sounds. The conclusion is obvious.
We are less certain about what brought about the change in the sound of the Model 12’s low-bass performance. It did not, and still does not, yield adequate deep bottom in rooms and locations that have been ideal for other similar-type loudspeakers, but we did find some other spots, and some other rooms, in which the extreme bottom of the Model Twelves was fuller, deeper and tighter than anything we have ever gotten from a pair of Janszen Z-600 systems. Indeed, we were introduced to some awesomely subterranean bass passages that we never knew were on the recordings. And this was achieved without any significant increase in mid-bass output.
We found that the Twelves wear exceedingly well, producing neither fatigue nor annoyance with prolonged listening. They still seem to us to lack much of the liquid transparency and detail of good electrostatic systems, and there is something about the velvet smoothness (from most recordings) of the electrostatic top that we find hard to resist.
Under the circumstances, we are obliged to revise our judgment of this speaker and to say that, under ideal conditions, we feel it to be one of the best stereo reproducers available, regardless of cost. This is not necessarily to say that you will prefer it to, say, the KLH Nine or the Altec A-7 under your own conditions of intended use, but it has reproducing capabilities that put it in the same general class as these and similar topnotch systems.—J. Gordon Holt
KLH Model Twelve loudspeaker Specifications
Description: Three-way dynamic system with frequency response contour control. Speakers: 12″ acoustic-suspension woofer, two 5″ cone midrange units, 1.25″ cone-type tweeter. Impedance: 8 ohms. Power capacity: 100W program.
Dimensions: 29″ H × 22¼” W × 15″ D.
Price: $275 each, $550/pair (1968); no longer available (2022).
Manufacturer: KLH Research & Development Corp., 30 Cross Street, Cambridge, MA (1968); KLH Audio, 984 Logan St., Noblesville, IN 46060 (2022). Tel: (833) 554-8326. Web: klhaudio.com.