A Spats of Speaker Systems, Large and pro, Good and Bad (Vol.2, No.2: 1979)

A Spats of Speaker Systems, Large and pro, Good and Bad: Axiom TLT-1 Beveridge System 3 B&W DM7 DCM 'Time Window' (further improved) Fried Model C (improved) Magneplanar Model MG-1; Onkyo Model F-5000; Perspective MK2 QLN I; Sound Lab R-1; Swallow CM70; Vandersteen Model IIA .

This is a motley group indeed, including some fairly impressive new developments and some major disappointments, names that you're unlikely to have encountered and names that have some cachet, with little or no relationship discernible between price and sonic accuracy.

By now we ought to be used to the idea that most speakers are pretty bad, regardless of price, and that the next one we test is statistically likely to be bad, too-but we just can't think that way. We still start every new evaluation with naively high hopes, only to end up being astonished once again that hardly any speaker designer appears to have a reference standard for reasonably accurate sound. We firmly believe that a side-by-side listening comparison with one of the few speaker systems capable of a fair imitation of music, such as for example the 25-year old (!) Quad electrostatic, would instantly send most of the new designs we examine back to the drawing board. Obviously, no such comparison takes place, the designer relying instead on some highly deficient or wishful private criterion.

Another source of constant astonishment in our laboratory is the obvious lack of a comprehensive test program in the vast majority of loudspeaker factories. The dozen or so simple, straightforward electronic and electroacoustic tests we listed in our cumulative reference issue (Vol. 1, No. 6, pp. 14-15), to which we routinely subject every speaker we review, reveal in nearly every instance vulgar design errors that typically would have cost no money at all to avoid or correct. What are these manufacturers measuring? Interestingly enough, the ''raw'' drivers most audiophile-type speaker houses buy from the limited number of trade sources (such as Audax, Becker, KEF, Peerless, Phi lips, etc.) are quite good; it's the way the complete systems are conceived and executed that shows a consistent avoidance of homework. Filter theory, in particular, seems to be foreign territory to all but a few speaker designers, and their crossover networks are lamentably primitive as a result. We hate to keep harping on this, but it happens to be perfectly true that the level of engineering competence currently accepted as normal in the hi-fi speaker business would get anyone fired in two weeks at a good aerospace firm. That's because aerospace vehicles with bad electronics kill people; bad speakers only kill the joy of music and only for the finicky listener.

Beware of speaker fuses.

The one new thing we want to add this time to our previously established principles in the area of speaker de sign is a warning about speaker fuses. Too few designers realize that only a heavier fuse (say, from 5 amperes up) can be considered as a straight-wire segment of the speaker leads.

Lighter fuses, especially very small ones of the order of 0.75, 1 or 1.5 amperes fast blow, constitute highly nonlinear, cur rent-variant circuit elements that can audibly modulate the signal. This is quite a different problem from the relatively high resistance of such fuses, which has long been of concern to audio purists because of its effect on damping factor and power dissipation. Here we're talking about a phenomenon that actually introduces veiling, edginess and other amplifier-type sonic crud. (See, for example, the Axiom and Magneplanar reviews below.) The question to ask, then, is whether or not you absolutely need all that protection for your oppressively fused speaker. Are you driving it with a huge amplifier? Do you ever drop the stylus with the volume control wide open? Do you play a lot of electronic synthesizer music with loud, continuous-wave signals in it? Is the speaker really so fragile? If the answer is no on all counts, you'd be better off shorting out the fuse or at least replacing it with a somewhat heavier slow blow or a much heavier fast blow. Also, check the speaker fuses that come incorporated in some amplifiers as part of the package. The signal obviously doesn't care at what point in its path the fuse sockets are located, be it at the amplifier or speaker end of the wire; what matters is the type of fuse and its value.

The only permanent solution will be a more sophisticated protective device that doesn't depend on temperature and melting to create an instant open circuit and at the same time looks like a straight wire to the amplifier.

Axiom TLT-1

Axiom Engineering Laboratories, 9601 Owensmouth Avenue, Chatsworth, CA 91311. TLT-1 Loudspeaker, $500 the pair. Five year warranty. Tested #100279 and #100280, followed by #100433 and #100441, on loan from manufacturer.

First the good news. This 38-inch tall floor-standing column speaker is very well constructed, considering its price; the exceptionally heavy particle-board structure incorporates a diagonal dividing wall that doesn't quite reach the bottom, thereby creating a 6-foot folded back-wave path for the 8-inch woofer/midrange, ending in a vent to the rear of the column. The 1-inch soft-dome tweeter is, mirabile dictu, in phase with the cone driver (though they are far from pulse-aligned, either geometrically or electrically). and the resulting sound has surprisingly good definition and focus, probably accounting for the speaker's enthusiastic following in some circles. The bass is quite solid, though the tuning of the enclosure evidences some anomalies; the - 3 dB point of the system is somewhere around 48 Hz.

Now for the bad news. We find the upper midrange and lower treble much too hard and edgy, often to the point of nastiness. We couldn't live with this kind of persistent ir ritation for any length of time. Initially we thought there may be something wrong with one or both of the tweeters; the manufacturer then sent us a second pair of speakers, this time with Ferrofluid injected into the domes. The revised TLT-1's showed peaks, troughs and mild ringing located at totally different frequencies than similar flaws in the first pair, but these were all within the range of the 8-inch driver; the basic edginess of the system above that range remained the same.

Since the Peerless KO-10 dome tweeter used by Axiom is known to us to be quite capable of producing smooth and musical highs (although Axiom puts a pinhole through the dome for wrongheaded reasons not worth discussing here), we concluded that the main cause of the ugly cutting edge in the region where the tweeter takes over is a miscalculated crossover. The tweeter cuts in at 2 kHz, which is much too low for a tiny 1-inch dome, and its drive level is set much louder than that of the 8-inch unit. (No tweeter control is provided.) Thus the tweeter tends to protest when the strings dig in or the soprano hits a high note-in other words, when there's a lot of energy near the bottom of its overextended operating range.3/4 ampere in our original pair of speakers, 1 ampere in the more recent pair. (See our comments on such fuses above.) When we substituted a much heavier fuse, the hardness and edginess were somewhat reduced but not eliminated, alas. Too bad, since the TLT-1 is a decently priced, audiophile-oriented product such as the mass-marketers would never dream of making, but that's still not sufficient reason for us to recommend it at its present stage of development.

Beveridge System 3

Harold Beveridge, Inc., 505 East Montecito Street, PO Box 40256, Santa Barbara, CA 93103.

System 3: unamplified electrostatic line-source speaker with integral woofers, $3500 the pair. Unlimited warranty on parts; five-year warranty on labor. Tested early production samples, on loan from manufacturer.

Here it is-the Beveridge electrostatic system without built-in electronics that owners of expensive power amplifiers and electronic crossovers have been waiting for. In some ways it's the best loudspeaker we've ever tested, in some ways it's woefully deficient, but luckily the deficiencies are in large part curable and, we're told, in the process of being cured.

The heart of System 3 is still the 6-foot Beveridge electrostatic ''line source'' with its ingenious cluster of acoustical waveguides. The electrostatic elements as well as the acoustical lens have been somewhat reworked since Sys tem 2SW; conventional transformer coupling to your choice of amplifier has replaced the integral direct-drive tube amplifier; the bottom of the line source's working range has been raised to approximately 200 Hz, below which two 10-inch woofers per side take over. The latter are now incorporated in the same very handsome 6'/2-foot high tambour cabinet as the line source; the cylindrical cabinet is split lengthwise, the back part forming the sealed enclosure for the woofers, one on top radiating upward, the other on the bottom radiating downward through a slot. A built-in 200-Hz passive cross over with 18-dB-per-octave slopes can be bypassed for biamped operation. Thus the new Beveridge (which doesn't replace its predecessor but provides an alternative to it) is a monolithic full-range electrostatic/dynamic hybrid system.

The net result is a substantial improvement in dynamic headroom over the 2SW, essentially the same uniquely persuasive radiation characteristics and spatial perspective, a few minor new problems in the line source, and completely unacceptable bass-at least in our very early production samples.

Let's dispose of the bad news first: the entire cylindrical cabinet is alive, its surface acting as an uncontrolled passive radiator that partially swamps the woofer output and may also (we aren't equally sure about this) interfere with the line-source output. Thumping the cabinet makes it go ''wa oom?' with a questioning inflection. Our measuring micro phone showed a tremendous amount of spurious energy right off the wall of the enclosure; at 152 Hz there was a very loud and buzzy high-Q resonance; furthermore, the 10-inch drivers themselves exhibited various anomalies that overlapped the cabinet resonances, making valid measurements rather difficult. For example, the electrically measured system resonance appeared to be at 33 Hz with a Q of 0.7, but the acoustically determined dynamic Q (in response to a step function) appeared to be much, much higher, in the region of 1.5 to 2.0 at low level and somewhat beyond 2 at high level.

The -3 dB point may have been at 33 Hz as predicted by theory, but too many other things were going on in the bass, including a rather sharp breakpoint or ''elbow'" at 26 Hz a few dB further down and various peaks and suck-outs all over the place. Rather messy, all in all, and the sound confirmed the measurements; we simply couldn't duplicate or even come close to the solidity and definition of our Janis W-1 reference woofers no matter where or how we set up the Beveridge.

Harold Beveridge has informed us that all of these bass problems are being taken care of. The cabinet has received more effective bracing and deadening (one of our spies thumped the latest version just before we sent this review to the typesetter and reported that the ''wa-oom?'' was gone); the drivers are also being revised; our major objections should be past history by the time we do our follow-up report for the next issue. Or so we hope.

Moving on to more pleasant observations, we still consider the radiation geometry of the Beveridge line source-the way it launches the wave front into the room-to be the most nearly correct of any speaker system developed so far and, combined with the inherent time-domain advantages of an electrostatic transducer, capable of the most successful imitation of a live sound field. As soon as we heard the first few bars of a choral recording through the System 3, we knew that we could never go back to the Koss Model One/A, not even as the midrange of our tri-amped reference system. There was just no comparison; the spatial characteristics were so much more distinct, three-dimension al and lifelike. Many music lovers will forgive the System 3 (or any other Beveridge model, for that matter) all its sins because of this one overwhelmingly persuasive quality. In comparison with the remembered sound of System 2SW, we gloried in the new freedom from strain at really loud levels and the absence of any built-in amplifier peculiarities (we used the Rappaport AMP-1), but we also noted some un mistakable midrange colorations well above the woofer pass band as well as a very slight softening of focus. As we said, the liveness of the cabinet may have spilled over into the working range of the line source, which is after all embedded in the same structure; in addition, we measured some new anomalies. Among the latter were bad ringing on tone bursts at 942 Hz and 1.28 kHz, and square-pulse response that was not quite as astonishingly flawless as in the 2SW, though still better than that of dynamic speakers. We must reserve final Judgment on this pending the promised cleanup of all mechanical resonances.

The frequency response of the line source can also be expected to be different in more recent production samples than what we measured. Beveridge supplies a passive equalizer network as part of System 3, to be inserted between the preamp and power amp or into the tape monitor loop. The network synthesizes a broad peak around 10 kHz to compensate for a drop in the unequalized response and also provides a few dB of variable boost at lower frequencies. It became obvious, however, that part of the reason for the declining frequency response of the electrostatic unit was the inter action of the latter's capacitance with the speaker wire's series inductance (see the full story in the wire and cable article elsewhere in this issue). This effect was further aggravated by the 1-ohm resistor Beveridge had wired in series with the transformer primary in early samples, with a view to limiting the amplifier-draining impedance drop at the highest frequencies. The resistor has been removed in recent production (the resulting impedance at 20 kHz is now only 1 ohm!);

the equalizer network characteristics were under review when we last inquired; the whole frequency response situation is still up in the air at this writing. One thing is certain:

there wasn't much to be heard or even measured above 12 kHz in our early sample, and that surely needs to be fixed before the speaker can be considered for reference applications.

You're beginning to get the picture: the early samples should never have been released. That's par for the course, however, among small manufacturers of exotic audio components. The first ten or twenty users are always the guinea pigs. As consumerists, we see red every time that happens, but the Beveridge is such a special product that we re willing to be a little more patient. We know we'll never be able to live for any prolonged period without that unique wave launch characteristic and the resulting sonic illumination of the room. In all other speakers the sound appears to dribble out of little holes, at least by comparison. So we're keeping the faith until we receive our revised samples; meanwhile, if you happen to be one of the early birds who just couldn't wait and bought the first pair available, we suggest you get in touch with the factory and discuss the possibility of a retrofit. In our opinion, they owe it to you.

B&W DM7

B&W Loudspeakers Ltd., West Sussex, England; distributed in North America by Anglo American Audio Company Inc., 1080 Bellamy Road North, Scarborough, Ont., Canada MIH 1H2, and PO Box 653, Buffalo, NY 14240. DM7 speaker system, $1190 the pair ($1350 the pair in special finishes). Tested #08813 and #08814, on loan from distributor.

Striking in appearance and impressively low in sonic colorations, this English import would be a strong contender for our ''Reference B'' selection were it not for the rather steep U.S. price, which reflects the extra middlemen involved in an export-import situation. In England it costs less, but in any country it would have to be considered a beautifully made and quite excellent-sounding speaker.

Not much larger than a very large bookshelf speaker, the DM7 stands on a columnar pedestal to reach a height of three feet, its bullet-shaped dome tweeter sitting externally on top of the cabinet, covered by a hemispherical wire basket. Not exactly conventional. The rectangular enclosure houses an 8-inch bass/midrange driver (with an unusual plastic cone that appears to be woven like wickerwork) and a passive radiator of approximately the same size. The cross over between woofer and tweeter is rather elaborate, with some obvious attention to phase relationships and delay compensation, although the tweeter is connected with re versed polarity relative to the woofer, so that absolute coherence is impossible by definition. Nevertheless, pulses are quite accurately reproduced (a la DCM and Tangent, with a little opposite-going preshoot); pulse shape retention is good to 0.14 m-sec width (only a very few speakers make it to 0.10 msec or less). Most remarkable of all, tone bursts elicit no ringing whatsoever at any frequency. That's truly rare and may account for the basically neutral tonality of the DM7.

The - 3 dB amplitude response corners of the speaker are at 50 Hz and 17.4 kHz, and everything in between is spectacularly smooth. On-axis tweeter response is down only 10 dB at 40 kHz; up to 13 kHz the tweeter radiation pattern is close to spherical. Really nice. The fourth-order Butterworth alignment of the vented enclosure is pretty much in the ball park; the Q of the system appears typical for this alignment but migrates a little with increasing drive. Allin all, we can't find anything very serious to criticize in the basic design of the DM7 except that out-of-phase tweeter connection; of course, it's still a fairly small two-way dynamic speaker, with all the inherent limitations of the breed.

The crisply defined, spacious, transparent and largely uncolored sound of the DM7 is equaled or excelled, among all the dynamic speakers we've tested so far, only by a last-minute arrival, the new Vandersteen Model IIA (but not by the Model II). The one thing that might place the DM7 lower than the IIA in our sonic ranking is just a touch of nasality, which is pitched higher and is therefore less notice able than the typical midrange ''nah-nah'' of most speakers.

The considerably lower price of the IIA clinches our preference. The DM7 is sufficiently impressive, in any event, to make us look forward to testing B&W's all-out Model 801 with a great deal of anticipation.

DCM 'Time Window' (further improved)

DCM Corporation, 670 Airport Boulevard, Ann Arbor, MI 48104. 'Time Window' floor-standing loudspeaker, $660 the pair. Five year warranty. Tested #12676 and #12677, on loan from manufacturer.

DCM has replaced the Philips woofer/midrange drivers in the Time Window with superior units of in-house design, has made some minor changes in the vented system alignment to correct the tracking of the vents with the woofers, and is now individually matching the Philips dome tweeters to each speaker. The result is greatly improved overall sound, considerably less colored by the characteristic lower-mid range thickness of earlier Time Windows and more clearly etched as well. The dynamic headroom, already very good from the start, has also benefited; with a good amplifier the improved version is capable of reasonably clean sound pres sure levels that will crack the plaster in most listening rooms.

If you're an SPL freak, this is the medium-priced speaker for you, without a doubt, unless you're willing to consider low-fidelity alternatives.

The amplitude response ''corners'' of the Time Window are now at 44 Hz and approximately 16 kHz, which isn't half bad for a not very large system of fairly high efficiency.

What's more, the speaker is quite smooth and free of any obvious ringing between those frequencies; square pulses retain their shape down to a width of 0.15 msec but show the inevitable opposite-going tweeter preshoot that goes with polarity reversal. (We've had endless friendly arguments with designer Steve Eberbach about this.) We observed no serious misbehavior by the speaker in any of our tests; the woofer voice coils do come out of the gap, of course, when you drive them extremely hard, at which point the Q goes to pieces, but there appears to be adequate control up to quite a high level. (Ultimately there's no substitute for a big, grown-up woofer, and DCM has come out with one, called the Timebase, which we plan to review in the next issue.) If it hadn't been for the last-minute arrival of the Vandersteen Model IIA, this latest version of the Time Window would have been our unequivocal ' 'Reference B'' selection, since it surpasses the Model II by about the same margin as the latter bettered the older Time Window. The IIA, however, is even more neutral, focused and generally accurate overall, with an airier and more three-dimensional rendering of space, although the new Time Window is also very good in all these respects, as well as more efficient and capable of playing considerably louder without breaking up. Our choice is the Vandersteen, but we know that some users will be happier with the DCM, especially in view of its significantly lower price. For the money, we don't know of any speaker that even approaches the Time Window in transparency, inner detail, spatial perspective, naturalness and just plain musicality. It's one of the few engineered products among a hundred amateurish cut-and-try boxes competing for the same consumer dollars.

Fried Model C (improved)

Fried Products Co., 7616 City Line Avenue, Philadelphia, PA 19151. Model C satellite monitor, $950 the pair ($400 the pair in kit form-everything but the wood). Tested samples on loan from manufacturer.

Fried has made a few minor changes in the Model C satellite, which have resulted in major improvements in both measurable and audible performance. This is a very good little speaker now, one of the better (though not the best and certainly not the best per dollar) dynamic systems around.

We still hear a slight edginess and some midrange coloration, but nothing to complain about vigorously; the overall sound is open, detailed and highly listenable in the speaker's working range, which is from 75 Hz on up. A subwoofer is, of course, mandatory.

It amuses us that one of the two irrational and ineffectual 1/2-inch holes we originally protested against has been plugged up, bringing down the dynamically observable Q (in response to a step function) from about 2.0 to 1.1 in sealed box terms. That's much better damping, but plugging up the one remaining hole makes it better yet! Maybe in the next 'improved' version ... The pulse response is even better than before; we ob served very good pulse shape retention down to 0.1 msec width. Tone bursts appear to cause no ringing to speak of, anywhere in the audio range. There still seems to be a peak in the amplitude response, at about 2.5 kHz and measurable only with some difficulty but easily apparent to the ear when the speaker is swept. This may explain the slight cutting edge on music. The culprit is the quasi-horn-loaded dome tweeter, which we never particularly liked, although Fried swears by it on account of its power-handling ability.

Incidentally, we listened again to (but didn't measure) a pair of the very similar Model B/2's, modified exactly as the Model C's. This time the smaller and cheaper B/2 didn't sound better than the C, proving that the C had indeed been meaningfully improved. The B/2 sounded fairly neutral, but from time to time an annoying upper-midrange hardness intruded, and the openness of the C was in no way equaled.

The B/2 needs, and benefits from, a subwoofer even more than the C; it opens up quite a bit when its small woofer isn't driven below 100 Hz. We still prefer, however, the Vandersteen Model IIA as well as the latest DCM Time Window to any Fried satellite speaker, with or without a subwoofer.

Magneplanar Model MG-I

Magnepan, Inc., 1645 Ninth Street, White Bear Lake, MN 55110. Magneplanar Model MG-I speaker system, $495 the pair. Tested #129680, on loan from dealer.

This junior version (one 5-foot panel per side) of the ''Maggie'' distributed-voice-coil dipole design had been advertised to us long before we tested it as the most natural sounding, musical speaker anywhere near its price, far preferable to our favored Vandersteens, DCM Time Windows, etc. We fully appreciate those aspects of the MG-I listening experience that incite such partisanship; what we don't quite understand is why these partisans can't hear the obvious flaws that make the MG-I an inaccurate and, on balance, unacceptable full-range speaker.

The qualities that seduce the MG-I fans, whether they know it or not, are the relatively large-area launch of the wave front, the upper-midrange smoothness and the excel lent coherence (we measured accurate pulse shape replication down to a width of 0.12 msec). We like all that, too; music in the real world doesn't come out of two or three different little holes, with a different delay through each, and it helps when speaker design is cognizant of that reality. The MG-I does indeed present a rather natural sonic perspective and a focused image, especially when you reduce the rear ward radiation with a blanket or overcoat. (Permanent pad ding to attenuate the back wave, as in the Quad and Sound Lab electrostats, would of course be the better solution.) But there's more to a good full-range speaker than just that-and the MG-1 hasn't got it.

For one thing, the speaker doesn't stop speaking when the input signal stops. (See also our more detailed discussion of the inherent energy storage problems of the Magneplanar design in Vol. 1, No. 6, pp. 19-20, under the Tympani I-D review.) Tone bursts excite easily measurable ringing in the MG-I throughout its range; at 3.04 kHz and also at 8.7 kHz the situation gets completely out of hand, resulting in some of the worst ring patterns we've ever observed in our laboratory. This explains the piercing and ultimately very fatiguing quality of the MG-I's highs. We certainly couldn't live with that very long. A step function causes totally uncontrolled ringing in the bass panel; if this were a sealed system we would guesstimate the equivalent dynamic Q to be some where between 10 and 20-that's how long those ripples keep coming. Luckily, the amplitude response of the speaker rolls off very rapidly below 60 Hz, so that all low-frequency anomalies come through heavily attenuated. (This is known as 'tight bass'' in some circles, meaning that only the accentuating harmonics of the bass notes are audible and the hard-to-handle, rumbling, high-energy fundamentals are missing.) On the top end the speaker is equally deficient; above 10 kHz the response drops sharply and becomes non existent past 18 kHz. All those apparent highs are actually lower-treble ringing. Thus the most appropriate use of the MG-I would be as a midrange driver between, say, 100 Hz and 2 kHz, a range in which its best features dominate.

By the way, the sound of the speaker improves quite noticeably when its very light protective fuse is shorted out.

(See our discussion of speaker fuses in the introduction to these reviews.) On the other hand, the manufacturer threatens to void the warranty should any tweeter trouble develop as a consequence of such tampering. In any event, the aggressive highs are still there without a fuse, but they burn the ear a little less and the overall sound is somewhat more transparent.

None of the above should be interpreted as denying the obvious first-time appeal of the MG-I when you switch to it from a conventional medium-priced dynamic speaker system of conventional radiation characteristics. To our ears, how ever, the appeal is of short duration, and our instruments provide ample evidence to back up our ears.

Onkyo Model F-5000

Onkyo U.S.A. Corp., 42-07 20th Avenue, Long Island City, NY 11105. Model F-5000 Phase Aligned Array Speaker System, $999.90 the pair. Tested #09072742 and #09072746, on loan from manufacturer.

It was the four-color ads that originally attracted our attention to the three-way, floor-standing Onkyo F-5000.

They were saying all the right things, about the importance of preserving the ' 'specific phase (time) relationships' of the original signal, about ''a total solution to the phase problem,"' about laser interferometry and computer analysis as design tools, about ' 'electrostatic-like clarity, definition and center imagery,' and everything else that's dear to our heart.

Just imagine-a marketing-oriented Japanese mass-producer with a truly sophisticated, up-to-the-minute speaker design philosophy! It was too good to be true. And, in fact, it wasn't true.

When a positive-going pulse is applied to the input terminals of the F-5000, the woofer moves forward, whereas the midrange and the tweeter move backward. How any manufacturer can talk about the decisive importance of phase accuracy in a speaker system and then not drive all voice coils with the same polarity is completely beyond us. The trade marked 'Phase Aligned Array' simply isn't phase aligned! It's quite incapable of a genuine replica of a pulse. There are also serious energy storage problems in the highly publicized planar (flat-diaphragm) drivers; the tweeter and the midrange liberate energy for a period three times as long as the pulse duration. This is time-domain integrity? The woofer is a little better in this respect, but on tone bursts it rings at many frequencies in its upper range; the midrange exhibits a particularly nasty ring at 2.035 kHz; and even the unusual 'direct drive' polyamide membrane tweeter, which amazed us with its flat-to-40-kHz range, rings at 13.2 kHz. This tweeter also does something we've never experienced before; when fed a sine wave at no more than a reasonably high power level, it synthesizes half that frequency-not a higher harmonic, but f/2! Would you call that a breakup or a breakdown of the diaphragm? All we can tell you is that it sounds really weird.

Where the 'phase aligned' Onkyo shines is, paradoxically, in good old garden-variety amplitude response. We'd call it essentially flat and very smooth from 40 Hz, which is its -3 dB low-frequency elbow, all the way to 40 kHz. At 50 kHz it's down only 10 dB! The damping of the sealed-box woofer is a bit on the loose side, our step-function test indicating a dynamic Q in the neighborhood of 1.5, reason ably independent of drive level.

The sound of the F-5000 is a tremendous letdown after all the ballyhoo. Music comes out hard, raspy and edgy, especially on female voices (it could be that 2 kHz ring); speech is hollow and garbled. Spatial focus is simply nonexistent, a sure indication of time-domain trouble. ''Center imagery' indeed! Let this be a lesson to those who question our emphasis on the products of small specialist companies; someone like Richard Vandersteen or Dr. Roger West could have told Onkyo in two seconds what was wrong with their sound.

Perspective MK2

Laboratoire d' Etudes et de Developpements Holophoniques, Paris, France; distributed by MBR Electronique, 15 r. Remusat, 75016 Paris. Perspective Model MK-2 floor-standing loudspeaker, $2750 the pair ( U.S.A. price). Tested samples on loan from distributor.

This almost four-foot high floor-standing column speaker, known as the Audience I in a slightly different earlier incarnation, is something of a cult item in France, where claims of towering and self-evident superiority to all other speakers are made about it with some regularity. That happens to be sheer fantasy; however, the main reason for this review, which is based on mid-1979 samples and could with some home stretch haste have made our last issue, is that the Perspective is no longer just an esoteric Gallic curio. It has meanwhile become a full-fledged U.S. of A. rip-off, being offered in stateside high-end salons for $2750, which is an insult to the intelligence of the American consumer. There's very little more material and labor in the Perspective MK-2 than in the Vandersteen Model IIA at one third the price. The cabinet of the Perspective is heavy and well made, it's true, but four drivers from the Audax catalog and a first-order crossover network couldn't possibly bring up the price to that staggering level, even with shipping, import duty and fat profits for all middlemen.

The design goal of the Perspective is the commendable one of time-domain integrity; as far as the alignment of the drivers is concerned, this is basically achieved, with every thing in phase and quite good pulse replication (though not nearly as perfect as illustrated in the Perspective literature, which doesn't show the trailing ripples of energy release after pulse cutoff that we observed). The rather eccentric and complicated woofer enclosure design doesn't particularly impress us; using a second woofer internally as a kind of compression and rarefaction valve for the air load behind the outside woofer is wasteful of amplifier power and doesn't really accomplish anything that couldn't be duplicated by any number of less devious approaches. The system doesn't know and doesn't care how its Q and its f3 are synthesized, and there's certainly no difference in sound quality between two systems of the same Q and f3, no matter how differently these were achieved, as long as the air-moving capabilities are the same. So the simplest way, not the most original, is obviously the best way.

In the case of the Perspective, the f;3 (the -3 dB frequency) is 30 Hz, which is rather impressive, but harmonic distortion ranges from 9% to 12% between 25 and 60 Hz at not unreasonable power levels, indicating that perhaps the woofer design priorities leave something to be desired.

As far as amplitude response is concerned, the entire system is very flat from 30 Hz to 30 kHz; we have no complaints on that count. The moment of truth, however, comes when the speaker is tested with tone bursts. About the Perspective you don't ask at what frequencies it rings. You ask, at what frequencies doesn't it ring? And the answer is--don't ask.

It's just awful. At every imaginable frequency we measured profound ringing. You have to hunt for a point in the audio spectrum where the tone-burst response improves to bare decency. In addition, at 760 Hz the woofer runs into a self-cancellation problem, accompanied by a particularly strange kind of ring.

The net sonic result of this is rather unpleasant. The speaker sounds shrill, aggressive and unfocused, so much so that we could hardly believe that our samples weren't defective. But if they were, how come both measured and sounded the same? One would expect defects to be randomized.

Both the designer and the distributor of the Perspective MK2 have expressed to us the opinion that Americans don't know how to listen to music or musical reproduction critically. If they're right, the speaker has a fair chance of succeeding in the U.S. market.

QLN 1

QLN Audio, Eklandagatan 6, Box 5061, 402 22 Goteborg, Sweden; distributed in the U.S.A. by Scandinavian Sounds, 233 Esplanade, San Clemente, CA 92672. QLN I two-way minimonitor, U.S. price NA. Tested preproduction samples, on loan from manufacturer.

A Swedish attempt to beat Rogers, Fried and all the others at the satellite/minimonitor game, the QLN I partly succeeds and may entirely succeed yet, as we have word at press time that both the bass/midrange driver and the cross over network have been improved since our tests.

The design consists of the KEF T27 dome tweeter, a Dalesford cone driver for the bass and midrange, a rather complex high-order crossover network with time-delay compensation, and a surprisingly heavy and rigid little sealed box, only slightly larger than that of the Rogers LS3/5A. The woofer Q is approximately 0.6 at all drive levels, indicating superb damping; the -3 dB point on the bottom end is 80 Hz; the frequency response is extremely flat and remarkably smooth all the way up to 35 kHz; tone bursts provoke no bad behavior worth discussing, at any frequency. So far so good.

Now for the bad news. The tweeter is wired out of phase with the cone driver, so that absolute time coherence becomes an impossibility. Nevertheless, the speaker reproduces pulses pretty accurately, with the typical opposite going tweeter preshoot also observable in the similarly flawed DCM Time Window, Tangent RS2, B&W DM7 and others.

A worse flaw is that the crossover design brings in the KEF T-27 tweeter at too low a frequency for comfort. We measured full passband level out of the little 3/4-inch (19 mm) button at 3.5 kHz; that's zapping it a bit too hard in our opinion. An even greater problem is that the crossover net- work appears to present much too difficult a load to nearly all amplifiers. The same amplifiers that handle immensely taxing loads like the Beveridge and Sound-Lab electrostats without batting an eyelash go into some degree of oscillation when driving the QLN I. We didn't bother to run a vector impedance curve on the speaker, but there must be a mind blowing capacitance across its terminals.

The end result of all this a a fairly accurate, transparent and focused sound, with none of the midrange thickness of so many dynamic speakers, but rather hard, bright and edgy, as well as slightly sibilant. After a while it becomes fatiguing.

The overdriven tweeter and the ringing amplifier are the most probable culprits; we're hoping against hope that the announced “minor change in the crossover” will correct these two problems, if not the tweeter polarity. The QLN I could then be a potential winner.

Sound-Lab R-1 (formerly SL-6000)

Sound-Lab Electronics, 5226 South 300 West, Suite 2, Salt Lake City, UT 84107. 'Renaissance Series' R-1 electrostatic loudspeaker, $2200 the pair (without woofers). Tested #0058 and #0059, on loan from manufacturer.

Now this one is a real contribution to the loudspeaker art. Even in its present, early production form. we prefer it by some margin to the similarly early Beveridge System 3 we tested, except for the latter's unique wave-front launch (i.e. radiation geometry). Needless to say, both designs can be expected to undergo refinement as the months go by, but at this writing the Sound-Lab electrostatic unit is considerably more neutral in sound and also measurably freer of assorted anomalies in response. The Koss Model One/A, which we never liked on the extreme top and bottom, must also take a back seat to the Sound-Lab in midrange neutrality and transparency. What's more, the Sound-Lab can play louder than any other electrostatic design we know of.

The technical brain behind the new speaker is Dr. Roger A. West, who worked for many years with Arthur Janszen back in the days of the KLH Nine and always wanted to make a better electrostatic utilizing all the advanced ideas accumulated along the way. The SL-6000, now rechristened the Renaissance Series R-1 (huh?) upon the advice of a marketing man, is the end result. The design is crossover-less from approximately 100 Hz on up; below that you can use any subwoofer you wish. Sound-Lab supplies on request an entirely conventional subwoofer for which they make no special claim; we used our trusty Janis W-1 with Interphase 1 electronics. The electrostatic transducer consists of five separate vertical strips forming a curved panel approximately four feet high by two feet wide. A rather attractive lattice work covers the panel and divides it into five times seven little rectangles, a grid of 35. A snap-on cushion fills out the concave rear of the panel to attenuate the back wave and reduce typical dipole anomalies to a minimum. (Cf. our comments on the Koss One/A in the context of our ''Reference A'' system in the last issue and on the Magneplanar MG-I in this issue.) The system is transformer coupled but presents a somewhat higher impedance to the amplifier across most of the audio spectrum than is optimum loading for typical transistor circuits, making up for the power loss by very high efficiency. The advantage is that even at the highest frequencies the impedance doesn't quite drop to the quasi-short-circuit values of the Beveridge and other highly capacitive speakers. (See the Beveridge System 3 review above.) That transformer is responsible for the most obvious idiosyncrasy of the speaker, a sharp peak in amplitude response at 23 kHz, apparently caused by some kind of LC interaction between the transformer secondary and the transducer element. With the misnamed ' 'roll-off frequency control'' of the speaker wide open (meaning all the way up to the 33 kHz position), this peak is at its maximum of 9 dB.

Playing around with the control, you can bring down the peak at the cost of creating a big dip just below it. The optimum setting-not much peak, not much dip-seems to be at the 25 kHz position. So it's a resonance damping trimmer, not a roll-off control.

This and other possible reactive peculiarities of the transformer may account for the unsettling effect the speaker has on many amplifiers. If an amplifier has the slightest tendency to go hard and edgy on program material with lots of upper midrange and lower treble energy. the Sound-Lab R-1 will bring it out. On the other hand, Class A or near Class-A amplifiers like the Bedini Model 25/25 and the JVC M-7050 remain sweet and smooth right up to clipping when driving the Sound-Lab. A pair of Amber Series 70s strapped for mono, though not Class A, would be another good match, as they hardly ever clip while remaining almost as sweet and smooth. We also noticed some strange wrinkles in the otherwise excellent pulse profile, centered approximately 0.16 msec from the leading edge and roughly corresponding to the half wavelength of 3278 Hz, which we determined to be a ring frequency in our tone burst tests. Whether this is a transformer problem or an energy storage and delay condition in the transducer, we're not sure; nor do we know whether it has anything to do with whatever amplifiers may find indigestible in the speaker. It may take us a while to sort out the complexities of this not exactly basic cookbook design. It isn't flawless, for sure.

Otherwise the Sound-Lab is rather impressive in measurable performance. From about 150 Hz on up to 21 kHz, the head-on amplitude response is almost dead flat; the low frequency elbow is at 120 Hz and a rather large resonance at 145 Hz, so that a somewhat higher crossover than 100 Hz would be desirable, but there is adequate output still at 100 Hz. Dr. West tells us that the bottom-end resonance of the system is subject to some judicious design manipulation and that this problem can be eliminated. The segmented geome try of the active surface is almost surely responsible for the lobes we observed in the radiation pattern; however, there was no serious ringing synthesized by the interference pat terns, the one frequency discussed above being the only genuine ring point as far as we could tell. The lobeyness of radiation necessitates extremely careful speaker placement; about 7 or 8 feet apart and turned 45° in toward each other seems to be the position that minimizes the problem to the point where it isn't at all obtrusive.

Despite this mixed bag of test results, the total sonic impact of the Sound-Lab speaker is unquestionably that of a SOTA contender. Extremely transparent, very low in typical loudspeakerish colorations, limited only by the amplifier in dynamic headroom, and fast enough on top not to make us yearn for the Pyramid ribbon tweeter, the R-1 is on balance our preferred reference speaker as we go to press. An occa sional touch of hardness or edginess intrudes on certain kinds of program material, but with the right amplifier and the right placement of the panels it isn't very obvious, certainly not enough to make us prefer the worse flaws of other expensive speakers.

One little fly in the ointment was the reversed wiring of the input terminals on our samples. Red was negative and black positive, messing up the phase alignment of our bi amped system until we caught the error. Without a pulse generator, a microphone and an oscilloscope it would have been very difficult if not impossible to catch; let all owners of early samples beware. The error has meanwhile been corrected, we're told. It just goes to show that nobody is perfect, not even the remarkable Dr. West and his obviously sophisticated little company.

Swallow CM70

Swallow Acoustics, England; imported by Sunrise & Company, 85 Columbia Street, Suite 19A, New York, NY 10002. Model CM70 compact monitor loudspeaker, $495 the pair. Tested samples on loan from importer.

This smallish bookshelf system has already attracted some surprisingly favorable attention in Great Britain and is just now coming into the U.S.A. It consists of a Dalesford 8-inch woofer and the KEF T27 dome tweeter in a sealed box (yes, it's another of those), with the apparent crossover at approximately 4 kHz. The two drivers are in phase for a change, but again it's our impression that the little T27 is being driven too hard. (See also our comments on the QLN I above.) Tone bursts show an unmistakable ring pattern at3.9 kHz, right in the crossover region, and the net result is a very edgy and fatiguing sound, even if the first-minute impression is one of clarity and focus.

The bass is peaked up a l'anglaise at 75 Hz, with a Q of approximately 1.3 (going up to 1.5 with increasing drive levels); frequency response is otherwise reasonably flat all the way to 25 kHz; pulse shape retention is good down to a width of 0.3 msec only, but there's relatively little trailing garbage after pulse cutoff. None of this is particularly relevant in view of the speaker's nasty cutting edge, which must be eliminated before the CM70 can be considered an audiophile product in our book.

Vandersteen Model IIA

Vandersteen Audio, 1018 South Mooney Boulevard, Visalia, CA 93277. Model IIA floor-standing 3-way speaker system, $920 the pair (3940 east of Denver). Matching 6' high metal stands, $60 the pair. Tested #3236A and #3237A, on loan from manufacturer.

This is a modification of the Model II described, and favorably reviewed, in our last issue. Since the basic design hasn't changed, let's just restrict ourselves to whatever is different in the 'A' version. For one thing, the dome tweeter is now an Audax, not a Peerless. The 8-inch woofer is also new and also by Audax; furthermore the tuning of the vented (passive radiator) enclosure has been changed and is now much closer to optimum, with its -3 dB corner at 40 Hz. The crossover frequencies have been slightly shifted. Best of all, the three drivers are now in phase, pushing and pulling together on all transients. We wish to take just a tiny bit of credit for nudging Richard Vandersteen in these directions.

The net result of these mods is that the Vandersteen has become the smoothest, best-focused, least colored, most balanced-sounding dynamic speaker system known to us, exceeded in transparency, definition and neutrality only by the top electrostatics. All this despite the fact that the delay between tweeter and woofer is too great to permit any kind of pulse shape replication. Obviously that's only one criterion of loudspeaker accuracy and possibly not the most important.

It must be emphatically added that the above ranking applies only in the case of music played at less than very loud levels. The speaker is basically a small-signal design, with clarity rather than dynamic headroom as its top priority.

When pushed, not even terribly hard, it begins to run into trouble. The stress becomes quite audible; it isn't just a subtle loss of quality. On the lowest organ pedal passages, the woofer flirts with total breakup unless a firm hand is applied to the volume control. The midrange and the tweeter also have their very definite dynamic limits, determined in part by the choice of 6-dB-per-octave crossover slopes. The DCM Time Window is unquestionably better in this respect, though not as free of sonic colorations.

All things considered, even the last-minute increase in price as we go to press, the Vandersteen Model IIA is our inevitable choice for the speaker end of our ''Reference B'' system, since it represents a fair approximation of what we consider accurate sound reproduction, in contrast to comparably priced speaker systems with a sonic personality of their own.

Recommendations

Interesting new speaker systems are waiting in the wings (such as for example the Dayton Wright XG-10, the revised Beveridge System 3 and the Pyramid Metronome Model 3), so don't carve any of this in marble yet. But here's how we see it currently.

Best speaker system: Sound-Lab R-1, biamped with Janis W-1 subwoofer (see also Reference A update in this issue).

Best speaker system at a three-figure price: Vandersteen Model ITA (see also Reference B update).

Best speaker value per dollar: DCM Time Window.

Best tweeter: Pyramid Model T-1.

Best subwoofer: Janis Model W-1 with Interphase 1.

Best subwoofer per dollar: The Bass Mint Model 10/24.

---------

[adapted from TAC]

---------

Also see:

A Genuine Breakthrough in Inexpensive Integrated Amplifiers: NAD 3020; NAD 3045

Why We're So Mean, Vindictive, Arrogant, Negative--and Truthful

Various audio and high-fidelity magazines