Sophisticated Speaker Systems, Large and Small: Continuing Our Survey (Vol. 1, No. 5)

Sophisticated Speaker Systems, Large and Small: Continuing Our Survey --- Beveridge 'System 2SW' (follow-up) ; Canton HC 100 ; Dayton Wright XG-8 Mk 3 ; DCM 'Time Window' (Improved) ; Hartley 24' Subwoofer ; Innotech D24 ; Spendor BC1 ; Tangent RS2 ; Ultraphase 2501.

-----

Part II: In which a new small box goes to the head of its class, a $3000 audiophile legend falls flat on its face, a "best buy' turns out to be even better, and our reference speaker remains unchallenged.

We have very little to add at this point to the general discussion that preceded our first batch of reviews in this survey (see our Number 4 issue). Our assumptions and our methods remained the same when we tested the speakers reviewed below.

The only further insights we have gained since the first go-around have to do with the specific sonic quality of certain measurable speaker characteristics. We have zeroed in a bit more precisely on a number of frequently observed anomalies.

For example, a steely, ear-piercing quality has generally very little to do with frequency response. It can nearly always be traced to ringing in the 2 kHz to 5 kHz range, where the ear is most sensitive. A tipped-up amplitude response at the higher frequencies may sound over-bright but doesn't "burn" the ear unless there's considerable ringing, identifiable with tone bursts.

Poor time response, as indicated by the inability to reproduce a recognizable pulse, is usually heard as a homogenized blending of inner textures and a lack of delineation of spatial detail. It is not irritating or unmusical if the amplitude response is reasonably smooth and ringing is moderate or absent. For that reason a lot of people throw their hats in the air and shout "State of the Art!" when they hear silk-smooth, wide-range audio without noticeable graininess or grit, unaware that the requirements of accuracy go well beyond that.

Genuinely lifelike sound reproduction is characterized by an open, focused, see-through quality that cannot be achieved without proper attention to time response.

Underdamped woofers also have a characteristic sound, a whompy heaviness that passes for "powerful" bass in some circles.

We continue to find most woofers delinquent when it comes to properly calculated damping; their Q is unnecessarily high. The whomp can sometimes be made more tolerable to the ear by placing the speaker away from all boundaries, i.e., far from the walls and off the floor.

(Admittedly not a practical solution under all circumstances.) The honky midrange quality that degrades the performance of so many expensive speakers is also due mainly to ringing, anywhere from 250 to 800 Hz. Often the woofer is required to reproduce these and even considerably higher frequencies, without any discernible attempt on the part of the designer to control cone behavior above the "piston" range.

In general, it's safe to say that pressure amplitude response (what is loosely referred to by audio people as "frequency response") is no longer the limiting factor of speaker performance as it used to be years ago. To day's main shortcomings are undamped oscillations (i.e., ringing), severe trade-offs be tween accuracy and dynamic range, and poor time response.

The Audio Critic as a consumerist force.

It may be interesting to our subscribers to learn that Part I of this speaker survey sent at least five speaker designers back to the drawing board. We have reliable reports on this in each instance. And we find the situation both encouraging and discouraging.

Encouraging because it shows that there are designers who are more interested in how their product will perform tomorrow than in how infallible their views were yesterday. Of such stuff is progress made. It also proves that consumerism can work even in the frenzied, ego-tripping world of high-end audio.

On the other hand, we consider it discouraging that our tests were news to so many people who earn their daily bread with audio engineering. The Audio Critic has so far developed no original concepts in testing. We may look into a lot of things that others neglect, but everything we do is based on information and procedures obtainable in any de cent engineering library, most of it in the last 14 or 15 volumes of the Journal of the Audio Engineering Society. We don't own a single laboratory instrument-or audio component- that can't be purchased or at least borrowed by any speaker company. The ears of our staff are well-schooled but in no way phenomenal.

And the laws of nature are the same at our address as elsewhere. So how come our test results are a surprise to so many professionals? Let's proceed to our reviews before we get too worked up over the industry's mentality.

Beveridge 'System 2SW' (follow-up)

Harold Beveridge Inc., 505 East Montecito Street, Santa Barbara, CA 93103 (note new address).

Beveridge Cylindrical Sound System, Model 2SW, $6000 the pair (new price, including HD subwoofers, plug-in power amplifiers and CM-1 control module). Unlimited warranty on all parts except tubes (one year); five-year warranty on all labor, including sonic updates. Tested #233 and #234, on loan from manufacturer.

Now that we have lived with this system for a number of months and used it as our principal reference speaker for evaluating pre amps, phono cartridges, etc., we believe more than ever in its overall superiority to other designs and at the same time are frustrated more than ever by its obvious deficiencies.

From about 100 Hz on up, the Beveridge is simply more accurate than any other speaker known to us; it reveals more about the signal fed into it than any other and just sounds more like real life. No question about it. We now feel that Harold Beveridge's idea of creating a coherent cylindrical wave front out of a planar wave by means of a limited number of wave guides (his amazingly simple "acoustic lens") is one of the major conceptual break throughs in the recent history of audio; further more, on the purely practical plane, his electro static transducer "sandwich" is the best in the business. These two points of superiority give the Beveridge system a substantial advantage right up front. But there are a few flies in the ointment.

As we pointed out before, the woofer is far from SOTA; a $6000 system (that's the latest price) ought to have tighter, deeper, more awesome bass. A 12-inch driver in a sealed box just can't do the ultimate job, even with the Q and all other parameters optimized (which they don't appear to be in this case).

What we'd like to see here is something like the sheer air-moving capability of the 24-inch Hartley, combined with the damping characteristics of, say, the Cizek or the Fundamental Research. After all, the woofers alone cost $500 apiece if you order them separately.

As for our complaints about the SPL limitations of the Beveridge, we've just made an interesting discovery. The sheer noise producing capability of the system is much greater than we would have believed on the basis of our original tests with musical reference material. For example, the railroad sounds on Side Two of The Power and the Majesty (Original Master Recordings, Mobile Fidelity Sound Lab MFSL 004) can be played loud enough to drive everybody out of the room without any obvious distress signals from the speaker-but then we must admit that our ears aren't attuned to subtle distortions in loco motives. On the other hand, the New Haven Brass Quintet on Side D of Volume Four of the Mark Levinson Acoustic Recording Series begins to sound edgy and clipped when played at a you-are-there level, whereas a couple of dB below that it's absolutely gorgeous. This leads us to believe that the frustrating dynamic limitations of the Beveridge may conceivably be due to nonlinearities in the amplifier rather than to the inherent SPL ceiling of the electro static panels. If we're right-and of course we may just as easily be wrong-this is very good news, since the amplifier could be much more easily redesigned than the electrostatic panels.

Harold Beveridge is well aware of these considerations and will undoubtedly come up with an answer-if there is one.

As far as the 8 dB peak at 76 Hz is concerned, it was a prototype boo-boo that has been fixed in the production model. Our early samples were modified accordingly and we can now give them a clean bill of health. The peak is measurably and audibly gone.

Which still leaves us with a speaker that we can neither exalt without a number of reservations nor replace with a better one.

It isn't as good as it ought to be; it's merely the best there is.

Canton HC 100

Adcom, 114 East 32nd Street, New York, NY 10016. Canton HC 100 miniature speaker system, $180 the pair. Five-year warranty; manufacturer pays return freight. Tested #002673 and #002674, on loan from manufacturer.

Teeny-weeny speakers with an audiophile appeal seem to be a German specialty; this one is Canton's answer to the Braun 'Output C' mini-system. The HC 100 is only very slightly larger, its longest outside dimension being 18.5 cm (7.5 in), but it has an irregular shape with a sloping front grille and is intended to be deployed horizontally. The driver complement consists of a 4" cone and a 1" dome, crossed over at 1700 Hz.

We happen to be somewhat obsessed with this type of speaker, always hoping against hope that one of them will turn out to be absolutely superb above 200 or 250 Hz and could then be used with a hidden subwoofer as an almost invisible system of near-SOTA performance. No such luck in this instance, even though the HC 100 has some very commend able qualities, just like the larger Canton LE 400 reviewed in Part I.

"The system resonance of the sealed-box woofer (if you can call a 4-incher by that name) is in the neighborhood of 200 Hz, and the system Q is approximately 2, which of course is too high-but all mini-systems try to fake extended bass response that way. Be tween 500 Hz and 20 kHz, however, the response of the system is spectacularly flat, very similar to that of the LE 400. The Canton people seem to have this part of the problem pretty well figured out. What's more, our tone burst tests proved the HC 100 to be remarkably free from ringing.

Our pulse tests, on the other hand, indicated that all was not well in the time domain.

The HC 100 appears to be totally incapable of reproducing pulses of any duration with even a semblance of coherence. And that makes it a lot less interesting to the audio perfectionist than it might have been.

The sound? Exactly what you'd expect.

Very smooth, sweet, balanced and non-fatiguing, remarkably wide in range, but without any real focus, air or spatial detail. You might as well be listening to mono. The Braun Output C is somewhat better in this respect, even though it rings more and is less flat than the HC 100. One day, either Braun or Canton will get this whole act together, but it just isn't happening yet.

We hear that one of the preferred uses of the Canton HC 100 is as a Mercedes Benz or BMW custom speaker. Since the drivers of these cars are already satisfied in the time domain, so to speak, that probably works out very nicely.

Dayton Wright XG-8 Mk 3

Dayton Wright Associates Limited, 350 Weber Street North, Waterloo, Ont., Canada N2J 4E3. XG-8 Mk 3 Series 3 full-range electrostatic loudspeaker, $2995 the pair. Five-year warranty. Tested #24224 and #2422B, with Model ST-300A matching transformer and bias supply #0723, on loan from manufacturer.

Whew! What a letdown. Here's the speaker that epitomizes exotic audiophile equipment: completely offbeat design, limited production, stratospheric price, fanatical cult following, ritualistic explanations. And the sound isn't even pleasing, let alone accurate.

We must admit that every experienced audiophile who heard our pair of Dayton Wrights remembered the speaker as having sounded better in earlier versions. We had the very latest Mark 3 Series 3 model, made entirely under the new Leigh Systems regime, after Mike Wright's departure. There was certainly nothing defective about our samples; someone at the company had warned us about the possibility of overinflation of the "gas bag", but that was definitely not the case.

(The 10 full-range electrostatic cells of the XG-8 are sealed in a special gas environment that permits much higher operating voltages than air.) We're fully satisfied that what we tested was representative of current production; what the speaker used to be is a moot point.

This particular pair of Dayton Wrights sounded honky and unclear in the midrange, aggressively hard, and almost instantly fatiguing. The only good things we heard were quite decent bass for a wooferless dipole speaker of one square meter (11 sq ft) area, plus excellent dynamic range. (Piano recordings could for once be played at the same level as a live piano.) But the colorations of the speaker were too severe for us to take it seriously as a reference-quality system.

Before we go into the specifics we want to reassure the cultists. We had the Dayton Wrights plugged in and charging for more than a week before we even went near them.

We then listened to them, measured them, and listened to them again. At length. After another month of uninterrupted charging, we tested them again. To those who believe that even this wasn't enough, that it takes six months of continuous charging before a Day ton Wright is ripe for evaluation, all we can say is that we were looking for a good loudspeaker, not a good cheese.

Our measurements revealed some serious problems in both the frequency domain and the time domain. The bass response of a dipole is hard to measure (see our comments under the Koss Model One/A in Part I); we estimate that the -3 dB point of the Dayton Wright is at 45 Hz, with some roughness in the 40 to 60 Hz region. From there on up we observed a constantly rising response to a peak at 700 Hz, with a falling response thereafter. In the 14 to 15 kHz region there's another bad peak, which comes from the 'piezoceramic ultra-tweeter" used at the highest frequencies.

(We don't believe that such a device belongs in a full-range electrostatic system, disagree with the rationale given for it in the Dayton Wright literature, and would censure it in greater detail here if it were the speaker's only flaw.) These two peaks dominate the sonic "signature" of the XG-8 Mk 3. The 700 Hz peak, which persists regardless of the angle of measurement, has an amplitude of approximately 7 dB as compared to the surrounding frequencies and is almost certainly the main cause of the speaker's honky coloration and un-clarity in the midrange. The tweeter peak has an amplitude of 10 dB with a fairly low Q (i.e., its base spreads over a fairly broad section of the spectrum) and is the probable cause of the zingy quality of the highs.

Even more significantly, severe ringing occurs at both of these peaks (namely at 700 Hz and 14.7 kHz), as well as at 1.8 kHz and 4.5 kHz, where we observed no comparable amplitude anomalies. In general, the speaker is highly prone to ringing through out the audio range, with many other trouble spots where tone burst reproduction is rather poor, though not as bad as at the four frequencies mentioned. Occasionally the ringing drops to reasonable levels, but on the whole there can be no doubt that the system's integrity is ruinously compromised.

Exploring the XG-8 Mk 3 with widely separated pulses revealed that the rise time in some instances is quite a bit slower than one would expect of an electrostatic transducer, and time smear was also evident at a number of frequencies. Not a glorious picture, all in all.

The way we see it now, the main and perhaps only permanent contribution of the Dayton Wright electrostatic system to the advancement of audio is the technology of operating electrostatic panels in a sealed, gas filled environment. By operating at voltages far higher than the ionization threshold of air, the Dayton Wright permits loudness peaks and a dynamic range comparable to, say, the Pyramid Metronome's. That's about it. We can't think of anything else to admire in the speaker, especially in view of its $3000 price.

Oh yes. We did use the new stands that raise the Dayton Wrights off the floor and tilt them slightly backwards. We didn't, however, drive them with the Threshold amplifier as the cult demands, but with a measly Bryston 4B.

Perverse, aren't we? (Those who feel that the Bryston causes peaks and ringing in the Dayton Wright, whereas the Threshold does not, are invited to send us their revisions of the laws of physics for immediate publication.) DCM 'Time Window' (Improved) DCM Corporation, 2275 South State Road, Ann Arbor, MI 48104 (note new address). 'Time Window' floor standing loudspeaker, $660 the pair. Five-year warranty. Tested #1853 and #1864, on loan from manufacturer.

A number of worthwhile improvements have been made in this already remarkable speaker, so that a reevaluation is in order.

The main visible change is the switch to Mylar domes in the tweeters (with mixed results, as we shall see), but the overall sonic character of the speaker is sufficiently different to indicate that other engineering details have been retouched as well. You can assume that all units with a serial number above 1600 incorporate these refinements.

The improved version of the Time Window sounds even more open and transparent than the original, with smoother and better-defined highs, and hardly any whomping or thickness in the upper bass and lower midrange. It's simply a more natural sound, immediately apparent in an A-B comparison.

The amplitude response of the speaker is now extremely smooth between 200 Hz and 14 kHz, within +3 dB we'd say. At 14 kHz there's a rather nasty tweeter peak that definitely wasn't there before; its amplitude varies with the measurements angle (9 dB on the axis of the dome, 6 to 8 dB in the composite response curves), but it's always present and cannot be "homogenized" out by moving the microphone. The tweeter response drops precipitously' after the 14 kHz peak; at 20 kHz it's good-bye and gone (down to -14 dB). Interestingly enough, this anomaly has relatively little effect on the speaker's audible performance, probably because the Q of the peak is quite high (i.e., the spurious energy is contained within a rather narrow band of frequencies). Tone bursts confirm the reality of the peak, exciting severe ringing at 14 kHz, even though the rest of the audio range is quite as free from ringing as before.

On the low end, the tuning of the vented enclosure is still highly suspect, although it appears to have been changed slightly. The vent still doesn't fill in correctly the null in the driver response (in this case at 32 Hz), and the composite bass response is still quite lumpy-we'd call it +5 dB from 40 Hz on up.

On the other hand, the woofers are well behaved under transient excitation; they shut up when the signal stops. In this respect we see (and hear) some improvement.

The main reason for the improved sound of the new Time Window, however, is almost certainly the further refinement of impulse response in the critical midrange and lower treble region. Pulses between 1 msec and 0.15 msec duration are now reproduced with even greater accuracy than before; the speaker has become virtually flawless in this range. Pulse form retention deteriorates rapidly with durations of 0.15 msec and shorter, possibly because of the tweeter problems observed.

There's a trade-off here: better pulses from 1 to 0.15 msec, worse from 0.15 to 0.1 msec and shorter. And it seems to be a trade-off that favors the sound.

Overall, we find the improved Time Window to be the nearest thing to a high-fidelity speaker in a single package of moderate size and affordable price. The Tangent RS2, re viewed below, is even more transparent and uncolored from about 200 Hz on up and costs less, but its bass is unacceptable without the addition of a subwoofer. The Time Window, on the other hand, can be enjoyed as is, straight from the carton.

Now if they'd only fiddle some more with the enclosure tuning and that tweeter peak . . .

Hartley 24" Subwoofer Hartley Products Corporation, 620 Island Road, Ramsey, NJ 07446. 24-inch Woofer-Driver, $375 each (without enclosure). Tested samples on loan from dealer.

This is widely accepted in audio extremist circles as the Mount Everest of subwoofers:

when size, convenience or price is no object, it's supposed to be the summit of the art. Well, it isn't.

We tested it in a completely sealed and heavily braced custom enclosure, 4 feet wide by 2 feet deep by not quite 3 feet high, with an estimated internal volume of 18 cubic feet.

It took four good men and true (admittedly not professional furniture movers) to lug a pair of these monsters from the van to the laboratory, and after the second one they swore they'd never do it again. In other words, the "infinite" baffling of the woofer was carried to the outermost limits of practicality, short of building it into the wall or constructing the enclosure in a room where it would have to stay forever. And, as it turns out, the Q of the woofer is much too high in a mere 18 cubic feet, suggesting that flat bass could only come out of it from the middle of an "infinite" wall.

We measured a 6 dB hump at approximately 48 Hz, indicating a system Q of 2.

The -3 dB point of the system was at 33 Hz: the corner frequency as normalized for the hump was estimated to be 36 Hz. So the Godzilla of the low frequencies turned out to be a 36-Hz box, somewhere between the 10-inch Cizek and the 10-inch Snell-and not nearly as well damped as either. In fact, pulsing the woofer with four-cycle tone bursts at 100 Hz, 50 Hz, 40 Hz and 30 Hz revealed a distinct half cycle of ringing in each instance. The system was obviously underdamped, and a woofy, whompy quality on bass transients was easily ascertainable by ear. This is one area where measurements and the sound of music always correlate.

In all fairness to Hartley partisans, we must point out that despite its inability to reach into the 16 to 32 Hz octave and its in adequate damping, the system could move the air in the room like no other subwoofer known to us and was able to generate truly lifelike midbass pressure levels. Piston area isn't some thing to be sneezed at. Since the quantity of low-frequency energy has just as much to do with realistic sound reproduction as the quality, the Hartley remains in the running against more accurate but less potent subwoofers. But that's only because no one has bothered to design mathematically an 18-cubic-foot sub woofer of optimum performance. If the problem were thus stated and solved from scratch, the required driver parameters (moving mass, compliance, flux density, voice coil dimensions, etc., etc.) would be quite different from the 24-inch Hartley's.

Quite aside from the Q problems, though, there are some funny things about the basic concept of the driver design. Would you believe a 1.25-inch voice coil moving a 21.5-inch cone? It's like pushing a shopping cart with your pinkie. Would you believe a cone ribbed and stiffened for good high-frequency response? The damn thing goes out to 3 kHz! Who would want to use it above 100 Hz? And who wouldn't welcome some sort of smooth mechanical roll-off above a couple of hundred hertz to help make the crossover design a little simpler? But that's not all. The 96-mm (3.5 in) aluminum tube that sticks out from the middle, purportedly to act as a heat sink for the voice coil, is alive! The axial response shows a double peak at 11.5 kHz only 10 dB below the bass reference level and a single peak at 29 kHz that's only 15 dB down. We don't think this high-frequency garbage could be activated through any reasonable crossover network, but we still wonder why the Hartley people threw in a peaky coaxial supertweeter at no extra charge.

One day, a nice, clean-cut boy who always did his physics and math homework and never consorted with long-haired, T-shirted audio mystics will graduate from engineering school.

On the back of the envelope his diploma came in, he will design in one sitting a large but utterly simple subwoofer capable of producing profound and accurate bass while happily obeying the laws of nature. It will be the dawn of a new era, just like the day Galileo dropped those iron balls from the Leaning Tower of Pisa.

Innotech D24

Innotech, 42 Tiffany Place, Brooklyn, NY 1123. Model D24 floor-standing speaker system, 3750 the pair. Five-year warranty. Tested samples on loan from manufacturer.

We can dispose of this review rather quickly, since the Innotech is directly competitive with the DCM 'Time Window' and doesn't sound nearly as accurate. Both are floor standing units of the same height (3 feet), with comparable elbowroom requirements; both incorporate four drivers; both go down a little bumpily to 50 Hz and then roll off quite rapidly; both are intended to appeal to the enlightened audiophile; and both cost roughly the same (the Innotech is actually a bit more ex pensive). But the Time Window is unquestionably the more sophisticated and more highly perfected design.

The main weakness of the Innotech is the 12-inch Mylar midrange dome, which rings severely throughout its range and interacts with the other drivers in various unwelcome ways, affecting both frequency and time response. With a better midrange driver the speaker might have turned out differently.

We also question the extremely close spacing of the four drivers, which appears to be the cause of severe lobes in the polar response.

The net result is a highly colored sound with a distinctly "canned" quality, the only re deeming feature being the absence of aggressiveness or "ear burn," probably because the tweeter is reasonably good.

A recent CBS Technology Center report on the Innotech D24, published in High Fidelity, contains the amazing assertion that "the tone-burst response is good." Did they check 2.5 kHz and 4 kHz, for example? And if that's good, what's bad?

Spendor BCI

Audio International, Inc., 1055 Thomas Jefferson Street NW, Washington, DC 20007. Spendor BCI vented-box loudspeaker, $570 the pair. Tested #13478 and #13479, on loan from distributor.

This nice-looking English box speaker, about two feet high by one foot square, comes with a good international audiophile reputation. We found it to be quite listenable; however, both the Rogers LS3/5A and Tangent RS2 are more refined examples of the same civilized BBC-monitorish sound and cost less to boot.

The Spendor BCI manages to achieve in a vented enclosure what we're already accustomed to in sealed-box speakers from England: completely underdamped bass. The composite response of woofer and vent shows a hump of approximately 6 dB, centering on 72 Hz. You can hear the hump, too; we found the bass boom decidedly objectionable regardless of speaker placement. From 100 Hz on up the response is very acceptable, especially at 45° off axis, where it remains within +5 dB all the way to 20 kHz. The axial response is ragged, though, with crossover troughs evident.

In the time domain the speaker behaves better than most; pulse form retention is very good to excellent between 1 msec and 0.1 msec, but the woofer needs 2.5 msec to re cover after a 0.5 msec pulse and the midrange driver keeps on going after a 0.1 msec pulse.

The clear, focused, neutral sound of the BCI is probably due to the generally good impulse response; the midrange, however, sounds slightly rough.

When we switched to the Rogers after extended listening to the Spendor, the Rogers sounded considerably smoother and more transparent; as for the Tangent, we prefer it even to the Rogers and therefore unequivocally to the Spendor. All of which doesn't mean that the Spendor BCI isn't a rather nice speaker (except for that bass boom), but as we've said before, we don't give consolation prizes.

Tangent RS2

Sound Physics Labs, a division of Inception Audio Ltd., 2 Carlton Street, Suite 919, Toronto, Ont., Canada MS5B 1J3. Tangent RS2 Reference Speaker, $480 the pair. Lifetime warranty (conditional). Tested #0771024 and #077102B, on loan from importer.

The ink was hardly dry on our enthusiastic review of the Rogers LS3/5A when we received our samples of this somewhat larger but still highly portable and basically similar British import. And, would you believe it, the Tangent is an even better speaker. Rule, Britannia . . .

The Tangent RS2 is just one of a whole family of interesting speakers designed by John Greenbank in England, but it's the only one we've been able to get our hands on so far. (The only North American source we're aware of at this writing is the Canadian importer listed above.) Model RS2 is a completely sealed box of approximately 19 liters (2/3 cubic foot) internal volume, containing the famous KEF T27 dome tweeter, an Audax 8" woofer and a "phase-corrected" crossover network. The basic concept of the design is to achieve a reasonable degree of time alignment between the two drivers while mounting them close together in an ordinary rectangular box, i.e., without stringing them out on a staggered baffle. This is the main difference between the Rogers and the Tangent; the former can be time-aligned only by turning the woofer end closer to the listener than the tweeter end, whereas the Tangent is acceptably aligned when listened to head-on.

It's interesting to observe how differently these two speakers utilize the same tweeter.

In the Rogers, the KEF T27 is covered by a perforated screen; in the Tangent it sits naked.

As a result, the Rogers goes out comfortably to 18 kHz, but the Tangent is flat to 32 kHz and doesn't show signs of quitting up to almost 40 kHz! We really don't know the reason for this difference in application or whether it has anything to do with the audible differences be tween the two speakers-but what a tweeter! If it were more efficient and able to handle more power, it would undoubtedly be every speaker designer's automatic first choice.

As a full-range system, the Tangent is very close to dead flat in response (especially at its "sweet spot'') from 300 or 400 Hz on up into the first ultrasonic octave. From about 200 Hz down, however, it's another story. Would you believe a hump of 8 dB at 75 Hz? It's that English bass again, grossly underdamped (the Q is about 2.5) and distinctly audible as a thick whomp that underlies the otherwise startlingly beautiful sound of the speaker. The only thing that could cure it would be more magnet, and more magnet would cost more pounds sterling. (We should never have given up India, sir.) Still, a very flat speaker overall.

In the time domain, the RS2 shines even more brightly. Pulse reproduction is truly excellent anywhere between 1 msec and 0.1 msec duration-and even a little beyond either end of that range. The angle at which pulse coherence is best is still somewhat off axis toward the woofer but far closer to a normal listening angle than in the case of the Rogers. On the whole, we haven't encountered a single electro dynamic speaker system that equals the Tan gent in this area, since even the improved DCM Time Window falters on the shorter pulses. One would have to go to the $6000 Beveridge electrostatic to get better impulse response.

On the other hand, both the woofer and the tweeter of the RS2 exhibit some ringing at various frequencies when tested with tone bursts-not nearly as much as most speakers but certainly more than the improved Time Window and a bit more than the Rogers as well. The excellent pulse form retention (i.e., freedom from time smear) appears to carry the day, however, since the Tangent sounds more accurate than the other two-as long as you can mentally tune out that bass boom.

(Pulling the speaker far away from all walls and raising it high above the floor will help considerably.) When we say more accurate, we mean both smoother and clearer, better focused, more spacious and detailed, less colored. The Rogers sounds a little hard and zippy by com parison (only by comparison!) and the Time Window a bit less open (only by comparison!).

The Tangent is the most electrostatic-sounding electrodynamic we've found so far.

As for adding a good subwoofer to eliminate that bass boom, it isn't as simple as it may seem. With an 8 dB hump at 75 Hz, you can't just insert a ready-made crossover, whether passive or electronic, at 100 Hz or even 150 Hz, and call it a day. The transition wouldn't be smooth enough. We're looking into this problem and may possibly come up with a recommendation in the next issue.

(We're no longer experimenting with cross overs or biamping for the Rogers, since we now prefer the Tangent.) Meanwhile, we suggest that you use the Tangent RS2 as is (if you can get a pair) and try to forget about the bass. If you're a fanatic about clarity, as we are, you'll find the trade-off worthwhile. If not, you'll probably prefer the improved DCM Time Window, which is a better balanced full-range speaker and also considerably more efficient. (The Tangent needs a fairly powerful amplifier for program material with a high peak-to-average ratio; in fact, we were able to make the Mark Levinson ML-2 clip on it with a master tape of piano music.) In any event, it appears that little boxes have come a long way.

Ultraphase 2501

Ultraphase, 2875 South Raritan Street, Englewood, CO 80110. Model 2501 floor-standing speaker system, 8596 the pair. Wooden stands, 332 the pair. Ten-year warranty. Tested #10056-25 and #10057-25, on loan from manufacturer.

This somewhat offbeat three-way system is about the size of the largest bookshelf speakers but much heavier, with unusually thick and rigid walls, and designed to be placed off the floor on a low pedestal. The sloping front baffle of the sealed enclosure holds a staggered array of three drivers: 1" dome tweeter, 22" dome midrange, and 8" woofer.

Of its type and its price range this is a very decent speaker, considerably more ac curate in response than what seems to be generally available. It lacks, however, that extra measure of engineering refinement that gives speakers like the Rogers LS3/5A, Tangent RS2 and DCM 'Time Window' their special appeal to the serious audio enthusiast. It doesn't quite have their smooth, transparent, unstrained quality in the midrange and on top, and therefore isn't quite as musical and listenable.

Where the Ultraphase shines is the lower frequencies. For once there is no boom, no woof, no whomp. No hump whatsoever in the bass response. This is a low-Q design; in fact the response profile suggests a somewhat overdamped condition, which of course is still greatly preferable to the usual underdamped boom box. Taking the 200 Hz level as 0 dB, the -3 dB point is at 70 Hz and the -6 dB point at 45 Hz. This is the kind of smooth, gradual roll-off that can be judiciously re equalized with a few dB of bass boost in the playback electronics while still preserving an acceptably low Q-a definite advantage in flexibility over the boom boxes. Needless to say (at least to those who believe that not even speaker designers can defy the laws of nature), the Ultraphase sounds considerably more natural and untroubled on heavy bass transients (plucked double bass, bass drum, left hand chords on piano) than any speaker with a high-Q woofer. On the other hand, Ultra phase's claim that "you will feel frequencies below audibility . . . you will feel a concrete floor vibrate" is the most arrant nonsense.

No 8" woofer, regardless of its small-signal response characteristics, has the large-signal capability to do that.

Above 200 Hz, the Ultraphase can be adjusted to be quite flat in amplitude response by fiddling with the midrange and tweeter level controls in the back. We doubt, how ever, that the consumer could do this accurately without laboratory instruments and we disagree with the entire philosophy behind such controls. (They can't really make the speaker response zig where the room characteristics zag-that takes infinitely more sophisticated equalization-and they destroy the impulse response of a speaker when in correctly set.) We found the best setting to be quite far down ( 10 o'clock) on the "mid" control and all the way up ( 3 o'clock) on the "high" control. That still didn't make the speaker as flat as a Snell or a Canton or a Tangent, but it was good enough to obviate all complaints.

In the time domain there was good news and bad news. The Ultraphase is claimed to be time-aligned, and its drivers are staggered ostensibly to that purpose. Our tests indicated that the leading edges of the wave fronts from the three drivers coalesce only when the speaker is tilted much further backward than is practical, with the bottom side partly exposed to the listener and the woofer facing not far from straight up. Pulses of 0.45 msec duration were the shortest that the speaker was able to reproduce at any angle; narrowing the pulse toward 0.1 msec resulted in increasingly severe pulse form deformation.

Pulses between 1 msec and 0.45 msec duration (probably more important from the listening point of view) were reproduced quite accurately.

The cavities that the thick front baffle creates below the midrange driver and the tweeter as a result of the staggered mounting arrangement (flush mounting on a stepped baffle would have been better) must take the blame for resonances and reflections that have exactly the same effect as ringing. Actual diaphragm ringing was observed only in the neighborhood of 5 kHz; above 7 kHz the reflections from the baffle cavities made accurate tone burst testing impossible. Over all, the midrange driver definitely appeared to be the least well damped. The woofer exhibited no ringing at all when driven moderately, but at very high levels its initially low Q (perhaps 0.4 or 0.5) rose to approximately 1.0, and there developed a moderate amount of ringing. This is to be expected in all cases where the number of voice-coil turns in the gap isn't absolutely constant regardless of excursion, The net audible result of all this is very good balance, imaging, clarity and focus, but with a definite edginess that we found fatiguing. It seems that some of the ringing occurs at the most irritating frequencies. The bass, as we said before, is very classy, but of course it's still 8-inch bass.

One thing that detracts severely from our basically positive reaction to this thought fully conceived speaker is the promotional literature Ultraphase is using to sell it. For a small company, it looks as if quite a few dollars were spent on a 2501 brochure, a more elaborate brochure on the four previous Ultra phase speakers, an "Audio Consultant's Hand book" and so forth-all of them full of the most egregious misinformation. The worst of it is that the literature takes an extremely de tailed, technically informative, almost educational approach, which is likely to mess up innocent heads very effectively and for a long time. Among the more flagrantly erroneous principles promulgated by Ultraphase are that a heavy woofer cone can't reproduce bass transients as accurately as a lighter cone . . .

that correct time alignment means positioning the voice coils along a single vertical line . . .

that an impedance peak in a speaker constitutes an undesirable load for the amplifier . . .

that a vented enclosure can't have flat frequency response . . . we could go on and on.

(We've already mentioned the bit about shaking a concrete floor with an 8-inch woofer.) If this stuff were in a national magazine ad, we'd give it a good roasting in The Admonitor; as it is, we won't go further than to wonder aloud where it all came from. We've spoken to Bill Kennedy, the engineer who designs the Ultra phase speakers, and he seems to know what he is talking about. Was an advertising copy writer with a pop-tech background foolishly given his head? The Ultraphase is a good speaker with some flaws that can probably be ironed out.

The last thing it needs at this stage of its evolution is a credibility problem created by its own promoters.

Recommendations

The only change here, as a result of this last batch of tests, is our preference of the Tangent RS2 over the Rogers LS3/5A. Our other two recommendations remain the same.

Best speaker system tested so far, regard less of price: Beveridge 'System 2SW' (with reservations about subwoofer and headroom- see review).

Best speaker system at a much lower price: Tangent RS2 (with strong reservation about bass-see review).

Best full-range system per dollar (without major trade-offs): DCM Time Window.

-----

[adapted from TAC, Vol.1, No.5 ]

---------

Also see:

Sophisticated Speaker Systems, Large and Small: A Comparative Survey

A Comparative Survey of Power Amplifiers: Part III ---Audionics CC-2 ; GAS 'Grandson' ; Mark Levinson ML-2

Various audio and high-fidelity magazines