To riff on Wes's conclusion, "If it ain't broke, it don't need fixin'." So I was somewhat puzzled when I first saw the Anat's successor, the Sonja 1.3, at the 2013 Consumer Electronics Show. The Sonja 1.3 comprises the Sonja 1.1, the top module, which can be used on its own and costs $38,800/pair (footnote 1). Adding the upper woofer module gives you the Sonja 1.2 ($72,800/pair), while completing the package with the lower woofer module results in the Sonja 1.3 ($106,800/pair).

Superficially, the new speaker looks identical to the older one: the same height, the same form—three separate modules constructed from aluminum panels mounted atop one another—and the same drive-unit array: a 1" silk-dome tweeter mounted within an almost circular waveguide between two 6" midrange drivers, and two 10" woofers. But on closer inspection, it became apparent that the Sonja 1.3 is, in many ways, a completely different design.

The Sonja . . .
The modules of the Anat Reference II Professional were flat-sided, with the HF/MF module rectangular and the two woofers truncated pyramids of different heights but identical volumes, which gave the speaker's appearance a Bauhaus flavor. By contrast, the Sonja 1.3's modules feature subtly radiused side panels, the increasing radius of the lower modules producing an elegantly concave profile. The woofers are no longer powered but passively crossed over, the crossover for both modules residing in the bottom one. (When just the Sonja 1.2 is purchased, the low-frequency crossover is housed in the 1.2 woofer module; this is replaced by a blanking plate when the 1.3 woofer is added.) The crossover filters are a proprietary topology, said to add in-phase in the overlap regions, and use expensive, close-tolerance Mundorf capacitors and "Zero Ohm" inductors. One toroidal air-core inductor is wound in-house by YGA, and the Sonja's internal wiring is all Kimber Select.

Most significant, whereas the Anat and its variants used paper-cone midrange units and woofers, the Sonja's drivers feature the BilletCore diaphragms that made their appearance toward the end of the Anat's life. Each cone begins as a circular blank of 6061 aluminum alloy, an aircraft-grade metal that finishes well and doesn't corrode like harder aluminum alloys. A five-axis CNC machine first balances the blank, then increases its rotational speed so that metal can be cut away to produce the finished cone. It takes about three hours of machine time to produce the woofer cone, which has a thickness before anodizing of 0.25mm, this down from an initial time of eight hours. To produce a midrange cone, which is 0.2mm thick, takes about 90 minutes (both times including setup). The finished cones are sent out for hard anodizing, then shipped to Denmark with the surrounds to be assembled into complete drive-units, the Danish company providing all the "soft" parts, such as the spider.

The final woofer cone weighs 46gm; for reference, the cone of the 10" woofer used by KEF in its R207/2, which I reviewed a few years back, including the voice-coil and its former, weighs 40gm. There is therefore a tradeoff between increased mass and reduced sensitivity with a machined-alloy cone. However, YGA's founder, Yoav Geva, feels that this is worthwhile, given that the BilletCore cone is going to be truly pistonic not only throughout its passband but well beyond it. The use of machined metal unstressed by stamping or extrusion makes the cones extremely rigid and strong—the midrange cone has a mass of only 8gm, but can take 1000 lbs of vertical load without flexing.

Not only the cones, but all the metal parts in the Sonja, including the trim rings around the drivers—even the biwiring binding posts—are made by YGA. The aluminum panels that form the enclosures are milled from large sheets of aluminum.

I visited the YGA factory just before writing this review. It's an impressive operation. Raw aluminum-alloy rods, bars, and sheets enter; finished speakers and bags of aluminum swarf and scrap leave, the former to dealers and distributors, the latter to a recycling center. YGA's two expensive CNC machines run flat-out for two shifts every day, even machining away the copper between the traces on the circuit boards for the crossover filters.

Having seen YGA's capital-intensive operation, I am surprised not only that the Sonja 1.3 doesn't cost more than it does, but also that it costs slightly less than the company's earlier flagship model.

Setup & System
YGA's Dick Diamond and Kerry St. James delivered the six aluminum flight cases containing the Sonja 1.3s and, much to my relief, insisted on setting up the speakers in my listening room all by themselves. With each complete speaker weighing 506 lbs, I had been wondering how they were going to move them.

The answer: To place the 1.2 module atop the 1.3 module, then crown the array with the 120-lb 1.1 module, they used the hand-pumped, handled suction cups used to lift large sheets of glass. Then, with the speaker assembled, they slid Teflon furniture-moving coasters, fitted with Delrin inserts, under the four spikes. It proved surprisingly easy to slide the speakers around on these coasters, to find the optimal positions in the room. Once Diamond and St. James had proclaimed themselves comfortable with the setup, they took the suction cups and coasters away with them, meaning that I was able to do no further fine-tuning of the speaker positions.

In view of the rising costs of materials and labor, as well as continuing advances in materials and production, Jones redesigned the Model-1. The result was the slightly smaller, still expensive ($78,000/pair), but equally impressive Reference One, which has been a great success since its debut at the 2006 Consumer Electronics Show. However, even at the time of the Reference One's launch, Jones was planning to design significantly less expensive speakers based on the same technology. He's since taken this in two directions. One is the Compact Reference CR1 ($37,000/pair), a stand-mounted three-way reviewed by John Atkinson in January 2012, and which he resoundingly approved. The other direction was the production of EX series, into which Jones trickled down most of the principles used in the Model-1 and Reference One to a price range within reach of many audiophiles. When I reviewed the top of the EX line, the S-1EX, in March 2007, I felt it was an outstanding speaker: well made, fully satisfying as a reproducer of music, and competitive with speakers costing twice its price of $9000/pair. But there was a catch: The S-1EX wasn't called a TAD.

I recall seeing an advertisement in the New York Times Book Review in the early 1970s that read, in banner print, "Anyone who says that you can't tell a book by its cover has never tried to sell one." The EX line bore the Pioneer badge, not TAD's. Despite the EX models' competitive excellence, it seems that putting a mass-market name on what were clearly high-end speakers discouraged buyers, and the entire EX line seems to be quietly disappearing.

When I saw the Evolution One ($29,800/pair) at the 2012 Consumer Electronics Show, I knew it was the real thing: a low(er)-priced TAD model wearing the proper badge. And in view of my experience with the S-1EX and the many excellent and entertaining show demos of the Reference One that I heard, I bet on the TAD DNA (and Andrew Jones) and jumped at the opportunity to review the E1.

Description
The Evolution One is a slimmed-down version of the Reference One, with which it shares not only its curved side and rear panels, which greatly enhance the cabinet's rigidity and reduce its resonances, but also its use of a very thick front baffle. This extends above the top panel to form a framing arch for the coaxial tweeter-midrange drive-unit, and stops a few inches above the base, thereby creating a wide port to load the woofers. The base contains the complex crossover, isolated from the main structure by compliant mounting. I was surprised to see that the removable underside of the crossover cavity is covered by a heavy, impregnated felt panel. Felt seemed an unusual choice of material, but it effectively damps any resonances. Four substantial, upright, multiway binding posts, suitable for single- or biwiring, rise from the rear skirt of the base.

The drive-units are arrayed vertically, with TAD's signature Coherent Source Transducer (CST) coaxial driver at the top: a 5½" magnesium midrange cone surrounding a 13/8" beryllium-dome tweeter. Magnesium and beryllium are extremely light and stiff; together, these two drivers provide a single-point source for all frequencies from 250Hz to 100kHz, per TAD. Below this driver are two 7" woofers with layered Aramid diaphragms. The coaxial driver is protected from curious fingers by a fixed scrim of what appears to be a fine, open weave of synthetic fiber. The woofers have removable covers.

All of these features are reminiscent of the Reference One and the S-1EX. The tweeter elements are the same in all, but the diaphragm material of the midrange driver is magnesium in the TAD-E1, beryllium in the Reference models. The midrange is also smaller. "Since magnesium is not as stiff and light as beryllium," said Jones, "the cone resonance would be at a lower frequency. By going to a smaller cone, we manage to put the cone breakup frequency back up higher, though not as high as the CR1 and R1 beryllium cone." That said, the Evolution One exudes the same quality of design and construction as its Reference brothers.

Setup
An hour after the Evolution Ones were delivered, Andrew Jones arrived to set them up. He first put them precisely where my main reference speakers sit. Then he hooked up his laptop to the Mytek Stereo192-DSD DAC via an AudioQuest Carbon USB cable. The Mytek's XLR outputs were routed to a Parasound Halo JC 2 BP preamplifier and, in turn, to a Halo A 31 three-channel power amp, bypassing my usual Meridian 861 digital processor. Then came the fun. As we tweaked the positions and toe-in angles of the TAD-E1s, Jones played lots of tracks, many of which would be familiar to those who've attended his entertaining show demos. Finally, as his favorite Boz Scaggs cut, "My Funny Valentine," played over and over, he finished. I hope he was happy with the final setup. I was.

The Model A-1 represents Sound-Lab's best effort at designing a system that appears to have a curved diaphragm. A truly curved diaphragm (à la MartinLogan) can have wide horizontal sound dispersion, but the problem is its limited linearity at low frequencies, where large excursions are required. When the diaphragm is pulled forward, it is stretched, increasing tension. When pulled back, its arc of curvature decreases, which causes it to lose tension. This difference in tension in the two halves of a push-pull cycle severely limits the linear excursion range where distortion will be low. What's a poor designer to do?

Rather than use a diaphragm that is actually curved in the horizontal plane, Sound-Lab opted to use an array, or mosaic, of small, flat-diaphragmed transducer cells. In such an array, a large sheet of Mylar is stretched over a rigid frame which subdivides the sheet into an array of flat cells. The diaphragm is locked onto the outer edges of the frame with a compound precision-molded clamp that assures no slippage of diaphragm tension over time. The cells are then arranged to form several long vertical facets positioned to approximate a curved surface. Sound-Lab's flat diaphragm ensures symmetrical behavior as it moves between the stators. Care is taken to keep the dead zone between facets down to 1/32". Further, to prevent an acoustic "venetian-blind" effect, the angle between adjacent facets is chosen so that the facets should integrate smoothly in the horizontal plane without lobes in the pressure response.

Any taut, underdamped membrane (eg, a drumhead) exhibits a significant fundamental resonance. Such drumhead resonances are undesirable in a speaker; the large peak (most often in the midbass) not only colors the sound, but also limits the speaker's usable dynamic range. The diaphragm can slap the stator electrodes and make nasty crunching sounds that sound little like music. Something must be done to control this resonance. Normally, resistive damping in the form of a cloth mesh or foam placed next to the stators controls diaphragm motion and reduces the Q of the resonance.

Sound-Lab's inventive approach, dubbed the "distributed bass resonance" principle, is protected under US patent. In this design, the diaphragm is physically subdivided into sectors in such a way as to stagger the sectors' resonance frequencies, spreading the bass resonance energy over a greater frequency range. As a result, the Q of the resonance is reduced, but as an added benefit, bass extension is improved. In the loudspeaker's nearfield, the bass response rises at a rate of about 6dB/octave, just enough to compensate for front-to-back dipole cancellation of bass energy for an octave or so below the diaphragm's drumhead resonant frequency. Farther into the room, in the far field, say 10' from the loudspeaker, the back wave wraps around the panel and neatly wipes out the array's inherent rising bass response. The result is a flatter, deeper bass response, marred only by the presence of the ubiquitous room modes.

The interface electronics package is housed in the speaker's base. An ultrasonic type of bias power supply is used, said to eliminate the power-line noise and hum typical of 50/60Hz supplies. A pair of step-up transformers is used to cover the audio spectrum. One transformer (actually two transformers connected in parallel) is primarily dedicated to the bass frequencies, while the other handles the upper range. The transformers overlap in the lower midrange. The interface was redesigned in Fall 1991 to improve the integration between the transformers and to increase the core saturation headroom. Better passive parts are also used. I can tell you from my own listening that an upgrade to the new interface is well worth the expense. Tonal balance in the lower midrange improved, as did midrange transparency. A similar interface upgrade is also available for the Sound-Lab A-3.

Three adjustment controls are provided on the back plate. The Brilliance control pot adjusts the treble level, the maximum position yielding the maximum high-treble sensitivity, counterclockwise rotation reducing the treble response. This control should be adjusted by ear, the optimum setting being a function of listening-room damping and front-end balance.

Control is also provided over the bass response, which can be shelved in 3dB steps relative to the midrange and treble. From the nominal 0dB bass setting, adjustments to +3dB, –3dB, and –6dB are possible by moving a jumper. If you're a bass hog, you might find heaven at the +3dB position. But for most of us, the goal should be to contour the tonal balance so that the bass response at the listening seat sounds right.

Finally, the Bias control allows adjustment of the DC bias voltage to the diaphragm. Think of it as a speaker sensitivity control: the higher the bias voltage, the more sensitive the speaker. That's good; the amplifier headroom requirements are lessened. Because the electrostatic array is tested at much higher voltages than those available at any bias control setting, it's perfectly safe to crank this control fully clockwise for maximum bias voltage. When you do, you'll very likely encounter crackling noise. That's normal, and is indicative of minor static discharges between the diaphragm and stators. Simply back up the Bias control until the crackling stops. After several hours of break-in, it may be possible to advance the pot to a higher bias setting. The ultimate bias setting achievable is a function of the local altitude and humidity. My listening room is 6800' above sea level; after six months, I'm still sitting pretty close to the minimum bias level.

Input connections are via five-way binding posts of average quality: a plastic body with a gold-plated metal core. C'mon, Roger—these speakers deserve much better connectors. WBTs would have been nice.

Planar power
Imagine a cylinder nearly 6' tall with a radius of 19.6". Slice the cylinder into equal quarters, as you might a pie. One of these quarters represents the A-1's radiation pattern. Thus, the A-1's horizontal dispersion is a full 90°; that of its smaller brother, the Model A-3, is only 75° (footnote 1).

More expensive, maybe, but still not expensive in absolute terms: the M350 costs $2498/pair. It features a 1" Extended Wide Bandwidth (EWB) soft-dome tweeter. This has a small dimple in the middle of the dome, and what's claimed to be a larger-than-usual radiating area for a 1" design. The tweeter handles all frequencies above 3kHz and is mounted below a 4.5" midrange driver on the front baffle. The latter has a polypropylene cone, mass-loaded to achieve a smoother frequency response at the top of the driver's passband.

Rather than a single, large woofer to cover the range below 400Hz, the M350 uses a vertical array of four 5.25" polypropylene-cone woofers, these fitted with aluminum shorting rings on their voice-coil formers to lower distortion. Boston's reason for using such an array, which is equivalent in radiating area to a single 9" cone, was that the multiple floor reflections from the units will both minimize the speaker's sensitivity to its position in the room while optimizing low-frequency reproduction. The woofers are reflex-loaded with a flared port, 2" in diameter, on the rear of the cabinet above the single recessed pair of binding posts.

The M350's elegantly proportioned enclosure is made from a material Boston calls Lo-Q: two layers of medium-density fiberboard separated by a thick layer of adhesive, to damp vibrations. A reinforcing brace near the top of the cabinet both increases rigidity and isolates the midrange cone from the woofers, and there are other cross-braces lower in the enclosure. The cabinet sidewalls are beveled at their tops and bottoms and finished in high-gloss black; the front baffle and the top and bottom panels are covered in a black faux leather material. A black, cloth-covered plastic grille is provided; I preferred the look and sound of the M350 without the grille. Four short aluminum pillars lift the enclosure above the pedestal; the result is a visually attractive tower.

The M350 was designed in the US but is manufactured in China.

Nothing's better left unsaid
As promised, the M350s seemed relatively insensitive to where they were placed in the room. On the other hand, no matter how I fine-tuned their positions, I couldn't eliminate an excess of energy in the upper bass. When I listened to the M350s at the speaker's press launch, Boston's Andy Clark was experimenting with foam inserts in the ports. These can be used either to fully block the port or, with the central cylinder of foam removed, to reduce the port's diameter. According to Clark, "in my experience so far with the M350 in a variety of rooms, unless they are in a relatively large room and off the rear wall, the the half-blocked port usually provides the balance I prefer. In really small rooms the full block may be beneficial, but I find myself using the half-blocked port most often."

I experimented with the foam plugs, which Boston includes free of charge with every pair of M350s. Yes, this mod went a long way toward addressing the problem. The double bass in "Killing the Blues," from Alison Krauss and Robert Plant's Raising Sand (24-bit/96kHz ALAC file transcoded from FLAC download, Rounder/HDtracks 11661), remained heavy-sounding but lost its oppressive character. On the other hand, Kurt Sanderling's 1972 traversal of the four Brahms symphonies with the Dresden Staatskapelle (CD, BMG Classics 69220-2), which has a rather lean sound, benefited from the lower-midrange warmth added by leaving the Bostons' ports fully open. I ended up doing much of my listening with the ports unblocked. For example, the synth-bass line in "The Trader," from the Beach Boys'Holland (24/192 needle drop from LP, Brother/Reprise K54008), though balanced a little high in the mix with the ports open, better balanced this recording's forward high frequencies.

And even as I write these words and "Sloop John B," from the HDtracks hi-rez release of the Beach Boys'Pet Sounds (24/192 ALAC files transcoded from FLAC), is playing, I reach for the foam plugs, that wonderful unison combination of plucked double bass and plectrum-played Fender bass sounding way too generous with the ports fully open.

Without the port plugs, the low-frequency warble tones on Editor's Choice (CD, Stereophile STPH016-2) were powerfully reproduced between 125 and 80Hz. The 32Hz tone was reinforced by the lowest mode in my room, but the 25Hz tone was weak and the 20Hz tone inaudible. Played at moderate levels, the low-frequency tones were free from audible distortion, and there was no wind noise coming from the port. With the ports half-closed, the region between 80 and 125Hz was less powerful, but the 63 and 50Hz tones now sounded weak. With the ports half-blocked, the half-step–spaced low-frequency tonebursts on Editor's Choice sounded commendably even, with good weight down to 32Hz. However, when I listened to the cabinet's side and rear walls with a stethoscope as this track played, I could hear a very strong resonance between 256 and 300Hz.

With pink noise, I got the smoothest balance when I sat with my ears level with the tweeters, 34" above the floor, though the overall balance was somewhat dark. When I sat up a little to put my ears on the midrange axes, which are 37.5" from the floor, I now heard an isolated band of presence-region energy. Playing "Whaling Stories," from Procol Harum's Live in Concert with the Edmonton Symphony Orchestra (ALAC ripped from CD, Mobile Fidelity Sound Lab), Gary Brooker's distinctively phlegmy voice (footnote 1) sounded uncolored, though that ascending bass line was not as clearly defined as I'm used to, and the choir sounded a little "hooded." However, it's fair to point out that the M350s suffered the misfortune of following the superbly transparent, astonishingly uncolored YG Acoustics Sonja 1.3 speakers in my room. Of course, the big YGAs, which I reviewed in July, cost more than 40 times the Bostons' price—sonic perfection is to be expected with a speaker having a six-figure price tag. In absolute terms, the M350's shortfall in lower-midrange transparency was not a serious fault.

The M350s' stereo imaging was well defined, if not up to the standard set by minimonitors like the BBC LS3/5a and KEF LS50. The central image with dual-mono pink noise was wider in the midrange than at high frequencies.

I recently downloaded pianist Angela Hewitt's 2008 recording of Bach's The Well-Tempered Clavier (24/44.1 ALAC files; CD, Hyperion CDA67741/4). There is more than four hours of music to come to grips with here, but already I particularly like her performance of the Prelude in e, from Book I. The melodic line in the first half, over that insistent arpeggio figure in the left hand, is almost song-like in its lyricism, while she keeps the lines in the double-speed coda clear, but without forcing the notes into a metronomic straitjacket. The piano's upper registers were clean and uncolored through the M350s, and while I could still hear that cabinet resonance excited if I listened with a stethoscope, I heard nothing untoward at the listening position, no emphasis of some notes but not others, no blurring. It must be noted, however, that this recording is on the dry side, and lacks the sense of space audible from Concert, my two-decades-old live recording of Robert Silverman performing the E-flat major and minor preludes and fugues from Book I (CD, Stereophile STPH005-2). The Bostons successfully presented the piano within the warmly supportive church acoustic, and Silverman is even more lyrical than Hewitt while matching her clarity of line.

The M350's high frequencies sounded smooth, though with a slight excess of energy in the top two octaves. Earlier this year, Stereophile reviewer Erick Lichte asked me to prepare the master from the original 24/88.2 files for a new CD, A Drop in the Ocean, by the Portland State Chamber Choir, conducted by Ethan Sperry. (Erick, one of the singers, had co-produced and edited the recording.) The first work on this CD, O Salutaris Hostia, by the young Latvian composer Eriks Esenvalds, features a duet for two sopranos and is one of the most beautiful pieces of choral music I have heard. The two women float an interwoven tapestry of twin vocal lines over a slow-moving chorale in the lower voices. Through the Bostons, the sopranos sounded pure and uncolored; perhaps a touch forward in the mid-treble, but without excessive sibilance, and with all the singers set within the luscious acoustic of St. Stephen's Catholic Church, in Portland, Oregon.

I finished my auditioning of the Bostons with Sibelius's Symphony 2, with Sir John Barbirolli conducting the Royal Philharmonic (ALAC files ripped from CD, Chesky Gold Series CG903). This 1962 recording, recorded by the dream team of Chuck Gerhardt and Ken Wilkinson for The Reader's Digest, played to the strengths of the Boston M350: the darkish balance reduced the audibility of the tape hiss; the warm lower mids added to the sense of authority of the lower strings, even with the ports half-closed; and the well-defined stereo imaging allowed the space of London's Walthamstow Town Hall to be readily apparent around the pizzicato strings.

Life is like a beanstalk . . . isn't it?
Whether or not Boston Acoustics' M350 will be a worthwhile purchase will depend very much on the speaker's excessive upper-bass energy not being a problem in the prospective owner's room and system. Other than that, the M350 offers a neutral tonal balance and is commendably free from overt coloration. Its only real failing is the congestion in the lower midrange, but this only occasionally got in the way of my enjoyment of the music, and must be put in the context of the M350's price.

I had a good time with these speakers; I predict that you would, too. Recommended.

When high fidelity took off during the early '50s, consumer loudspeaker systems were nothing more than cheaper, scaled-down versions of those professional systems, with much the same sonic flavor but even rougher highs. Then, in 1956, Edgar Villchur patented a new kind of speaker system designed specifically for consumer use.

Recognizing the fact that, in the home, bass extension was more important than efficiency, Villchur's "acoustic suspension" system achieved tremendous LF extension by (among other things) reducing mid- and upper-range sensitivity to match the system's bass sensitivity. By comparison with professional speakers, Villchur's AR-1 had a muted, almost withdrawn middle and upper range, which was far more pleasant to listen to in small rooms. It also launched high fidelity loudspeaker design on a tangent from "pro" sound, and independent evolution in those divergent directions has made the two increasingly different since then. While professional systems grew ever more efficient and forward, so-called audiophile speakers became more laid-back, polite. (Our Bill Sommerwerck dubbed this kind of sound "Boston bland," in honor of its place of origin.) The greater the schism, the more the audiophile community sneered at the "brashness" of professional speaker systems (footnote 2).

But that brashness has proven to be exactly what's needed for music normally heard at very high levels, such as hard rock. The result has been the evolution of two distinctly different kinds of "audiophile" speaker systems: highly efficient, up-front systems for rock enthusiasts, and the laid-back, "polite" systems with exaggerated depth and spaciousness favored by many classical listeners and most audio perfectionists.

Enter JBL
One of the leading manufacturers of professional speakers during the '40s was a firm called Altec Lansing. The Lansing part of the name belonged to a young engineer named James B., who parted company with Altec to form his own firm, JBL. Through the years, JBL has always tried for high efficiency and a forward, gutsy sound in their consumer line, thus earning the undying scorn of all audio perfectionists, who consider "JBL" synonymous with sonic trash. This never bothered JBL as long as "high end" was a mental aberration afflicting only a miniscule part of the population. The masses preferred JBL's kind of sound.

But two developments in the past three years have prompted JBL to reconsider their public image. First, the masses discovered the snob appeal of the High End, and learned from the purists that liking the sound of JBL speakers was—well, it just wasn't done! And second, JBL recognized that digital recordings, whose lack of surface noise is an invitation to play them at high volume, put their own products at a decided advantage over those of most audiophile manufacturers. JBLs could play loudly without burning out; most audiophile loudspeakers can't. The Ti Series of loudspeakers is JBL's bid for a new image—respectability.

The Ti in the 250Ti's model designation stands for titanium; some techniques JBL developed for working with this tricky metal made their new tweeters possible. Titanium has an extremely high mass-to-stiffness ratio. A very thin sheet of it has enough stiffness to behave very much like the ideal piston radiator, with its entire surface area moving in unison. But its stiffness also makes it very brittle, and likely to split when subjected to the kind of forming processes necessary to produce an effective radiating surface. JBL claims to have found a solution, and has even found a way of embossing a pattern of diamond-shaped ribs into the surface of the tweeter dome to increase its stiffness. The result is the new tweeter which shows up on the high end of all their new speaker systems, the 440Ti.

The cone drivers in the 250Ti are described in JBL's literature as "evolutionary rather than revolutionary," which is to say, they're the latest improved versions of earlier designs. But it's obvious that JBL has done their homework.

For example, this is the first time JBL has used polypropylene as a cone material for their midrange drivers, even though most high-end speaker manufacturers have favored it for some years. But JBL's is not the usual polyprope material; it's "doped" with an additive that increases its stiffness and reduces breakup. (One of the recognized shortcomings of polypropylene is that it is too flexible for its own good; its popularity stems mainly from its high internal damping, which minimizes colorations due to resonance.)

These are also the first speakers from JBL whose large crossover capacitors are bypassed by small ones, reducing the effects of dielectric absorption. This is strict perfectionism, but JBL ascertained to their own satisfaction that bypassing did, in fact, improve the sound. The other characteristics of the 250Ti—the very high power-handing ability, the long-throw woofer (5/8" displacement within 10% linearity!), and the rigid enclosure construction—are pretty much old hat; they've been earmarks of JBL products for many years.

Sonics
I have never much cared for most aspects of the traditional JBL sound: boomy midbass, shallow soundstage, vague imaging, lack of deep bass or really high highs, and (often) piercingly shrill middle highs. But I have always admired their midrange performance. Whether it was their professional or consumer lines, JBL's speakers always had a punchiness and detailed immediacy, an ability to make a voice or a solo instrument sound right in the room, that has not been equalled by any audiophile speakers I know of. "Wouldn't it be great," I thought, "if JBL has retained that middle-range performance, and just augmented it with state-of-the-art performance in those other areas?" I should have known better!

JBL did their homework, all right. In their bid for the perfectionist audiophile market (and who else would pay $3400/pair for loudspeakers?), JBL has produced nothing more than yet another high-priced "Boston bland" behemoth. Every vestige of the middle-range immediacy, aliveness, and incredible detailing that characterized their previous speakers is gone. Instead, what we have are outstanding lows, superb highs, excellent soundstaging, very good imaging, and a total inability to make anything sound real. It's another instance of the baby going down the drain with the bath water.

Perhaps I'm being too hard on the 250Ti. I acknowledge that it is capable of producing prodigiously high listening levels without a trace of strain, and that its highs are simply gorgeous!—silky smooth, open, completely effortless, and free from steeliness, even at the highest listening levels any sane person would tolerate (footnote 3). I acknowledge its remarkably smooth low end (usable to around 35Hz), and its ability to float a wide, deep soundstage around the instruments in simply-miked recordings.

But I also admit that my unusually brusque dismissal of JBL's not-inconsiderable design efforts on behalf of this system stems more from frustrated expectations than from anything else. I expected a system that would combine the old JBL systems' best attributes with those of today's state-of-the-art audiophile systems. I did not expect a total sellout to the audiophile taste for unctuous blahh!

I am, however, constantly reminded by everyone here that my stubborn insistence that reproduced music should sound alive is not shared by everyone—nor, perhaps, by most audiophiles. After all, it wasn't the terminally deaf that JBL surveyed prior to designing this system, but the people who apply perfectionist standards to reproduced sound. So, for those perfectionists among our readers who don't give a sow's teat about aliveness, here is an impassive description of how the 250Ti's sound when driven by the kind of amp they like best: a top-notch solid-state amp like the Electron Kinetics Eagle 2.

I got the most natural sound over-all with the midrange drivers at their maximum setting and the tweeters strapped for 1dB of attenuation. The highs, as I said, are just superb—better in smoothness and freedom from steeliness than those of many electrostatic systems, and in perfect balance (in my listening room) with that 1dB of attenuation.

In general, the sound of the system is warm, a little laid-back, and a shade heavy and loose through the midbass. Bass detail is good but not excellent, having only moderately good delineation of pitch. The speakers' sound varies markedly according to the vertical angle of your ears relative to the midrange drivers, and is most neutral with the ears almost exactly on the axis of the lower-midrange driver. Below that, the sound becomes even more laid-back; above, a pronounced dip develops in the upper middle range. The 250Ti throws a very wide, deep soundstage, with stable but not very specific imaging. Mono sources produce a vague ear pressure suggesting substantial random-phase content, and center bunching is not very tight.

Summing Up
Overall, this is a very pleasant-sounding system that can produce some very impressive sounds, but it lacks the feeling of life that makes the difference between excellent reproduction and literally accurate reproduction. For $3400/pair, I expected more.—J. Gordon Holt

The first of these reasons: The three-way 804 Diamond takes up no more floor space than the two-way 805 while also having two 6.5" woofers, which greatly expand its power handling and bass extension. In fact, it was my wife who, when I consulted her about choice and placement, asked why one would even consider the smaller speaker when the bigger one took up no more floor space and looked better. Second, in the 804, the same 6" Kevlar-cone mid/woofer used in the 805 is relieved of all bass responsibilities and works purely in the midrange, as it does in the 800 Diamond.

The third reason—perhaps a corollary of the second—is that dedicating the Kevlar driver to the midrange means that it can be used with B&W's proprietary Fixed Suspension Transducer (FST) technology, which was designed to better control and define the breakup patterns of relatively nonrigid diaphragms. The use of Kevlar in the 805's mid/woofer demands a more traditional surround that imposes the uncontrolled radial breakup patterns that the use of a Kevlar diaphragm was intended to avoid. Note, also, that B&W uses diaphragms of Rohacell, which is stiffer than Kevlar, in the woofers (ie, not mid/woofers) of all its 800-series models, including the 804.

Finally, although the 804 Diamond is not endowed with B&W's iconic Marlan head, as in the 800 Diamond, the 804's midrange driver is still enclosed in an internally tapered enclosure, and scores over the larger but similarly configured 803 Diamond in having a narrower cabinet, which, potentially, would not have as great an impact on midrange dispersion. It might also be effective in minimizing the kink in the horizontal off-axis radiation pattern that John Atkinson has discovered in the transition from the woofers to the midrange drivers of the Marlan-topped 802 and 800 Diamonds.

Arrival The 804 Diamonds arrived in substantial packaging that lacked the forklift-accessible plinths needed for the 800 Diamonds. I was grateful for the unpacking instructions printed on the outside of the box; I was able to unpack and set them up with little effort and no assistance. Both spikes (for carpet) and plastic feet (for hardwood floors) are provided. I chose the latter. The slim, graceful cabinet is oval in cross section, except for the flat front panel; my review samples were finished in Rosenut veneer. (Cherry and Piano Gloss Black are also available.) A black front grille attaches magnetically—when the grille is removed, no securing devices mar the speaker's appearance.

On top of the cabinet, lying in a shallow niche, is B&W's iconic enclosure for its diamond tweeter and its tapered tube. Below that is the yellow, woven-Kevlar diaphragm of the FST midrange driver, which is installed through the front panel and secured by a shaft to a compliant support in the rear, just as the FST midranges in the 802 and 800 Diamond models. (A plastic disc must be removed from the rear of the midrange enclosure, which opens to the speaker's rear, before listening.) Below that are the two Rohacell-cone woofers and a low-turbulence port, dimpled and flared, and similar to the one hidden on the underside of the bigger models. Protruding from near the top of the rear of the cabinet is the adjustable mount for the midrange; near the bottom are two pairs of speaker binding posts of a new design that accommodates easy tightening by hand. Biwiring and biamping are thus made possible; jumpers are also provided. To meet EU requirements, the center bore of each binding post is occupied by a plastic plug; I removed these in order to use cables terminated with banana plugs.

Set-up
I was able to lift and lower each 60-lb 804 Diamond into position by using its bass port as a grip. At first, I set them up in the precise spots just vacated by my 800 Diamonds. In these positions, the 804Ds seemed to sound somewhat thin and bright, but some expectation bias is inevitable: I was consciously aware, from both sight and aching muscles—I'd just moved the big 800Ds out of the way—that I had just replaced two very familiar, very large speakers with a pair of small towers fresh from the farm. Apparently, I adapted as I adjusted the setup. The 804Ds ended up about a foot closer to each other than where they started out, toed in by no more than 10°.

Comparisons can be odious
That doesn't mean that comparisons smell bad, but making direct comparisons can lead to various problems. In the case of speakers, it's all too easy to describe how one speaker sounds different from another, but 1) that doesn't tell us a lot about which speaker might be better or more accurate, and 2) a specific character of one speaker might constrain an accurate description of another. So, although I can't ignore comparing the 804 Diamond to its 804S forebear or to its big brother, the 800 Diamond, both of which I had on hand, I'll first describe the 804D's sound as I found it, independent of comparisons.

I listened casually to the 804 Diamonds for a couple of weeks before sitting down to do more careful listening. During that time they evinced good tonal balance and great stereo center fill. FM broadcasts sounded good, with no emphasis of hiss with weak signals, and Internet radio sounded fine without obvious dulling due to the limited bandwidth. As I sat down with my favorite discs and downloads, my expectations were rising.

The 804 Diamond was quite beyond criticism in the treble, with clarity and fine detail. Cymbals and triangles sizzled and tingled appropriately, but, more important, E-string fiddling was sweet and pure. Voices, too, were lifelike, whether solo or in chorus. One of my new favorite vocal recordings is a 24-bit/192kHz PCM download of Marianne Beate Kielland's recital disc Come Away, Death (SACD/CD, 2L 2L-064-SACD). Following a tonally convincing introduction by pianist Sergei Osadchuk, Kielland's silken mezzo-soprano appeared eerily between the 804 Diamonds with such presence that I got up to check that my (presumably) idle center speaker was, in fact, silent. The effect expanded with multiple voices—as on Dixit Dominus, a disc pairing Handel's and Vivaldi's settings of Psalm 109, with David Bates leading La Nuova Musica (SACD/CD, Harmonia Mundi HMU 807587): the choristers' voices were arrayed in space between and above the 804 Diamonds.

I was impressed by the SS-AR2. Somewhat mellow-balanced, it had a superlatively natural way with voices; clean, extended low-bass frequencies; and offered a transparent window on the recorded soundstage. So when, at the October 2012 Rocky Mountain Audio Fest, I first heard Sony's new SS-NA2ES speaker—which, at $10,000/pair, costs only half as much as the SS-AR2—I put it on my "to review" list.

The SS-NA2ES
The SS-NA2ES—"NA" stands for "Natural Acoustics"—appears very similar to the SS-AR2: a beautifully finished three-way tower, approximately 3' tall, with drive-units custom-made by Scan-Speak in Denmark and designed in collaboration with Sony: twin 6.5" aluminum-cone woofers and a 5" midrange unit with what looks like the same spiral radial cuts in its paper cone. (These cuts control the cone's out-of-band behavior to give a smooth low-pass rolloff above the crossover frequency.) The 'NA2 is actually a little more than 2" shorter than the 'AR2, however, and while those spiral cuts in the 'AR2's midrange were real, and subsequently filled with sealant in a time-consuming and expensive manual process, pressed into the new speaker's midrange cone are spiral radial grooves. (The drive-unit is exclusive to Sony.) The pressed grooves are not quite as effective as the filled cuts, but they allow Sony to keep the SS-NA2ES's bill of materials down.

The biggest external difference concerns the tweeter. Or tweeters. The 'AR2 had a conventional if high-quality, 1" silk-dome tweeter with a motor that featured a circular array of six neodymium magnets that left the rear of its diaphragm unobstructed. The 'NA2ES also uses a 1" silk-dome tweeter sourced from Scan-Speak, but mounts it vertically between two 0.75" fabric-dome "assist" tweeters. All three tweeters share the same faceplate and cover the same frequency range. I asked Sony's designer, Yoshiyuki Kaku, what he was trying to achieve with this three-tweeter array, which Sony calls the I-Array.

Mr. Kaku explained that he'd been listening to a Sony "lifestyle" speaker, the Sountina, that had been developed by another group within the company. The Sountina had an unusual cylindrical tweeter with a 360° radiation pattern. Though the Sountina was not a "hi-fi" design, the sound of its tweeter was remarkably lifelike, and very different from that offered by normal directional dome tweeters. That experience set Kaku thinking about how he might get the same result from a conventional speaker, to widen its radiation pattern at high frequencies to better match that in the midrange and bass. The target was therefore to design a tweeter with a directivity similar to that of the midrange driver and woofers. Simply making the tweeter's dome smaller would increase the top-octave dispersion but compromise power handling. The solution turned out to be a vertical array of a single 1" tweeter flanked by two smaller-diameter tweeters, these having wider dispersion above 10kHz but slightly lower sensitivity. The benefit would be more top-octave energy in-room without affecting the flatness of the on-axis response.

I asked why this wouldn't introduce severe lobing in the vertical plane in the frequency region where the outputs of all three tweeters overlapped.

The answer: To ensure that the array would act as a single wide-dispersion tweeter without compromising the performance in the vertical plane, the distance between the main tweeter and the two assist tweeters had to be optimized, in conjunction with the shape of the frame and the depth of each diaphragm. A chance meeting in a local Starbucks with a colleague in Sony's R&D department proved fortuitous. The researcher had actually been working on ways of reducing fan noise in laptops, but was intrigued by the problem. He developed computer simulations of how spaced tweeters could work together without unwanted peaks and dips off axis. The solution involved making the shared front plate slightly concave so that the central dome would be recessed a little compared with the assist domes. (Sony has applied for a patent for this.) In conjunction with the spacing of the three tweeters, making the crossover point for the assist tweeters a little higher than that of the primary tweeter, and the assist tweeters' lower sensitivity, this would eliminate lobing in the vertical plane.

The tweeters' front plate is rabbeted into the top of the front baffle to give a smooth acoustic environment for the tweeter diaphragms. Like the SS-AR2, the 'NA2ES's enclosure is not designed to be completely inert; instead, its vibrational behavior is controlled in a manner consonant with the music. Like the 'AR2, the cabinet is made of Scandinavian birch ply, which has a dense grain structure. However, while the SS-AR2's enclosure has gracefully curved walls, the SS-NA2ES's cabinet is straight-sided, again to lower cost. Whereas the 'AR2's baffle was maple, the 'NA2ES's is also Scandinavian birch ply, 1.375" thick, with chamfered vertical edges to control lateral diffraction. Internal birch bracing is used to provide strategic stiffening. Although the wood is prepared in China, final machining and assembly of the enclosure, and overall manufacture, are performed at Sony's factory in Japan. The dark-brown gloss finish allows the wood grain to show through; the result is an elegant tower.

The twin woofers are reflex loaded with an offset, slightly flared port near the base of the rear panel. The midrange driver is mounted in its own internal chamber, isolated from the woofers' back pressure, and is vented with a small port just below the top of the rear panel. Electrical connection is via a single pair of high-quality binding posts on the rear panel, just above the port that loads the woofers. The speaker is supported by four small cones screwed into the base.

Listening
Like the SS-AR2s before them, the 'NA2ESes were set up in my listening room by Yoshiyuki Kaku, assisted by Sony Electronics' Motoyuki "Yuki" Sugiura. While this ensured that the speakers were optimally set up, I also continue to learn about speaker setup from observing others do it in a room that I know so well. I tried the SS-NA2ES with all three pairs of solid-state monoblocks I had on hand: Classé CTM-600s, MBL 9007 References, and Lamm M1.2 References (the Lamm's driver circuit includes a single small-signal tube). Like the YG Acoustics Sonja 1.3 loudspeaker, which I reviewed in July, the Sony proved fussy about which amplifier brought out its optimal performance.

The Vittora is designed and built by Greg Roberts (footnote 1), a longtime audio enthusiast who bought his first pair of Klipsch La Scala loudspeakers when he was 14. (He has owned a number of pairs of Klipschorns in the years since, having settled on an especially nice-looking set from 1967.) A woodworker by training and a homebuilder by trade, Roberts began, in 2001, to offer his services as a commercial restorer of Klipsch's "heritage" products: the Klipschorn, the Belle Klipsch, the La Scala, and the Heresy. In time, restoration turned to modification, as Roberts developed a midrange horn and other components to improve the performance of classic Klipsches that hadn't always been built to perfectionist standards. Not long after that, Roberts decided to incorporate what he'd learned into a completely new, if unabashedly Klipschian, loudspeaker of his own design. Thus, over several years, did Volti Audio and the Vittora loudspeaker come into being.

The Vittora is a three-way, fully horn-loaded loudspeaker in two enclosures per channel, both made entirely of Baltic birch plywood. The bass cabinet is a single-fold bifurcated horn in which a rear-facing woofer fires into a splitter that, according to Roberts, took considerable time to develop—as did the shape of the bass horn: "The size of the mouth is a big determinant for lower bass: It is what it is," he says. "But I found that upper bass was something I had control over, and I used cheap OSB [oriented-strand board] to build multiple prototypes." The result is a design in which the sides of the enclosure—and thus one surface along each of two paths—are curved. The high-sensitivity, 15" bass driver has a stiffly suspended paper cone—Roberts estimates its Q as approximately 0.3—with a free-air resonance in the neighborhood of 40Hz.

The cabinet's curve is repeated in the sides of the upper enclosure, which houses the Vittora's midrange and treble horns, both of which are derived from Roberts's modifications of Klipschorns. The rectangular midrange horn, made of plywood and bendable hardwood, has a tractrix flare, and is driven by a 2" compression driver (a BMS 4592) with a phenolic diaphragm. The elliptical treble horn is made of composite and is driven by a 1" compression driver with an aluminum diaphragm. The two horns fit side by side, the latter secured in an opening that Roberts designed into the former.

The upper enclosure also contains the Vittora's crossover network, which is user-adjustable for treble output: By substituting different preassembled resistor modules—which work within the Vittora's capacitor- and autoformer-based network to create different L-pad configurations—the owner can suit room or taste by raising or lowering the tweeter's output across its operating range, from 6kHz up. The crossover network is accessed through a panel on the back of the upper cabinet, and the resistor modules are connected with integral gold-plated spade lugs, making soldering unnecessary. Roberts says that the bass portion of the crossover also includes an adjustable contour filter—a notch filter, really—that helps flatten out a known response peak.

The Volti Vittora is built in a shop—as opposed to a garage, a driveway, or somebody's mother's basement—solely dedicated to the production of loudspeakers and loudspeaker components, and which Roberts has equipped with state-of-the-art power tools and an air-filtration system. Cutting and shaping are done with high-tech European table saws and bandsaws. Wooden parts are bent to shape in a vacuum-bag system—also used to apply veneer—and catalyzed polymer finishes are applied in a separate, room-sized spraybooth. The build quality of my review pair, finished in bosse cedar, equals that of the finest American loudspeaker cabinetry I've seen, DeVore Fidelity and Thiel Audio included. Roberts makes his own wooden cabinet feet, and even irons and applies his own vintage-style grillework—it all contributes to one of the best-built audio products I've had in my home. Forgive the ham-fisted cliché, but even my wife, who was at first put off by the idea of a speaker that takes up more space than a front-loading clothes-dryer, was impressed.

It rained. Of course.
Janet was also impressed with Vittora's sound, going so far as to call it the best horn speaker she's heard. But that's getting ahead of myself—before any listening got done, Greg Roberts and I had to get the Vittoras through the door, which meant that we uncrated them in my driveway. It rained. Of course.

The crates themselves were well made, each containing a single channel's bottom and top enclosures, separated from one another with sheets of sturdy foam. Carrying inside the 60-lb top enclosures wasn't too terrible, but the 127-lb bottom enclosures gave us a spot of trouble on the way up to my porch, especially as the enclosed stairway is 31" wide and the uncrated enclosure's depth (its smallest dimension) is 29". A few knuckles were scraped that day, a few curses cursed.

Once inside, the setting-up was fairly easy. Roberts shares my preference for using felt pads on the bottoms of his loudspeaker feet (provisions exist for those who endure in preferring spikes), so the heavy lower enclosures were easy to slide on my hardwood floors. The upper enclosures are fitted with spikes, the points of which correspond with dimpled discs atop the bottom cabinets; fitting together the two enclosures is a two-person job, but neither person need be terribly clever or strong, merely possessed of good depth perception. (I scarcely filled the bill.) In order for me to have the complete Vittora experience, Roberts also brought a matching sample of its optional subwoofer ($2900 without its corresponding Marchand amplifier/crossover), beautifully finished in the same bosse-cedar veneer. That said, we began our fine-tuning and our first few hours of listening without it.

The Vittoras
Bass extension with the Vittoras alone (Greg Roberts says they reach down to 50Hz, in-room) was superb from the get-go: The bass horn loaded the room exceptionally well, with no egregious dead zones. Our work was confined to selecting the optimal distances between the speakers and the front and side walls; we noted, without surprise, that when those two dimensions were too similar, bass notes lost some of their distinctness of pitch and clarity. Our best results were had with the cabinets only a few inches from their respective sidewalls, and with about 26" between the back of each cabinet and the wall behind it. A gentle to moderate amount of toe-in was preferred, the handed enclosures arranged so that their treble horns were on the outside edges of the midrange horns.

At the far end of my room, driven by the 25W Shindo Corton-Charlemagne amplifiers, the Vittoras sounded nothing short of wonderful. Their trebles were smoother and altogether softer than those of my metal-horned Altec Valencias, while their bass range had the same touchtone, vintage magic: a little less sharp and a little more colorful than the Altecs, and just as big, just as full of impact and nuance and feeling.

Design
The three-way Diamond 10.7 has a 1" (25mm) soft-dome tweeter with a neodymium magnet, mounted in a cast-alloy surround and covered by a metal diffusion grid designed to iron out high-frequency perturbations for a smoother treble response. In addition to its 2" (50mm) midrange dome, the Diamond 10.7 has two 6.5" (165mm) woofers with woven Kevlar cones, the weave's diamond pattern continuing into the surrounds to damp standing waves. The upper woofer has a polished phase plug in the center of the cone; the lower woofer, which is rolled off above 150Hz, has an inverted dustcap. The cabinet has curved sidewalls, and a front baffle made from a composite material finished in piano black, on which are mounted the drivers. The biwirable 10.7 has a rear-firing reflex port. In the US, the speaker is available in Blackwood, Cinnamon Cherry, or Quilted Rosewood. The Quilted Rosewood of my review samples was quite attractive.

I listened to the 10.7s with and without their grilles and found the tonal balance unchanged. Without the grilles, however, I heard a slight increase in transparency and resolution of detail, so I left them off for most of my listening.

Listening
Well-recorded voices on original pressings of vintage LPs enabled the Wharfedale Diamond 10.7 to show off an uncolored midrange. In "The Night They Drove Old Dixie Down," from Joan Baez's Hits/Greatest & Others (LP, Vanguard VSD 79332), her voice was reverberant and dimensional. Similarly, Tony Bennett's rendition of "Autumn in Rome," from his This Is All I Ask (LP, Columbia CS 8856), was reproduced with the master's voice in all its silky, voluptuous glory. Higher in the audioband, the Wharfedale reproduced well-recorded pianos with clarity and extension and no trace of coloration. The upper register of Anat Fort's piano, in her A Long Story (CD, ECM 1994), was reproduced with a great deal of sparkle and air. However, I've heard other speakers reproduce Steve Nelson's vibraphone solo in Jerome Harris's arrangement of Duke Ellington's "The Mooche," from Editor's Choice (CD, Stereophile STPH016-2), with more sparkle.

The Diamond 10.7's bass performance was natural and extended, particularly with jazz. In Anat Fort's A Long Story, Ed Schuller's double bass was warm, deep, and rich on all tracks, with no coloration. I also listened to rough mixes made by John Atkinson of a recent recording of a classical piece I wrote for my quartet Attention Screen, "Recessional," which I performed on the Greenlaw Memorial Pipe Organ of The Community Church of Douglaston, Queens. The recording has considerable energy below 50Hz, and the Wharfedales reproduced the most difficult passages forcefully, with no sense of strain, distortion, or rolloff. The fortissimo passages moved a lot of air, and the sound of the organ was reproduced with the realism of a live performance. I look forward to reading JA's technical analysis of the Wharfedale's bass extension.

Piano recordings revealed the speaker's excellent ability to articulate transients. Ahmad Jamal's solo passages in the title track of his At the Pershing: But Not for Me (LP, Argo 628) were clean, lightning-fast, and with no hint of smearing. On the other end of the Gershwin spectrum, Earl Wild's reading of Rhapsody in Blue, with Arthur Fiedler conducting the Boston Pops (LP, RCA Living Stereo LSC-2367), was light and airy in his rapid-fire upper-register passages.

Percussion recordings showcased the Diamond 10.7's super rendering of low-level dynamic phrasing. I was able to follow every detail of drummer Joe Morello's famous solo in "Take Five," from Dave Brubeck's Greatest Hits (LP, Columbia CS 9284)—his technique breathed in a linear organic fashion, as in a live performance. Similarly, Jack DeJohnette's busy but delicate drumming during the ensemble passages of "Third World Anthem," from his Selected Recordings (CD, rarumECM 8012), covers a broad range of percussive textures in the range of ppp to mf, and the Wharfedale captured them all. However, I've heard other speakers render the high-level dynamic passages in this track with a bit more slam and drama. With rock recordings played loudly, however, the Diamond 10.7s were capable of an impressive level of high-level dynamic realism. When I cranked up Roger McGuinn's Cardiff Rose (LP, Columbia 34154) fairly high, every track had rock-'em, sock-'em dynamic slam—the drums and bass guitar were punchy and forceful.

The Wharfedale's ability to render detail made it very easy to hear differences in recording quality. I recently acquired original pressings of early recordings by Herb Alpert and the Tijuana Brass, and was floored by the high quality of their sound through the Diamond 10.7s. As I listened to every track of Whipped Cream & Other Delights (LP, A&M SP4110), I was smitten by the perfectly crisp detail and transients, the warm midrange timbres, the clean, clear bass. I'd always thought this music was cheesy, but the Wharfedale let me analyze the high quality of the compositions, arrangements, and musicianship on these recordings. The music was so involving that I began to reflect on how the then-unknown Herb Alpert had started A&M Records with his friend Jerry Moss, with no roster other than his own group and the Baja Marimba Band, and was able to build it into a recording powerhouse that, 10 years later, they sold to PolyGram for $500 million. Though Alpert still performs, he now dedicates much of his energy to musical philanthropy.

The album that brought together all of the Wharfedale's strengths was my original British pressing of King Crimson's Larks'Tongues in Aspic (LP, Island ILPS 9230). Every timbre in "Easy Money" was pristine and natural, with layers of detail and dynamic subtleties revealed. Each transient was crisp and clean, with no artificial sharpness or harshness. My listening notes read: "Drama! Detail! Clarity!"

Comparing
I compared the Wharfedale Diamond 10.7 ($1299) with the Epos M5i ($899 when last offered), the Dynaudio Excite X12 ($1200), and the Monitor RX6 Silver ($1250). (All prices per pair.)

The Epos M5i had more detailed high frequencies than the Wharfedale and even better low-level dynamics, but the Diamond 10.7's lower midrange was richer. The Epos's midbass sounded cleaner, but the Wharfedale's bass extended a bit deeper. Higher-level dynamics were slightly better through the Diamond 10.7.

The Dynaudio Excite X12's midbass was warmer than the Epos M5i's but not as warm as the Diamond 10.7's. However, the Dynaudio's highs were more involving and more delicate than the Wharfedale's, and its midrange and high-frequency detail were superior. The Excite X12's and Diamond 10.7's high-level dynamics were about equal.

The Monitor RX6 Silver's highs were more detailed and extended than the Wharfedale's, but the Monitor also had the deepest, cleanest bass, and the best high-level dynamics, of all four speakers.

Summing Up
I've always found that Wharfedale's Diamond bookshelf speakers excel at providing excellent sound per dollar, and can now say that that extends to their more expensive floorstanding models. The Diamond 10.7 is an impressive speaker that provides many of the attributes of pricier floorstanders in an attractive, small-footprint cabinet at an accessible price. It should be on the short list of anyone shopping for something at or near the competitive price point of $1299/pair.

The new Venere line represents Sonus Faber's effort to bring all of this to a lower-priced segment of the market currently dominated by rectangular boxes with OEM drivers. In what's become common practice, SF does all their design and engineering at their home base in Arcugnano, Italy, but has their new speakers made in China. The result are models that retain SF's traditional Italian styling and technology, but are available at prices that more audiophiles and music lovers can afford.

When I first saw the Veneres, at the 2013 Consumer Electronics Show, I was willing to accept that they might comprise SF's lower-priced range, but I was still amazed at what the prices actually are. The subject of this review, the Venere 2.5 ($2498/pair), is the first of a line that includes a larger floorstander, the 3.0; the 2.0 and 1.5 stand-mounted models: the Wall, a wall-mounted two-way with a passive radiator; and the Center, a center-channel speaker.

Description
As its name implies, the floorstanding Venere 2.5 is a 2.5-way design with Sonus Faber's own coated-fabric tweeter (which is not cooled with ferrofluid), and two 7" woofers, also by SF, the lower rolled off above 250Hz. The drivers are inset into a gently curved front baffle and protected by a stiff but lightweight black grille, magnetically attached. A cutout at the bottom of the grille reveals the Sonus Faber logo engraved on the baffle; below that, at the bottom, is the semirectangular slot port. The side panels curve gracefully around to meet a very narrow rear panel just wide enough to accommodate a staggered array of four multiway terminals. The cabinet sits on a thick glass base plate that accepts four substantial spikes, the front pair nearly an inch longer than the rear, which tilts the entire cabinet back to align the tweeter with the deeper woofer diaphragms. This rake is continued in the top panel, which has a glass inlay and from the front slopes up toward the rear of the speaker, strikingly tapering to a near point. The effect, in the sleek Black Lacquer finish of the review pair, was of an elegantly dressed gentleman in provocative slouch. Certainly, no one will be tempted to place a cocktail glass on it.

My Connecticut system was channeled through the Marantz AV-8801 A/V pre-pro, which includes Audyssey's MultEQ XT32 room-optimization software. However, I couldn't assume that such equalization would be available to all readers, so I listened to the Venere 2.5s either via a direct analog route or, with multichannel sources, with Audyssey set to bypass the front L/R speakers (the Sonus Fabers). I also spent some time listening to the Veneres via a Trinnov MC Optimizer, in both stereo and multichannel.

Setup
I connected the 2.5s with the biwire jumpers in place and proceeded to position them. This was somewhat impeded by the necessity to keep the pads (provided by SF) under the spikes to protect my hardwood floor. I ended up with the Veneres in roughly the same places as my regular speakers, but with a toe-in that aimed their tweeters directly at the main listening position. Not all speakers work well in this orientation; usually, they require less toe-in, to avoid the tinge of HF brightness that often accompanies on-axis listening. I suspect that has to do with such speakers having a mild upper-midrange depression or peak, and that aiming them off axis mitigates the MF/HF imbalance.

Since I used the Venere 2.5s while reviewing the Trinnov Optimizer for my column in the September 2013 issue, I can say that the uncorrected on-axis response in my room did show a mild 1–2dB emphasis from 300 to 1000Hz, followed by a similarly mild 1–2dB trough from 1.5 to 8kHz. This wasn't remarkably flat, but the Venere 2.5s matched within 1dB, and gave the subjective impressions of great midrange clarity and ideal balance when listened to on axis. Below 300Hz, room modes dominated the frequency response, as the two speakers' different relations with the room boundaries imposed a much more widely differing response.

Perhaps due to the pair's close match throughout the midrange and treble, I was immediately struck by the spaciousness of the Veneres' rendering of stereo orchestral recordings; despite the notable midrange and treble detail, the overall balance was warm, with a soundstage that began just behind the speaker plane. Bass was adequately extended, but, with no room EQ, I was aware of excessive midbass emphasis. As I said, I don't ascribe this only to the Veneres, but to my room setup as well. Nonetheless, it was a balance that many would find satisfying, and reminded me of the slightly rising bass response and mildly attenuated treble incorporated into many "house curves."

But what about speakers that sound good at audio shows? Well, a speaker can't sound better than its inherent mechanical/electrical characteristics allow, but the rest of the system may have been hand-picked to be the most synergistic possible, and include ultra-expensive components. The setup may have been tweaked to a fare-thee-well, and the demo material chosen to show off the speaker's positive characteristics and minimize its deficiencies. Based on listening under these conditions, you may be sufficiently impressed to buy the speaker, then disappointed when you can't replicate in your home what you heard at the show.

And yet, while I appreciate these caveats, I, like most people, make judgments of speaker sound quality based on my impressions at the Consumer Electronics Show and regional audio shows. In fact, I choose most of the products I review based on what I've heard at such events. If I don't care for the way a speaker sounds there, I'm unlikely to review it, even though I recognize that its poor showing may have been due to one or more of the reasons listed above. I would rather not take a chance on spending months listening to a speaker whose sound I don't like, and instead select a speaker that impressed me positively. I believe that this sort of selection bias explains why Stereophile reviews tend to be positive. Stereophile's bias is to mostly seek out equipment that its writers think a) is very good, and that b) its readers will enjoy.

I first heard PSB Speakers' Imagine T2 Tower at the 2012 CES. What I heard was good enough that I wanted to hear more.

Description and Design
In designing its speakers, PSB uses the anechoic chamber and measurement facilities of Canada's National Research Council (NRC), in Ottawa. PSB designer Paul Barton is very much a "hands-on" designer; he told me that last year he spent five months in China, supervising the manufacturing of PSB speakers.

The Imagine T2 Tower is a lovely-looking speaker, well proportioned and with the sort of finish that, just a few years ago, would have been the exclusive purview of speakers made in Italy by master craftsmen. In fact, the T2 is made in China, using advanced technology as well as individual attention to detail. The cabinet is teardrop-shaped in cross section, which is more visually pleasing than a plain rectangular box, and has the acoustical advantage of having no parallel internal walls. (A number of speaker manufacturers, including B&W and Wharfedale, have adopted this sort of cabinet shape.) The cabinet panels are formed of seven layers of MDF, pressed into shape, while the front baffle is made of 2"-thick MDF. Considerable effort is made to match the grain of the wood veneers, to create a "cathedral" appearance.

The Imagine T2 Tower evinces similarities to PSB's Synchrony One, reviewed by John Atkinson in the April 2008 issue, and might even be described as a scaled-down version of that model, much of whose technology it shares. However, all the drive-units were developed specifically for the Imagine T2, whose crossover takes into account those drivers' characteristics, the distances between them on the baffle, and the height of each driver from the floor, to minimize interference between direct sound and the floor bounce.

The Imagine T2 Tower's driver complement comprises: a 1" titanium-dome tweeter with a neodymium magnet, similar but not identical to the tweeter used in the Synchrony One; above that, a 4" clay/ceramic-filled, polypropylene-cone midrange unit; and below it, three 5¼", clay/ceramic-filled polypropylene-cone woofers. Each woofer has a distortion-reducing phase plug of aluminum, as in the Synchrony One, as well as its own internal chamber and rear-firing port. Compared to speakers from other manufacturers, an unusual aspect of the T2 Tower's design is what Barton calls a "transitional" crossover in which each woofer is crossed over to the midrange at a different frequency: the one closest to the floor at the lowest frequency, and each of the other two at an incrementally higher frequency (but see "Measurements" sidebar—Ed.). As well as each woofer having its own port, PSB also supplies two rubber plugs, to optionally block the outputs of the two lower ports.

Setup
The Imagine T2 Towers seemed quite happy being plopped down in the general area where speakers normally sit in my listening room, but benefited from my tweakings of the distances between them, and from each speaker to the wall behind it. In their final positions, the speakers were aimed so that their tweeter axes passed just outside my ears when I sat in the listening position, forming an angle of about 60°. I then installed the supplied spikes, which somewhat tightened the bass and increased the specificity of imaging.

As mentioned earlier, the Imagine T2 Tower is supplied with plugs that permit selective blocking of the woofer ports. I tried the speakers first with no plugs (corresponding to a fully ported design), then with the lowest port plugged, and then with plugs in the two lowest ports. (I didn't try other combinations.) My preference, when listening to material that had a good amount of mid- and low bass, was for the two lowest ports being plugged. The bass in this configuration was tighter, and subjectively more extended than with no plugs—even the midrange seemed cleaner.

The Imagine T2 Tower is provided with a curved metal grille that's attached to the speaker at several points, thus reducing possibility of the grille rattling. I listened to the speakers with the grilles on and off, and was surprised by the degree of veiling the grilles introduced. I strongly recommend leaving them off. They're easy to attach and detach, for those occasions when you're entertaining small children or adults who may be tempted to poke the drivers.

I had three amplifiers on hand to try with the Imagine T2 Towers: my McIntosh Labs MC275LE (75Wpc, tubed, paired with the Convergent Audio Technology SL-1 Renaissance preamp); a Simaudio Moon Evolution 860A (200Wpc, solid-state, paired with Simaudio's Moon Evolution 740P preamp); and a PrimaLuna ProLogue Premier integrated amp (40Wpc, tubed). My description of the overall sound of the Imagine T2 Tower represents a kind of "averaging" of its performance with these three amps, with differences as noted.

Unlike the Wharfedale Jade 7, which I reviewed in the May 2013 issue, and which had to be played for about 150 hours before it sounded its best, the Imagine T2 Towers seemed to require no break-in at all: the only changes in their sound during the review period resulted from tweakings of their positions. (The review samples had gone through the measurement regimen at the NRC, which may have served as break-in.) They did benefit from a good warm-up, though: their sound was more "relaxed" after about a half-hour's play.

Sound
In his book The Audio Glossary (extracted here), Stereophile founder J. Gordon Holt defines accuracy as "(1) The degree to which the output signal from an active device is perceived as replicating all the sonic qualities of its input signal, and (2) The ultimate objective of an ideal system, which everyone claims to want but nobody likes when he hears it." Gordon's definition of euphonic is "Pleasing to the ear. In audio, 'euphonic' has a connotation of exaggerated sweetness rather than literal accuracy."

The Baton uses the tried-and-true two-way dynamic design, with a 7" coated-paper woofer and a 1" fabric-dome tweeter. The tweeter comes with a little Marigo dot stuck to its center to shape its response. It's not a physically easy task for a woofer to reproduce (well) all the frequencies from about 60Hz up to about 2 or 3kHz. One that succeeds is a nice find, though, because the sound has a nice coherence to it when most of the music is coming from the same driver. But don't take my word for it—just look how many zillions of two-way speakers there are out there.

The cabinets of these speakers are quite dead, meaning they shouldn't add too much garbage to the sound coming from the drivers. The walls are 1" thick and braced on the inside, and the edges of the cabinet are rounded to reduce detrimental diffraction effects. In a nice touch, Swans puts the crossover inside a sub-compartment to isolate the components from internal sound waves. The drivers are sonically matched in pairs at the factory prior to assembly, after which the speakers are submitted to lengthy break-in calisthenics. After break-in, the speakers are tested again. Then they're shipped out, and you get to test 'em.

Sound
When I set up the Batons in the "minimized room modes" position that worked well with my B&W 804s, their bass was a little boomy. I realized that, since these speakers are rear-ported, I'd have to check all the wall distances to the ports, as well as to the woofers. This meant twice as much work and trickier problem-solving than was required by the Thiel CS1.5 and Unity Audio Signature Mk.3 speakers that I also review this month, whose woofers, ports, and passive units were lined up more or less vertically. Not to worry, though—an inch this way, a couple inches that way, and all the numbers worked out oh-taay.

I found the Batons sounded best tilted way back. It didn't sound like there was driver time coherence until I got my eyes on the axis of the lower grille anchors, below the woofer. (Be careful: if you tilt them back much more than this, they might fall over.) With the amount I used, though, they still seemed to have enough stability. The threaded spikes (included) are nice and large, and have no nuts attached. These spikes were easy to adjust, looked plenty strong enough, and gave lots of maneuvering height to reach a good tilt angle. I wish more speakers had spikes like these.

After a bit more listening it became apparent that the Batons should be toed-in so that they point directly at the listening position. This configuration gave the most detail, and the Batons needed all the detail they could get. The sound was less involving if they were angled away from me at all. Also, I preferred listening with the little grilles removed. (I've found that grilles in general interfere too much with the top octave.)

Steve Tibbetts' recent masterpiece The Fall of Us All (ECM 1527), is a wonderfully weird album with one of the most diverse collections of instruments I've heard: rock'n'roll guitars, New Age-y synthesizers, and ethnic Asian instruments and drums. Tibbetts' gift for composition really turns me on, and his creative and masterful use of his analog studio doesn't hurt, either. Through the Swans, I heard less detail than with my reference B&Ws. The Swans seemed to give everything a warm, wooden quality, which helped the hand drums to sound quite realistic, but metallic instruments came out sounding a bit muffled. The lower-midrange detail was actually fairly good, and the bass was fairly tight, but there appeared to be information missing from the upper mids on up.

The Batons' bass sounded less extended than that of the Thiels or B&Ws. They went down almost as low, but their midbass was a bit recessed. With Dorian Records' Sampler II (DOR-90002), I had the impression that the bass was slightly too lightweight. Though in general the Batons' bass had good speed, with little overhang, their upper bass somewhat obscured the midbass, making the bass viols and organ pedals difficult to pick out. The bass coloration overall is not that large; I'm just pointing out minor flaws.

I noticed some colorations in the midrange. Some of these were admittedly due to the room, but others were present regardless of where I placed the speakers. I'm not going to try to name particular frequencies; I'll just say there were peaks and dips slightly bigger than expected in speakers of this price. These colorations, in addition to the lack of detail overall, made me want to stop listening to classical music on the Batons. The instrumental timbres were changed enough to detract from the level of realism that normally excites me with well-recorded acoustic music. I reached for rock'n'roll instead.

Now for the good news. I put on Toto's Isolation (Columbia CK 38962) and immediately started to enjoy myself. This ain't no cave, and my name ain't Bruce Wayne, so out with Bat-on, and in with Rock-On! These speakers worked wonders with this recording. Much of the overbright balance, recording artifacts, and excessive cymbals were mercifully absent. Through the "Rock-Ons," this recording sounded more natural than usual, thanks to the speaker's slightly soft treble and rolled-off top octave. This is not high fidelity. I think I'd call this Fidelity Band-Aid. I also think I like it on some recordings, even though I know better.

Moving on to Journey's Greatest Hits (Columbia CK 44493), I wrote, "Now I'm on to something here. Steve Perry's voice has a warm, natural-sounding roundness, but I've never heard this recording sound this good before. The electric guitars sound particularly good." I had never heard "Don't Stop Believin'" sound as good as it did on the Swans "Rock-Ons," and it made me wonder if the speakers the recording engineer had used had a similar balance. I was having a lot of fun with this synergistic match between recording and speaker. I hadn't known that it was possible to strap a pair of speakers on to the end of a transparent, neutral playback system, and have them systematically transform bright, gimmicked recordings into more natural-sounding music.

The main drawback left with the Journey was the speakers' tendency to reduce ambience, even though there isn't much on these recordings. What would normally sound like a medium-sized recording studio sounded a bit smaller and more heavily damped. The left/right image placement was well-defined, and the images themselves gave a fairly good illusion of being in the room with me, but they were more atmospheric than well-fleshed-out. Everything sounded a little softer than real, and the depth was foreshortened a little.

As my Zen amps have a moderately high output impedance of about 0.8 ohm, I wanted to try the Swans with a more characteristic transistor amplifier. I hooked up the NAD 2100X, which has an output impedance less than 0.1 ohm, and the treble sounded a little improved. The difference was not that large, but it was definitely a change for the better. Later, I performed the opposite experiment, connecting 2.2 ohm resistors in series with the Zen and the speakers. This simulates an amplifier impedance of 3 ohms, such that you might find in a single-ended tube design. The treble sounded worse with this setup. I definitely recommend using amplifiers with low output impedances with the Swans. I didn't use the NAD with them all the time, however, because it couldn't compete with the Zen in terms of sound quality. (It's a good amp for the money, though.)

The renamed "Rock-Ons" performed admirably in the Greenberg Memorial Loudness Test. (As Corey Greenberg doesn't write for Stereophile anymore, I figured we should name something after him.) I played some Megadeath and some Warrior Soul, and turned the volume up beyond the point where I had to put cotton balls in my ears to be comfortable. The Swans uttered nary a complaint. No woofer bottoming out. No tweeter bits on the floor. Not even any kind of strain or excessive distortion. This refusal to expire (or even perspire) is impressive for a pair of 7" two-ways, and convenient considering how loudly their preferred program material wants to be played. So bring on your poorly-recorded-yet-lots-of-fun Rap, Heavy Metal, Alternative, Industrial Dance, and Disco, and turn it up.

Conclusions
I wanted to like the Swans Baton a lot. It's so well made and looks so good, you just automatically expect great things from it. But I can't recommend the Baton for music-lovers in general. Their sound just did not offer a clear enough view into the recording—or, by extension, into the original performance—for them to be competitive at this price. Their lack of treble detail, reduced ambience, and midrange colorations take them out of contention for neutral transducer honors (footnote 1).

Nevertheless, people who have very specialized collections of nothing but over-bright, over-processed rock and pop music might want to give the Batons a listen. Why should you buy the relatively wimpy Swans as your dedicated rockin' speakers, when you could have 15" woofers? Because, silly, those monstrosities are usually crossed-over too high in frequency, so that they're dragging their slow 15" butts all over the midrange. Well, not in my midrange you don't! Kickdrum, guitars, and voices all sound way better when most of their sound comes from a single, fast woofer. And remember, swans may look pretty, but they can attack if provoked. Rock On!

I first heard the Alexia at the 2012 Rocky Mountain Audio Fest, but didn't decide that I needed to get a pair into my listening room until the following January, when, at the 2013 Consumer Electronics Show, I heard a pair of them driven by Dan D'Agostino amplifiers and a dCS Vivaldi digital source, hooked up with Transparent cables. As I wrote in my online report from CES 2013, when I heard the hi-rez master file of Cantus singing Eric Whitacre's Lux Aurumque, "not only was the relationship between the sounds of the singers and musicians and the surrounding ambience of the recording venue breathtakingly real, so was the relationship between the musicians and the music. I have played Lux Aurumque on dozens of systems—never have I heard it sound so real, so musically involving." (footnote 1)

The Alexia
The floorstanding Alexia bears a strong family resemblance to the two earlier Wilson speakers I reviewed. Its truncated pyramidal profile resembles the Sophia's, while, like the WATT/Puppy, it comprises separate enclosures: the lower, rectangular one holds the two woofers, and the upper, pyramidal one the midrange unit and tweeter on its sloped-back front. And the Alexia is finished in high-gloss automotive paint.

There the resemblance ends. The Alexia is larger than either of the earlier speakers. The latest (and best-sounding) iteration of the WATT/Puppy, the W/P Sasha, reviewed by Art Dudley in July 2010, stands 44" high and weighs 197 lbs. At almost 54" tall, the Alexia is a head taller than the Sasha and weighs a backbreaking 256 lbs. But the newer speaker wears its bulk well, and its footprint in the listening room is not significantly greater than those of the smaller speakers.

The more significant difference concerns the drive-units. Like Wilson's various MAXX and Alexandria models, the Alexia's two woofers are of different sizes. An 8" woofer is mounted above a 10", both loaded with a large, 3"-diameter, aluminum-lined port on the cabinet rear. The idea behind using two woofers of different sizes is that the radiation pattern of the smaller one at the top of its passband better matches that of the midrange unit at the bottom of its passband. Whereas the XLF's woofer cones are made of Focal's proprietary W sandwich material and the Sasha's twin 8" cones are of polymer, the Alexia's bass drivers have paper cones.

The Alexia's 7" midrange driver, mounted in its own enclosure, is the same as that used in the XLF; it features a composite pulp/carbon-fiber cone, with a substantial half-roll rubber surround to confer greater dynamic range at the bottom of its passband. The midrange unit is resistively loaded by two foam-lined vertical slots on the rear panel, and the enclosure can be moved back and forth and have its tilt adjusted with spikes of various size that couple its rear to a stepped metal plate on the top of the woofer module. The Alexia's tweeter is a variation of the sophisticated silk-dome model Wilson used in the Alexandria XLF. This, too, is mounted in its own enclosure, which engages with both the top of the midrange enclosure and the underside of the black-anodized aluminum "bridge" that covers the module via spikes that sit in grooves machined into metal plates. The tweeter module can also be moved forward and back, and tilted, with respect to the midrange enclosure on which it sits. Wilson calls this ability to fine-tune the upper-frequency drivers Aspherical Group Delay; I refer you to Michael Fremer's explanation.

Electrical connection is via a single pair of brass binding posts mounted the rear of the woofer enclosure and standing proud of the cabinet. Two pairs of heavy-gauge cables emerge from the top of the woofer enclosure, and connect to two pairs of binding posts on the rear of the midrange enclosure: one pair each for the midrange and tweeter. All four of the Alexia's drive-units are connected with series power resistors in their feeds from the crossover, which is inside the woofer enclosure. The resistors for the upper-frequency units are accessible behind a metal plate on the top rear of the woofer enclosure. These are mounted on a heatsink and held in place by Allen-head bolts. Replacement is easy, and different values can be substituted to adjust the levels of the tweeter and midrange. Barrel resistors adjust the woofer level and damping; these are not intended to be changed by the buyer.

In common with the W/P Sasha and Alexandria XLF, the Alexia's enclosures are far removed from the usual rectangular monkey coffin. A subtle wave motif is machined into the woofer enclosure's side panels, and a pair of vertical walls flank the midrange module. The midrange enclosure is fundamentally a truncated pyramid, but with shoulders machined into the side panels and enclosing sidewalls rising either side of the small tweeter module, to support the metal bridge mentioned above. The walls of the woofer enclosure are built entirely of Wilson's proprietary X-Material, a mineral-loaded phenolic compound that is both extremely stiff and difficult to machine. Laser interferometry was used to optimize wall thickness and the placement of internal braces. Cloth-over-frame grilles are provided for each of the three modules; these have pins that plug into matching sockets in the front baffles.

The Alexia was designed by Dave Wilson and Vern Credille, Wilson's lead acoustic and electrical engineer. It is gorgeous to look at, but what matters most to audiophiles is how it sounds.

Setup
Wilson Audio is known for its attention to detail, even when that detail might seem unnecessary. For example, the manual states, "Place the ODD numbered modules in the LEFT channel section and the EVEN in the RIGHT channel position." Even the packaging is superbly thought out, and complete sets of tools and accessories are provided. Perhaps as should be expected at this price level, Wilson works hard to maximize the purchaser's pride of ownership. The head units are both contained in one crate; each woofer cabinet rolls out of its individual crate on sturdy wheels, which allow the speakers' positions in the room to be easily fine-tuned despite their size and weight.

When you buy a pair of Wilson Audio Specialties loudspeakers, the retailer will install them in your home and perform that fine-tuning. In my case, Wilson's Peter McGrath did the deed. Having adjusted the position and tilt of the tweeter and midrange modules for the height of my ears in my listening chair and their distance from the speakers—the exact settings are detailed in the manual's "Propagation Delay Correction" table—he rolled each speaker back and forth, and from side to side, until he was confident that they were close to their optimal positions. McGrath then placed two strips of masking tape at 90° to one another, each marked with a ½"-spaced grid. Listening carefully to each speaker in turn, he moved the enclosure in ½" steps in both planes and adjusted the toe-in until each Alexia sounded its best. (The grilles were left off.) He then replaced the wheels with the spiked feet and declared himself satisfied.

McGrath had fine-tuned this setup using a single track: "So Do I," from singer-songwriter Christy Moore's This Is the Day (CD, Sony 5032552). I asked him what was so helpful about this recording. "This track features two acoustic guitars, double bass, and voice. There's a little bit of reverb on the voice, enough elements to make it sufficiently complex, but not overwhelmingly so," he explained. "The voice has been recorded without microphone proximity effect, so if you hear what sounds like proximity effect, it's actually due to a room mode."

Listening
When Peter McGrath left, I began my own listening. Pink noise sounded smooth and uncolored when I was sitting in my chair on the exact listening axis set by McGrath, using the position adjustments for the tweeter and midrange modules and the precise amount of toe-in he'd settled on. However, if I moved slightly above or below that position, a narrow band of brightness became apparent.

System Audio originated in 1984, when guitarist and electronics technician Ole Witthoft grew dissatisfied with the lack of realism he heard from most home audio systems and figured he could do better. He built some speakers for himself and for a few friends, with encouraging results. It's a familiar story: we all know competent hobbyist speaker builders. A few of them gain a bit of local notoriety, but most never venture further than making a few units for friends and relatives.

But Witthoft's reputation grew rapidly, and so did his business. Fourteen years later, his little startup has become a serious player in the loudspeaker market, with annual production in excess of 18,000 units.

With growth have come accolades from reviewers and users in Europe, the UK, the US, and elsewhere. System Audio's 1050 was designated a "reference" by the German hi-fi magazine Stereoplay. Columbia Pictures uses 15 pairs of the company's 5010 Signature model in its Los Angeles studios. Copenhagen's Academy of Modern Music has 16 pairs of the 2010 in its classrooms. The 1070, which resided with me for several months, was named 1997 "Loudspeaker of the Year" by Scandinavian hi-fi journal Lyd & Billed.

I first encountered Ole Witthoft and his diminutive minimonitor, the 905, at the 1994 Winter CES. The little loudspeaker, driven by Densen electronics, sounded both musical and amazingly dynamic. There was something so seductive about it (and the other Danish products being shown that year) that I kept making excuses to go back and hear it. It had an open, effortless quality that I found extremely appealing—a quality shared by the PBN Montana SP (reviewed in January 1997, p.225). (The 1070, in fact, resembles a small-scale version of the SP.) Laborious-sounding, insensitive loudspeakers are not on my wish list.

Fast-forward to HI-FI '97. Strolling past the Nordost suite, I heard that same effortlessness again. There, in the midst of cable displays and product posters, was a pair of 1070s pumping out some really infectious rock. This was a good sign. It meant the manufacturer wasn't afraid to play real-world music at real-world levels, and proved the speaker could handle plenty of roughhousing and still come back for more. Best of all, the 1070 sounded really good in the small hotel room, a space not much different from what most people have to devote to home entertainment.

Lars Kristensen, who always seems to be enjoying some private joke, gave an enthusiastic and entertaining demonstration of both the System Audio speakers and an assortment of Nordost cables. His affable associate Joe Reynolds arranged to get a pair of the speakers to me for review. A few weeks later, while touring their dealers in Northern California, Sonic Integrity's Peter Hansen and Pat Mulcahey showed up with a pair of rosewood 1070s. Peter and Pat are friendly, helpful, and extremely knowledgeable guys who didn't object to spending the better part of an afternoon positioning the speakers for the best sound. A couple of weeks later, Lars and Joe, on a similar venture, showed up and repeated the exercise. These Nordost guys are serious about music.

Sound Quality
My relatively large listening room (+400 sq. ft.) allowed the speakers to stand well away from both the side and rear walls. For most of my listening, the 1070s were out 3'–4' from the back wall and about 7' apart, with a toe-in of approximately 15°. This seemed to produce the most coherent imaging from the little columns, but the bottom end was more extended when they were closer to the wall behind them—typical behavior for rear-ported speakers. As with most speakers, the 1070's bass response can be tailored to a certain extent by moving it in relation to the wall.

Regardless of position, the System Audio's low-end output was consistently quick and punchy. Rock music especially benefited: dynamic pieces like Joan Osborne's "St. Teresa" (Relish, Blue Gorilla/Mercury 314 526 699-2) and No Doubt's "Don't Speak" (Tragic Kingdom, Trauma INID-92580) were repeatedly delivered with real impact. Softer recordings were equally effective—such as audiophile favorite Diana Krall's "Boulevard of Broken Dreams" (All for You, Impulse! IMPD-182), which proved to be as thoroughly involving through the 1070s as through many other, more expensive speakers. With good equipment, this excellent recording can produce quite a three-dimensional soundstage. The 1070s were very good in this department.

Sara K's cover of "Brick House" (Hobo, Chesky JD155) is another great recording for both imaging and pacing, two of the 1070's strong suits. Vocal presentation was clear and surprisingly uncolored—a mild surprise, given that the two little bass/midrange drivers cover a fairly wide segment of the audio spectrum. By comparison, the midrange was more "pure" than both the $850/pair Spica TC-60 and the $8000/pair PBN Montana EPS (review soon), but not by a big factor.

Spec'd at ±4dB, the 1070 diverges from an ideal "flat" frequency response in a most musical manner. Paradoxical as this may seem, while a flat frequency response is absolutely desirable in electronic gear, truly flat-response loudspeakers can sometimes sound dull and lifeless. Why this is I'm not sure, but musically involving speakers often depart from "accuracy." This doesn't particularly bother me if the result is musically engaging and the presentation believable. Like movie fans and opera goers, audiophiles must indulge in a certain amount of willing suspension of disbelief. For me, all a loudspeaker has to do is make that exercise easy, and it can reel me in like a fish on a hook.

The 1070 made this especially easy with guitar music—not surprising, considering that designer Ole Witthoft is a guitarist. Lyrical guitar instrumentals like Stevie Ray Vaughan's "Little Wing" (The Sky is Crying, Epic EK 47390) or Joe Satriani's "The Forgotten" (Flying in a Blue Dream, Relativity 88561-1015-2) were delightfully seductive and involving. Likewise teenage blues sensation Jonny Lang (Lie to Me, A&M 31454 0604 2). B.B. King's Deuces Wild (MCA MCAD-11711) was in very heavy rotation hereabouts during the 1070's stay.

Rock and electric blues aren't the only types of music these speakers do well. Great jazz recordings like Stan Getz's Bossa Nova (Verve Jazz Masters) or Chet Baker's My Funny Valentine (Pacific Jazz CDP 52826 2) never grew tiresome, even with repeat playing. And leave your smug, ironic commentary at the door: I'm a sucker for the stunningly bombastic, like Toni Braxton's "Unbreak My Heart," from her self-titled CD (LaFace/Arista 26020-2). Overproduced? Ask me how little I care. The girl's got an incredible voice. If overwrought opera stars can raise bombast to high art, aren't pop divas entitled to do the same? Braxton consistently delivered the goods through the 1070s.

System Audio's speakers also worked well in a two-channel bedroom audio/video system (Marantz DVD810, Sony TA-E77ESD, Kenwood KM-106), where they displaced a venerable but near-mint-condition pair of KEF 104/2s. Movie dialog was rendered with clarity, music with emotion, and sound effects with kick.

Conclusion
Overall, the System Audio 1070 was a rugged, reliable performer that never let me down, regardless of the material I threw at it. With its punchy bottom, lyrical midrange, and extended top end, it was capable of providing sustained satisfaction with a wide variety of music. It was also compatible with a wide assortment of supporting electronics—I experimented with everything from bargain-basement gear (RadioShack 3400, Chase Technologies RLC-1, Dynaco Stereo 70) to fairly rarefied high-end electronics (PS Audio Lambda 2 transport; Threshold T-1D DAC, T3i preamp, and T-100 amplifiers; Nordost SPM cables) and was never disappointed.

The 1070 was efficient and easy to drive, and should work well with moderate amounts of power: 100Wpc is probably all that's necessary in any space where it's likely to be used.

A limited bottom octave was the little speaker's only serious drawback—substantial output from 20 to 40Hz is asking a lot from something that's closer to a minimonitor than to a full-range loudspeaker. Even so, the bass this speaker produced was quite satisfying. If deep bass is your top priority, the PSB Stratus Goldi (reviewed by John Atkinson in October '97, p.199) or the TDL T-Line 3 (reviewed by Muse Kastanovich in March '98, p.123) are better choices in the $2–3k price range.

But both of those products might overwhelm the System Audio's "natural" environment: a small listening room. If you're among the myriad of rock, pop, and jazz fans living in apartments, townhouses, or condos, the 1070 is strongly recommended. It works well with a wide variety of electronics, is physically unobtrusive, and looks pretty with or without its grille. Its light weight makes it easy to position, and, best of all, it sounds great. I got tons of enjoyment from this loudspeaker. You might, too.

To get these overgrown bookends to stand up, you slide little black boards into the slots in their bottoms. Each board is held in place by two set screws, and sticks out to support the speaker with two of the four spikes. The board also tilts the speaker back a little. How do they get sufficient bass out of such slim cabinets? Unity Audio is glad you asked. They call it "Inverse Force Vector Coupling," and are very proud of their development, isolated through "extensive physics research." Basically it consists of one woofer pushing while the other woofer pulls. They can tune this system by varying the dimensions of the tube that runs through the middle of the cabinet. Unlike an acoustic suspension system or a ported system, this type of bass alignment is said to be able to work with an extremely small internal volume.

The Signature 3 has a 7" polypropylene-cone, cast-basket woofer mounted at the top of the front baffle, and a second one mounted at the bottom of the speaker's back wall. This rear woofer puts out midrange as well as bass, so whatever is behind the speaker will affect the sound more than with an ordinary front-firing speaker. It also has a 1" silk-dome tweeter. My first samples were old enough to not have the ceramic coating on the tweeters, and they looked like the same units as in the Swans Baton reviewed in this issue, but without the Marigo dots. Unity Audio apparently has done quite a bit of resonance-structure research for the development of their new tweeter, investigating different ways of getting the best transient speed from the dome while trying to avoid ringing. They found a lightweight ceramic material more than 10 times as rigid as aluminum, and which can be deposited in very thin layers. Multiple layers of this are used on the dome, alternating with thin layers of a very effective damping material. All recent production of the speaker has featured the new tweeter.

Remember what I said about packaging costs? The crossovers of these beauties feature copper-foil air-core inductors, Kimber Kap polypropylene capacitors, and pure silver wire. The cabinets are internally braced, and small enough to not want to resonate much anyway. They did seem a little wobbleriffic, though, so I placed some disks on them. Was it those neato bingo thingos? No, these were 2kg of cast concrete (with calcium!) encapsulated in a luxurious no-scratch plastic casing and delightfully contoured for maximum domestic appeal. That's right, kids, it's Orbatron! I bet, like me, you have some hiding in your basement that you can fetch if you need it for resonance damping. Just remember to ask your parents first. (By the way, I put the weights on the speakers' bottom braces.)

The Way the Signature 3 Used to Sound
When I was taking a look at the first pair, I saw a little bump on the woofer surround. It felt hard. I removed the woofer and found some nice things inside. The input wires are soldered directly to the woofer terminals, eliminating crossover components. The woofer crossover is accomplished mechanically, by mass loading the cone with a coating of hot glue.

Which brings me back to the bump. It was a stray piece of glue that had accidentally splashed onto the surround and stuck there. By reducing the elasticity of that side of the surround this could theoretically be bad for the sound. In my listening, though, I had difficulty detecting any difference between the sounds of the two speakers. At least this little defect proves that Unity Audio doesn't give preferential treatment to speakers that will be sent to reviewers, which is nice to know. I only found such bumps on two out of the eight woofers of the two review samples; they don't seem to affect the sound much, but Unity's glue people should be more careful.

I substituted the Signature 3s after listening to the Swans Batons for a week. It was immediately obvious that the Unity Audio reproduced well-recorded music in a more neutral fashion than the Canadian design, and this impression continued throughout my listening. I listened with the grilles off, mostly with the speakers in the same positions as the Batons. I found the bottom of the woofer to be a good listening axis—though this is a low 27" from the floor—and preferred a slight toe-in.

With Steve Tibbetts'The Fall of Us All, I got the sense that the Signature 3 was well-balanced, with better top-octave air and a much more spacious presentation than the Swans. The only part of their presentation that seemed to be substandard was the midbass: there wasn't much. True, the upper bass was of high quality and at a good level, but a certain visceral push was missing. The bass-frequency tones on Stereophile's Test CD 3 (STPH006-2) went missing in action below about 60Hz.

Despite being endowed with what looked like the same tweeter as in the Batons, this pair of Signature 3s had a brighter presentation. Instrumental timbres were more natural with source material of the trustworthy, unequalized type. They also had more treble detail to offer the listener, although I felt they couldn't quite equal the Thiel CS1.5 and B&W 804 in this department. Partly because of their nice top end, they provided a large sense of space from their reproduction of the reverberation info on the recording. Their spaciousness was also augmented by the rear firing mid-woofer. While this was a bit too much to be strictly realistic, it was very pleasing, and certainly closer to reality than the Batons' persistent tendency to reduce soundstages to the equivalent of a Sonexed bedroom.

This first pair of Signature 3s had a good midrange: fairly smooth in terms of tonal balance, but with a few small departures from linearity. The really cool thing about their midrange, though, was the amazing transient speed. With no crossover to get in the way, the Signature 3 reproduced all the mid detail that the amp could feed it. After all of my mod madness, and other listening escapades, I know how well simplicity can take me closer to the original event of the performance. With this setup there were only two resistors, two capacitors, one MOSFET, and four mechanical connections between the CD player's output and the midrange's voice-coil. I was a happy little pew-wrist.

The Sound of the Current Signature 3
Towards the end of the review period, Unity submitted a revised pair of Signature 3s (footnote 1), these featuring the revised ceramic-coated tweeters.

These new tweeters were great. They seemed to do well with transient details such as twinkly little sounds. Maybe they were a little too twinkly, but not much. Because the woofer has no electrical crossover, the upper end of its range invades the tweeter's territory. Nevertheless, the crossover region sounded quite smooth with the speakers arranged for only a small amount of toe-in. I felt the top octave was a little clearer with the new tweeter. Timbres sounded a bit too bright and steely, but this was to a very slight degree. Overall, the Unity Audio crusaders seem to have succeeded in their quest to lower the excess tweeter-dome ringing.

I flashed back to one of the first times I'd heard the Nautiluses—in 1994, at the home, nestled against England's South Downs, of B&W's then-owner, Robert Trunz. Listening to the Nautiluses, quad-amped by Krell monoblocks in Trunz's system, was an ear- and eye-opening experience. The four-way speaker, designed by Laurence "Dic" Dickie, was an all-out assault on the state of the loudspeaker art, with aluminum-diaphragm drivers intended to behave as perfect pistons throughout their passbands, each loaded by a true transmission line to absorb its backwave. It is the coiled line for the woofer that gives the Nautilus its distinctive shape, resembling that of a giant snail.

Trunz was soon to sell his stake in B&W to the Equity Group and retire to South Africa, where he started a world-music record label, MELT2000. Dic Dickie also moved on, to design high-quality sound-reinforcement loudspeakers for companies like Turbosound. In 1999, Trunz introduced Dickie to two South Africans, Philip Guttentag and Bruce Gessner, who were thinking of starting a company to manufacture high-tech loudspeakers. The result was Vivid Audio, and in 2004 came Vivid's first loudspeaker, the Oval B1, which I reviewed in October 2011. The B1 was based on developments of the ideas Dickie had first used in the Nautilus, which reached full fruition in Vivid's Giya G1, which Wes Phillips reviewed for Stereophile in July 2010.

John Marks examined Dickie's ideas in his October 2012 "Fifth Element" column; "Dickie's designs for Vivid [minimize] cabinet resonances and diffraction through the use of unique cabinet shapes and high-tech materials and fabrication techniques. He also aims to achieve uncolored, distortion-free sound by making all drivers from the same proprietary alloy, and by keeping driver behavior as pistonic as possible through careful driver and crossover design." In his G1 review, WP mentioned that "[Dickie's 'Tapered Tube loading' concept] consists of coupling a driver to an exponentially tapered tube filled with damping fiber. . . . All drivers have cylindrical magnets to leave a large-diameter vent behind the diaphragm in order to couple to the tapered tube, [and] are decoupled from the baffle by ring mounts."

The Giya G3
The G3 resembles the G1, but on a smaller scale, and costs $39,990/pair instead of $64,990/pair. To cover the range above 220Hz, the G3 uses the same three aluminum drive-units as the G1: a 1" catenary-dome tweeter operating above 3.5kHz; a 2" catenary-dome midrange unit covering the region from 800Hz to 3.5kHz; and a 4.9" lower-midrange cone covering 220–800Hz, all three set into a shallow depression on the front baffle. The crossover filters are all high-order. The ring-magnet topology permits the top three drivers to be closely spaced. It's just 10.5" from the top of the tweeter dome to the bottom of the lower-midrange cone—above the 220Hz crossover to the woofers, the G3 is basically a minimonitor, with all the attendant advantages for dispersion and imaging.

Whereas the Giya G1 has two 11" aluminum-cone woofers, mounted on opposite sides of the teardrop-shaped enclosure near the base and loaded with a unique vented, tapered line, the G3 has twin 5.5" woofers, still with aluminum cones and still loaded with a vented, tapered line. Along with the lower-midrange unit, the three cone drivers have chassis carefully shaped and cast to offer minimal obstruction to the cone's backwave.

The enclosure, molded from a glass-fiber–reinforced, balsa-cored composite material, is still teardrop-shaped and, as in the G1, its top section curves over to merge with the cabinet's rear. However, because of the G3's smaller height—45.7"vs 66.3"—the tubes loading the top two drive-units, which were internal in the G1, emerge from the rear of the enclosure and extend to meet the woofer line. The internal tapered line that loads the lower-midrange unit now extends to the end of the woofers' line, where the latter meets the enclosure. The result is an attractive speaker that, to my eyes, looks more elegant than the larger G1.

Electrical connection is via two pairs of binding posts housed in a recess on the speaker's base. This was the only gripe I had about the G3's design: connecting cables required the speaker to be tilted to one side, and, with the thick cables preferred by audiophiles, there is only just enough clearance in the recess to tighten the terminals, even when the six carpet-piercing cones are fitted to the base. But as Laurence Dickie pointed out when I bitched to him about this awkwardness, it actually affects only dealers and reviewers. Once the dealer has installed the G3s in the customer's home, the owner need never be bothered by the problem.

Listening
Vivid's US distributor, Philip O'Hanlon of On a Higher Note, set the Giya G3s up in my room. They ended up a little farther way from the listening chair and closer to the sidewalls than the Wilson Alexias that had preceded them, and O'Hanlon preferred the high-frequency balance and imaging with the speakers toed-in, though not quite to the positions of my ears. This worked for me, and I performed all of my listening with the speakers set up that way. I also left off the vestigial wire-mesh grilles, the dome units still being protected by cruciform grilles.

As always with a speaker review, I began my critical listening by playing dual-mono pink noise from my Editor's Choice (CD, Stereophile STPH016-2), which gives me a handle on a pair of speakers' frequency balance and imaging. Pink noise sounded smooth and seamless when I sat with my ears on the Giya G3s' tweeter axes, the two speakers sounding identical. The noise signal acquired a hollow coloration when I stood up, and sounded bit midrange-dominant when I slouched in my chair so that my ears were below the tweeter axes. There was no top-octave emphasis; in fact, the G3's treble was free of grain, and sounded sweet without being rolled-off or dull. High-frequency clarity was superb.

With a dual-mono signal, a pair of perfect loudspeakers should produce an infinitely narrow image precisely midway between them. The G3s got close to this ideal with a central image that was narrow, well defined, and didn't "splash" to the sides at some frequencies as a result of frequency anomalies or resonances.

The low-frequency, 1/3-octave warble tones on Editor's Choice were reproduced with full weight down to the 40Hz band, and with the 32Hz warble significantly reinforced by the lowest-frequency, diagonal mode in my room. The 25Hz tone was both audible and free from obvious distortion, though the 20Hz tone was inaudible at normal listening levels. With the half-step–spaced toneburst track on Editor's Choice, the tones spoke very cleanly down to 63Hz, below which there was a slight lack of energy apparent before a big boost at 32Hz.

This exaggerated low bass is, as I explain in the "Measurements" sidebar, due to the coincidence of the G3's port tuning frequency and the frequency of the lowest resonant mode in my room. The exaggeration is just low enough in frequency that it didn't add significant coloration, though it did contribute to the "magnificence" I noted with some orchestral recordings. But it made the choice of amplifier critical. For example, the rumbling bass drum in the third movement of Mahler's Symphony 2, "Resurrection," with Benjamin Zander conducting the Philharmonia Orchestra (24-bit/192kHz ALAC file, Linn CKD 452), needed more control than the Pass Labs XA60.5 amplifiers I reviewed in January could provide. Replacing them with the MBL Corona C15 monoblocks (review to appear in a future issue) brought the low bass under better control without affecting the Giya G3s' superbly defined imaging and transparent, grain-free midrange and highs.

The control required by the G3's woofers will make amplifier choice critical for getting the best from this speaker. But even with the MBL amplifiers, the double bass on my recording of Mendelssohn's Piano Sextet from the 1997 Santa Fe Chamber Music Festival, on Encore (CD, Stereophile STPH011-2), was a touch on the gruff side, though pianist Christopher O'Riley's insanely fast spiraling arpeggios were reproduced with crystalline clarity. And the solo kick drum that jump-starts "Slow, Happy Boys," from Gov't Mule's The Deep End, Volume 2 (ALAC file ripped from CD, Advance Music AT00006), was reproduced by the Vivids with superb definition of the leading edges of its sound, plentiful weight, and no low-frequency hangover. My Fender bass guitar in the channel-identification tracks on Editor's Choice spoke cleanly, with excellent definition, even at high sound-pressure levels.

The Metropolitan
The developmental history of Vortex speakers provides a meaningful framework for the design of the Clearfield offerings, especially the Metropolitans, or Mets. Like the Vortex designs, the Mets are three-ways with transmission-loaded bass. Like the Kevlar Reference Screen (reviewed by Robert Greene in The Abso!ute Sound's "double-issue" 83/84, December '92), the Mets use Kevlar-coned midrange units from Focal that cover a broad range from 125Hz to 2kHz. What's dramatically different is the overall driver layout. While all Vortex speakers use a stepped baffle for time alignment (like the Vandersteens and Thiels), the Mets begin with a flat front baffle and a D'Appolito—vertical midrange/tweeter/midrange—configuration. Clearfield believes that time-aligned configurations do not perform as well off-axis, although they can be outstanding performers on-axis. To achieve an acceptable range of off-axis performance, Clearfield opted for the D'Appolito arrangement, which mimics a point source.

The Met also differs from the Vortex designs in having doubled-up woofers, the two 8" woofers having an area equivalent to an 11.5"-diameter driver. Von Schweikert credits David Wilson and his WAMM for convincing him that you need a lot of drivers to move a meaningful amount of air to realistically re-create sound. The most meaningful (to me) decision regarding the woofers is the very low 125Hz crossover point. In many respects, the Met's bass drivers are configured to operate as a subwoofer. The two woofers are mounted in a transmission-line enclosure forming the bottom half of each cabinet, with the three-driver satellite setup in the top half.

The Metropolitan is impressively large, 2' wide and standing over 5' tall. The Met's cabinet is uniquely shaped. Seen from above, the cabinet looks like the letter V with the point flattened. This truncated point of the V becomes the 10"-wide front baffle and is covered by a grillecloth running the entire height of the speaker. To the left and right of the baffle the cabinet slopes toward the rear, a styling feature intended both to reduce diffractive effects of the radiated sound and to cut down standing waves inside the cabinet by minimizing parallel surfaces. The combination of cabinet shape, driver alignment, and specifically designed equalization circuitry leads to what Clearfield calls a "controlled directivity response" for improved soundstaging and off-axis performance.

The back of the cabinet tells a great deal more about the speaker. For starters, there are two circular ports for the transmission enclosure located near the bottom (footnote 1). Each port comes with a foam-rubber "Q" cylinder. While the speakers are intended for placement out in the listening room, Clearfield has correctly assumed that some small or problematic listening rooms may simply not allow such placement. In such cases, the bass output could be excessive. Inserting the Q cylinders into the ports modifies the bass output. The very detailed manual provides numerous suggestions concerning room placement to deal with this and any number of other potential problems.

Higher up on the back of each cabinet are two sets of five-way binding posts, along with gold-plated brass jumpers. Using the supplied jumpers, the speakers can be run with a single set of speaker wires from a single stereo amp (or monoblock pair). Since the speaker's sensitivity is quoted as a high 90dB/W, amplifiers with power ratings as low as 75W are recommended. My listening tests with the 50W Metaxas Iraklis amp confirmed this assertion. I was able to get a big sound out of the Mets with very little power. With the jumpers removed, the speakers can be bi-wired or bi-amped. Once again, the manual provides illustrations and explanations of each of these alternatives, including both horizontal and vertical bi-wiring (the latter, using one stereo amplifier per speaker, is my preference, as it only puts the demanding bass load on one channel of each identical amplifier). Though I tried them in every way possible, I used the Mets in a bi-wired configuration with various amplifiers for the majority of my listening.

On the back of each speaker is a second set of jumpers, used to adjust the level of the 1" aluminum-dome tweeter. Each speaker has five discrete balances (+2, +1, 0, –1, and –2dB). Since Clearfield feels the Mets can sound bright when they are breaking in (and they did), the manual suggests attenuating the tweeter levels during this period. The primary purpose of the tweeter-level controls is the same as that for the Q cylinders. Counterpoint recognizes that all rooms aren't the same, the various equipment likely to drive the speakers can vary tremendously, and people often simply have different preferences. The tonal balance of the speaker can be adjusted accordingly.

The back plate, where the binding posts and other controls are located, is made of a combination of two different-thickness aluminum plates sandwiched around an adhesive damping layer of Imoplex-G. Each cabinet is bolted to a massive base plate, or plinth (finished in splattered black), which provides added stability for the tall, heavy speaker. Spikes are mounted in the bottom of the plate.

The last significant piece of the puzzle can't be seen from the outside. The internal cabinet is built with a stressed monocoque technique borrowed from the aircraft industry. Rapping anywhere on the box yielded only the dullest of thuds, attesting to the virtues of this heavily cross-braced design. The speaker is available in a number of finishes, including light oak (the review pair), natural walnut, or black oak, with rosewood available for an extra charge.

The research that has gone into the development of the Clearfield speakers has been intense, not at all what would be expected from a "new" speaker manufacturer (though, as will be seen, the design of the speaker continued to evolve during the review period). In addition to the excellently written manual, Albert Von Schweikert sent me a 12-page, single-spaced letter with numerous attachments describing how each element of the design had been determined. From the Counterpoint side came extensive experience with high-quality capacitors, resistors, wire, and circuit boards, which are made of mil-spec fiberglass with 4oz copper traces.

Careening down
There was a major change made to the Metropolitan after production began: To improve soundstaging and image focus, the tweeters have been laterally offset from the center of each cabinet and are no longer symmetrically located (as is shown in the company's extensive advertisements). This turned the Mets into mirror-imaged, or handed, pairs. Since my pair arrived with this change already implemented, I can't comment upon its effect on the sound.

A number of pairs of the completed production speakers went out to dealers, as well as to myself and Tom Norton in Santa Fe. Since Tom and I had the speakers in for review, we commented neither to one another nor to Clearfield about their sonic performance. Unfortunately, we both found that performance disappointing. Had those first Met samples been all that were to be reviewed, the results would have been very negative due to the speakers' bass-heavy character. Counterpoint's dealers, however, didn't hold back from sharing their frustrations with the company. The assessments coming back from dealers were identical to the problems I was having with the speakers. Counterpoint responded promptly to the dealer feedback and made still further revisions to the Mets. These were reflected in the second pairs of speakers that I (and JA and TJN in Santa Fe) received.