Of the small number of times I have been totally swept away by listening to recorded music, a significant proportion have involved loudspeakers from Wilson Audio Specialties. It was my experience of their X-1/Grand SLAMM in the listening rooms of reviewer Martin Colloms, then-retailer Peter McGrath, designer Dan D'Agostino of Krell, and manufacturer Madrigal Audio Labs, that led me to name it my "Editor's Choice" for 1995 and join my vote with those of the Stereophile scribes to make it the magazine's "Loudspeaker of the Year." I wrote in my December 2001 "As We See It" about how a cross-country road trip had begun with a listen to the Cantus CD on the Wilson WAMMs in their designer's Utah listening room. And, as I wrote in my April column, auditioning Peter McGrath's 24-bit Nagra-D master tapes on Wilson MAXXes in the Halcro room was, for me, the highlight of the 2002 CES.

But yes, all three of these speaker systems are very expensive, ranging from $38,900/pair (the MAXX) through $70,000/pair (the X-1) to over a quarter-million bucks (the ultimate WAMM). As much as I appreciate and desire the attainments of speakers like these, my own music-making, like almost all Stereophile readers', has to be based in the real world of mortgages, car payments, and school and college fees.

So when I laid eyes and ears on Wilson's Sophia, priced to sell at a relatively affordable $11,700/pair, at the 2001 CEDIA Expo last September, I began salivating about how they would work in my room and my budget.

Sophisticated
In its general appearance, the Sophia looks like a single-box cousin of Wilson's WATT/Puppy (last reviewed in its System 5 incarnation in November 1995). And its 41"-high, floorstanding, three-way design concept reminds me of the now-discontinued WITT (reviewed in our January and July 1996 and January 1998 issues). But the Sophia is all new, from its handcrafted enclosure to its custom-built drive-units.

I usually begin my description of a speaker with its active parts: the drive-units and crossover. But having witnessed the labor-intensive craftsmanship that goes into each Wilson speaker, I'll start with the Sophia's intricate cabinet.

This is mainly crafted from what Wilson terms "M" material. According to Vern Credille, the company's R&D director, this composite material consists of a matrix of cellulose fibers bound together with phenolic resin to form basic sheets. These are then laminated together to make the Sophia's enclosure panels. For the speaker's front baffle for the woofer section, and for the base, Wilson's even more rigid "X" material is used, which is a high-density, mineral-loaded, methacrylate-based composite. The Sophia's assembled enclosure is spray-coated with automotive paints and clear coats, and rubbed down between each application, to produce an immaculate, high-gloss finish. (If you live in Utah and can't get your car repainted to the original factory standard, it's because Wilson tempted away the state's best workers for its production line.)

All this attention to detail is to ensure that the enclosure doesn't emit sound, thus providing an optimal environment for the components that do. Handling the highest frequencies is the same version of Focal's inverted-dome tweeter used in Wilson's home-theater speakers, its titanium diaphragm coated with dark-gray "tioxid." This crosses over to a 7" midrange unit sourced from ScanSpeak. The paper cone has a radiating diameter of approximately 5", and cone and dustcap are scored in a radiating pattern, the score marks filled with damping material. The long-throw woofer is a 10" aluminum-cone unit, reflex-loaded by a 3"-diameter aluminum port on the rear panel. The midrange unit is also reflex-loaded with a 1" port, again on the rear panel, presumably to increase its dynamic range at the bottom of its passband.

Setup
Peter McGrath, now with Wilson Audio Specialties, set the Sophias up in my listening room. If you're concerned that this means a reviewer is being given special treatment, I'm told that the dealer from whom you purchase a pair of Sophias will provide this service as a matter of course.

Before we did any listening, Peter marked off a grid in the probable speaker positions with masking tape and, standing in the center of each grid and crouching so his mouth was at the approximate height of the Sophia's midrange unit, began the setup procedure by speaking evenly. He and I were listening for the position where the coloration added to the sound of his voice by the room acoustics was minimized. This would be the starting point for deciding on the optimal placement of each speaker.

Once we had found those positions, we experimented moving the speaker ½" at a time in both horizontal planes, noting whether the midrange tonal balance became less or more even. (Half an inch may not sound like much, considering the wavelengths of sound in the midrange, but it can produce a surprisingly large change in the perceived balance.) At the end of this iterative procedure, the speakers ended up 56" out from the wall behind them—a little closer to the room boundaries than the Revel Performa M20s, which had preceded them. The spikes were then fitted, effectively locking the 160-lb Sophias to the floor.

Sound
I've mentioned the phenomenon before in these pages: Loudspeakers tend to affect your choice of music to listen to. Even though the small Revel M20s had excellent low-frequency extension in my room, I found that I was playing a lot of vocal and chamber music with them. Yes, this was to great musical effect, but it wasn't until I played back a hard-drive copy of Peter McGrath's 24-bit live recording of "Der Abschied," from Mahler's Das Lied von der Erde, on the Sophias that I realized how little orchestral music I had been putting on the player. Without a speaker that can reproduce the wide dynamic sweep of a symphony orchestra in full measure, the musical sense becomes diminished. But over the Sophias, as conductor James Judd—in one of his last performances with the Florida Philharmonic, the orchestra he had bullied, cajoled, and persuaded in the past decade or so to become a truly world-class ensemble—I felt there were no dynamic limits, that there was almost nothing between me and the event captured by Peter's microphones.

One of the major strikes against small speakers when it comes to reproducing the symphonic repertoire, of course, is their lack of low bass. This has always been one of the first things I will sacrifice in favor of natural midrange tonality, accurate imaging, and stable soundstaging. But that doesn't mean I don't value true 20Hz extension when I can get it in addition to those more important (to me) aspects of sound quality.

In many of his scores, British composer Edward Elgar indicates a part for organ marked ad libitum. You'd think that this gives them free rein, but I'm always surprised by how discreet organists are, adding only enough of the instrument's majesty to flesh out the sound when appropriate. But when the organ pedals are used to underpin a work's foundation, a speaker like the Sophia gives the bass fundamentals full measure. Our June 1998 "Recording of the Month" was a CD of the Bournemouth Symphony Orchestra performing Elgar under veteran conductor George Hurst (Naxos 8.553564). The main work on the disc is the popular Enigma Variations, and the organ is held back until the final, eponymous "E.D.U." variation, where it thunders out a descending bass line under the final climactic reiterations of the tune and reinforces the final cadence. The Sophias' wide dynamic range and extended lows allowed this music to be presented in as realistic a manner as I have experienced, particularly when I drove them with the Australian Halcro amplifiers. It was only when the last chord died away and I heard the phone ringing that I realized how loud I had been playing this CD.

But it was not so much the quantity of the Wilson speaker's high frequencies as their quality. In real life, where acoustic music comprises many different sound sources, those sources are perceived as discrete entities. By contrast, with reproduced music, the acoustic objects tend to blur into one another. Think of a thin rubber sheet pulled down over a series of physical objects. The objects are now perceived as projections thrust up out of the rubber matrix; they may appear discrete when viewed from certain angles, but they are actually connected at an underlying level. This is exactly analogous to the creation and perception of acoustic objects in a stereo recording, with the rubber replaced by all the artifacts and grunge added by the recording and playback processes. The worse the audio equipment, the higher the level of grunge, the less differentiation there is between the perceived objects, and the harder it is to make sense of the music.

Don't let the fact that I kept returning to orchestral recordings in my listening sessions leave you with the impression that the Sophia was a one-trick pony, optimized for just one kind of music. In fact, its high frequencies were delicate enough, its midrange neutral enough, that chamber works were reproduced in a manner to rival the finest minimonitor. Peter McGrath had left me with a copy of a CD (Naxos 8.559067) he had engineered of a piano trio and quintet and a violin sonata by the American composer Charles Wakefield Cadman. Despite his being voted 1930's "Most Popular American Composer" by the National Federation of Music Clubs, Cadman wrote in an unashamedly Brahmsian idiom. There was no sense via the Sophias that these intimate works were emanating from loudspeakers. Instead, the images of the instruments and the space around them defined a soundstage that was stable yet completely unlocalized from the speaker positions.

The Sophia's high frequencies were smooth and grain-free, though perhaps not featuring as much top-octave air as the similarly priced Revel Ultima Studio. When I mixed the master for Mosaic, my recently issued CD of the Brahms and Mozart clarinet quintets (Stereophile STPH015-2), choosing the correct top-octave balance for the final mix was an arduous task. For the main pickup, I had used ½" omni mikes with acoustic equalizers over their capsules to boost the on-axis response in the top octave. By applying a complementary cut above 10kHz in the mix, I would suppress any HF grain by the same amount. However, for the secondary pickup I used tubed Neumann cardioids with 1" capsules, which roll off above 15kHz or so. So choosing the exact amount of top-octave rolloff to apply was a subjective decision that depended both on the balance between the two pairs of mikes and the speakers used. When I auditioned Mosaic on the Sophias, the balance that had sounded correct on the Revel Studios made me wish I had gone for just a little more energy above 10kHz in the mix.

In fact, my only real criticism of the Sophia, and one that is relatively minor, concerned an octave or so below 10kHz: the speaker's tonal balance was slightly forward in this region. This was not nearly enough to add any hardness to the sound, but the sense of clarity into the perceived soundstage that was aided by this forward balance was offset by a soundstage that was not as deep as I have experienced from the Meridian DSP 8000s (reviewed in November 2001, Vol.24 No.11), for example, or my venerable B&W Silver Signatures.

As I am a bass guitarist, it should come as no surprise that one of the things I am most fussy about is how loudspeakers reproduce my own instrument. In fact, this is partly why I am so prepared to sacrifice low bass: In their attempts to reach down as low in frequency as possible, so many speakers destroy the leading edges of the sound, turning recorded bass guitar into a "puddingy"-sounding instrument instead of one that is struck and strummed percussively. But, like the Sony SS-M9ED, which I reviewed last August (Vol.24 No.8), the Sophia is one the few speakers that gets right the sound of the 1964 (pre-CBS) Fender Precision Bass that I had used to prepare the channel identification and speaker phasing tracks on Stereophile's Test CD 2 (STPH004-2).

When he had set up the speakers in my room, Peter McGrath had brought with him a recording that I have been unsuccessfully trying to get hold of for years, by bassist Brian Bromberg, brother of guitarist Dave. On the solo cut "My Bass," Bromberg offers a compendium to modern electric bass technique that a mundane player like me can only marvel at. He plucks, taps, slaps, and double-stops in a technical tour de force of chords and counterpoint, hammers-on and harmonics, that makes it hard to remember that the instrument has but four strings and the player but two hands. Via the Sophias, there is no sense that the sound of Bromberg's bass is emanating from loudspeakers. Instead, there is a stable, palpable sense of a sound source hanging in the space between them. And when Bromberg punctuates his music with low open Es, the Sophia's excellent LF extension allowed the speaker to fill my room with sound. Awesome stuff.

I finished my listening sessions with another bass player, the inimitable Marcus Miller, on Miles Davis' We Want Miles (Columbia 469402 2). The two iterations of "Jean-Pierre" state the simplest motif you can imagine, a child's tune, and take it for a ramble. I have owned this album on LP pretty much since it came out, and had been put off by the CD reissue, which had always sounded bright, hard, and messy. But on the Sophias, the CD started making sense. The music-making on this live album is all about the spaces left between the notes by the musicians, and on ordinary speakers those spaces are diminished by grunge. Via the Sophias, even when drummer Al Foster is riding a splash cymbal that would otherwise fill in all the gaps on lesser speakers, the individual acoustic objects are sufficiently well-differentiated that the music continues to communicate. Just as it does in real life.

Conclusion
At $11,700/pair, the Sophia is the best value for money speaker to come from Wilson Audio, but there will probably be shocked intakes of breath at Wilson's Utah HQ when David Wilson and his team read that I think the Sophia might well be Wilson's best speaker to date. "Best," not in the sense that the Sophia exceeds all their other models in every area—it clearly doesn't—but "best" when considered as an integrated package at its price, where its performance in each area optimally balances those in other areas. In that sense the Sophia raises the bar for the twice-the-price WATT/Puppy, which is probably why that system was scheduled to be relaunched in its "version 7" reincarnation at Home Entertainment 2002, still two weeks away as I write this review.

I enjoyed my time with the Sophias immensely and was sorry to see them leave my listening room. I wanted to hang on to them for a little longer, but the review pair had been purchased by Stereophile publisher John Gourlay. I envy him the musical enjoyment he is going to get from this loudspeaker. Perhaps he'll invite me over.

Description: Three-way, floorstanding, reflex-loaded loudspeaker. Drive-units: 1" (25mm) inverted titanium-dome tweeter, 7" (178mm) paper-cone midrange unit, 10" (254mm) aluminum-cone woofer. Frequency response: 29Hz-22.5kHz, +0/-3dB. Sensitivity: 89dB/2.83V/m. Impedance: 4 ohms nominal, 3 ohms minimum. Recommended power: >12Wpc.
Dimensions: 41" (1042mm) H by 12" (305mm) W by 18" (457mm) D. Weight: 160 lbs (72.7kg) each.
Finishes: High-gloss automotive paint.
Serial numbers of units reviewed: 0073 & 0074.
Manufacturer: Wilson Audio Specialties Inc., 2233 Mountain Vista Lane, Provo, UT 84606. Tel: (801) 377-2233. Fax: (801) 377-2282. Web: http://www.wilsonaudio.com/.

Digital sources: Mark Levinson No.31.5 CD transport; Mark Levinson No.30.6, Chord DAC64 D/A processors; dCS 972 upsampler; Meridian 800 DVD-V/CD/CD-R player; Technics DVD-A10 DVD-A player; Musical Fidelity Nu-Vista 3D CD player; Macintosh 8100/80 fitted with two Sonic Solutions SSP-3 digital audio workstation cards; Dell 866MHz Pentium III fitted with RME Digi96/8 Pro and Digital Audio Labs CardDeluxe soundcards, running Windows Me, WinAmp 2.5, SoundForge 5.0, Acid 2.0, and CoolEdit 2000.
Preamplification: Mark Levinson No.380S, Z-Systems rdp-1 digital control center (updated to handle 96kHz sources).
Power amplifiers: Mark Levinson No.33H, Theta Citadel, Halcro dm58 monoblocks.
Cables: Datalinks: Kimber Illuminations Orchid AES/EBU, AudioQuest SVD-4 S/PDIF. Interconnect: Madrigal CZ Gel-1, balanced. Speaker: Synergistic Research Designer's Reference². AC: Synergistic Research Designer's Reference², PS Audio Lab Cable.
Accessories: PS Audio Power Plant 300 at 90Hz (preamps only), Audio Power Industries 116 Mk.II and PE-1 AC line conditioners (not power amps), ASC Tube Traps, RPG Abffusors.—John Atkinson

Other than impedance, for which I used an Audio Precision System One, all acoustic measurements were made with the DRA Labs MLSSA system and a calibrated B&K 4006 microphone. To minimize reflections from the test setup, the measuring microphone is flush-mounted inside the end of a long tube. Reflections of the speaker sound from the mike stand and its hardware will be sufficiently delayed not to affect the measurement.

My estimate of the Sophia's voltage sensitivity came in a little below the specified figure, at 88.3dB(B)/2.83V/m, but the difference is negligible. The speaker's impedance drops to a minimum value of 3.26 ohms at 220Hz (fig.1), but the electrical phase angle is very low at the same frequency, meaning that a good 4-ohm-rated amplifier will have no problem driving the speaker. The phase angle does reach a fairly high value of 40 degrees, capacitive, in the midbass, but the impedance is high enough in the same region to ameliorate any drive difficulty.

Fig.1 Wilson Sophia, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

Other than a wrinkle at the tweeter's ultrasonic resonance frequency, the traces in fig.1 are free from the glitches that would indicate the presence of mechanical resonances in the enclosure. I investigated the panels' vibrational behavior with a plastic-tape accelerometer and found little that could be considered sonically significant. As shown by fig.2, a waterfall plot calculated from the accelerometer's output when it was attached to the back panel, there were two modes present, at 360Hz and 539Hz, but these are low enough in level and high enough in frequency to be not worth worrying about.

Fig.2 Wilson Sophia, cumulative spectral-decay plot calculated from the output of an accelerometer fastened to the cabinet's rear panel level with the midrange/woofer transition. (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz.)

The saddle centered at 26Hz in the magnitude trace indicates the tuning frequency of the big, rear-facing port, which in turn implies good low-frequency extension. The colored traces in fig.3 show the nearfield responses of the lower port (green), the woofer (red), and the midrange unit (blue). (I haven't shown the response of the upper port, as it appeared to be a clone of the midrange unit's output but lower in level.) The minimum-motion point in the woofer's output is a little higher in frequency than I expected from the impedance graph, and, most unusually, the midrange unit has a notch in its output at almost the same frequency. (This is well down in level, however, due to the action of the crossover.) The port itself covers the bandpass between 18Hz and 70Hz, but its output is a little suppressed compared with the woofer, which peaks up sharply between 50Hz and 100Hz.

Fig.3 Wilson Sophia, anechoic response on-axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with the complex sum of the nearfield woofer and port responses (black), the nearfield midrange response (blue), the nearfield woofer response (red), and the nearfield port response (green) plotted below 300Hz, 500Hz, 1kHz, and 500Hz, respectively.

Because of this peakiness, the Sophia's overall output in the bass (fig.3, black trace below 300Hz) is boosted somewhat in this midbass octave. This woofer tuning will lead to problems in some rooms—Martin Colloms noted a problem with it in his review of the Sophia in the May 2002 issue of Hi-Fi News—but it was not an issue in my room, which has a general lack of energy in the 63Hz band. Fig.3 also looks a little worse than it should because of the 3dB boost given low frequencies by the nearfield measuring environment compared with a true anechoic measurement.

The acoustic crossover frequency between the midrange unit and the woofer appears to be set at around 150Hz. The inverted-dome tweeter has a small rise apparent above 15kHz, but rolls off sharply above the audioband, while the speaker's overall response in the midrange and treble is basically flat, though broken up by small peaks and dips. As these tend to be equally spaced, with peaks balanced by dips, the perceived balance will tend to be neutral.

Aiding this is the fact that dips tend to fill in and the peaks become suppressed to the speaker's sides (fig.4). In typical rooms, therefore, the Sophia will tend to sound evenly balanced in the highs. In the vertical plane (fig.5), the speaker's balance doesn't change much as long as the listener sits with his or her ears below the top of the enclosure. Stand, however, and a large suckout appears at the upper crossover frequency, which appears to be just under 2kHz.

Fig.4 Wilson Sophia, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90 degrees-5 degrees off-axis, reference response, differences in response 5 degrees-90 degrees off-axis.

Fig.5 Wilson Sophia, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 15 degrees$bn5 degrees above axis, reference response, differences in response 5 degrees-10 degrees below axis.

In-room, the Sophia's spatially averaged response (fig.6) is impressively smooth and flat, though a slight excess of presence-region energy is apparent at the bottom of the tweeter's passband. I suspect that this contributes both to the speaker's clarity and to its impatience with less-than-stellar recorded balances. The slightly sweet top octave I noted in my auditioning is probably associated with the tweeter's limited dispersion above 12kHz. Note that this graph reveals the Sophia to offer superb low-frequency extension in my room, which does offer good support in the lowest audio octave, and that the nearfield excess in the Wilson's midbass is not apparent.

Fig.6 Wilson Sophia, spatially averaged, 1/3-octave, freefield response in JA's listening room.

In the time domain, the Sophia's step response (fig.7) reveals that its tweeter and woofer are connected in positive acoustic polarity, the midrange in inverted polarity—which is what is needed, in conjunction with the phase shift provided by the crossover, to ensure that the outputs of the drive-units add to give a flat response in the farfield in the crossover regions.

Fig.7 Wilson Sophia, on-axis step response at 50" (5ms time window, 30kHz bandwidth).

The cumulative spectral-decay plot (fig.8) is generally clean. Though some delayed energy is apparent through the presence region, it doesn't appear as the ridges typical of resonant behavior, which suggests that it might instead be due to reflections.

Fig.8 Wilson Sophia, cumulative spectral-decay plot at 50" (0.15ms risetime).

All in all, this is excellent measured performance for which no apologies need be made, though the woofer alignment suggests that the Sophia will work better in some rooms than others.—John Atkinson