NHT 3.3 loudspeaker
NHT prospered. Along the way, it was acquired by International Jensen, which also owns AR and Advent. But NHT retained its design autonomy (footnote 1). The product line grew, but its most expensive member other than the six-piece VT-1 video theater loudspeaker array—the $1100/pair 2.3A—remained well short of second-mortgage territory. However, Kantor had ideas which he could not realize at this price point. He wanted to further push the concept of designing a loudspeaker whose room performance could be reliably predicted. His objective was not, however, simply to build the best $2000, $3000, or $4000 loudspeaker. It was to build the best loudspeaker he presently knew how to build, a loudspeaker which, by his criteria, could not be further improved in any specific area without compromising its performance in others. Price, he says, was not an overriding concern.
The resulting loudspeaker is the NHT 3.3. The fact that it came out costing $4000 and not $40,000/pair should not go unnoticed, given Ken's objectives. Whether or not he succeeded is his to decide. If someone says, "This is the best I can do at the present state of the art, given the inevitable design tradeoffs," you can hardly argue the point—after all, "Wherever you go, there you are," in the immortal words of Buckaroo Banzai. Given the track record of NHT designs, however, I expected, at the minimum, something special from the 3.3. I was not disappointed. But I'm getting ahead of myself.
This price point is a big step up for NHT; the cabinet alone costs more to build than an entire 2.3a. Glancing at the two large, black blocks jutting out from one end of the Stereophile listening room, I was reminded of no other loudspeaker I've ever seen or heard of. The photo at the beginning of this review will tell you more than any written description ever could. I find the style intriguing, but its WAF (footnote 2) is questionable. The optional wood finishes (light oak or rosewood) might be a bit less technoid-retentive than the metallic-looking black laminate. Beauty is, as ever, in the eye of the beholder.
The four-way 3.3's top three drivers, clustered at the top of the front baffle, were carefully selected to meet the designer's objectives. The tweeter and upper-midrange drivers are made by SEAS, the woofer and bass-midrange by Tonegen, the latter a major Japanese driver manufacturer whose products are widely used by US and European manufacturers. The bass-midrange and upper-midrange drivers are each mounted in isolated chambers. The anti-modal upper-midrange enclosure was designed to minimize colorations in the critical 320Hz-3.5kHz frequency band covered by its driver. The latter's frame has also been specially damped for the same reason. The tweeter is the only metal-dome unit used to date by NHT—Kantor says that it's the first aluminum-dome tweeter he's found that he likes.
The narrow, angled front baffle is the least surprising part of the design. This reflects NHT's "Focused Image Geometry" trademark. FIG is intended to minimize interaural cross-correlation—the acoustic mixing of left and right loudspeaker signals—by angling the baffle by a specific, experimentally determined amount. NHT also claims that the nonparallel baffle reduces internal cabinet reflections and standing waves, minimizes side-wall reflections, and stabilizes the image by time-intensity trading. (That is, as you move off to one side, you move off-axis to the nearer loudspeaker, reducing its audible prominence.) Of course, three of these objectives are accomplished routinely with conventional loudspeakers by simply toeing them in. Though it's possible that NHT's inherent cabinet shape could more likely achieve the geometry required to minimize cross-correlation.
But the real shocker in the 3.3's design is the cabinet's 31" depth. This enables the woofer—mounted on the rear inner side of the cabinet—to be placed near the floor and front-wall boundaries for deep-bass reinforcement. At the same time, the upper-range drivers extend into the room, clear of the wall behind them—the classic minimonitor position. The placement of the woofer in the 3.3 causes it to "see," to an extent, a corner of its own "room" consisting of the floor, the front-wall, and the large side panel of the loudspeaker cabinet. This "room" is, in the nearfield at least, predictable by the designer. The recommended positioning of the 3.3s is flush against the front wall. However, they can be moved out from that wall, if necessary, to "tune" the bass balance to the actual listening room, which can never be made to completely disappear from the equation.
The cabinet itself is thick and heavily braced—attested to by its 123-lb weight. The "knuckle-rap" test on the large side panel produced a slight resonance. But it sounded reasonably well-damped and above the woofer's frequency range, thus unlikely to be easily excited. Stabilizer feet are provided and should definitely be used. By itself, the tall, narrow cabinet lacks reassuring lateral stability under normal household use, particularly on a carpeted floor.
The 3.3's 26-element crossover network is not unduly complex, though impedance compensation is used for the tweeter and upper-midrange drivers. It is configured for bi-wiring, if desired. Crossover components are chosen for their specific functions in the circuit, not on the supposed superiority of a particular "type" of capacitor or coil. Kantor believes in using the best parts where they count—eg, premium capacitors in the series feed to the tweeter—and in keeping costs down by using less expensive parts where they are unlikely to have any sonic impact. It is common high-end practice to use the "best" parts everywhere, on the theory that it can't hurt. And if the chosen parts have no adverse side effects, it won't. But by the time all multiplying factors are taken into account (footnote 3), it will have significantly jacked up the retail price—though not necessarily the level of performance.
The NHTs were auditioned in a system consisting primarily of the Mark Levinson No.31 transport and No.35 D/A processor, Rowland Consummate preamp, and Krell KSA-250 power amplifier. A Kimber AGDL digital link (S/PDIF) connected the transport and processor. TARA Labs Master RSC linked the processor and preamp, and Cardas Hexlink ran between preamp and amp. Straight Wire Virtuoso loudspeaker cables (bi-wire) tied the Krell to the NHTs.
I was concerned at first that the layout of Stereophile's listening room would cause a problem with the setup of the 3.3s, as our normal (and only practical) setup arrangement is not ideal for the 3.3s' recommended positioning. The wall behind the loudspeakers has diagonally cut-off corners and a long block-and-plaster seat beneath a large, centrally located window that extends to create the hearth for a small, partially blocked-off fireplace in the left-hand corner. (There is a complete description of this room in Vol.14 No.10, including a diagram and photos which clearly show what I've just described.) The window seat, in particular, makes it impossible to place the backs of the 3.3s flush with the front wall.
Not to worry. Ken Kantor was on hand in the initial setup stages to familiarize us with the 3.3s. During his visit we determined that placing them out from the banco (Santa Fe-speak for window seat) by several inches actually worked better in this room than placing them as far back as possible. After Ken left, I determined that moving the 3.3s even farther out into the room—their backs now a good foot from the banco—improved matters even more. This may not be the case in all rooms, but it does indicate that experimentation with setup can pay off. Start with the manufacturer's recommendations (NHT's manual is brief but quite thorough), and work from there to fine-tune loudspeaker/wall proximity, loudspeaker spacing, and listening distance. The angled front baffles make the last two especially important in optimizing NHT's intended "Focused Image Geometry."
NHT states that the macro placement of the 3.3s—ie, where they are located in the listening room—is relatively uncritical, but that their micro positioning is very important. The latter simply means that, in addition to fine-tuning the loudspeakers' spacing, their perpendicularity to the wall behind and, in particular, their exact parallel spacing, are important. (When Ken Kantor, one of the most objectively oriented and non-tweaky designers I know of, says that a fraction of an inch is important, you pay attention.) We carefully set up the 3.3s when he was here. When I had to temporarily move them out of position, I took similar precautions in setting them up again to make certain everything was squared away before getting down to serious listening. Experimentation with laterally spacing the 3.3s is also worthwhile in focusing the image. But in my listening room I found this effect to be less dramatic than optimizing the listening distance. The latter, of course, accomplishes the same thing but changes other variables.
I had no problem achieving the correct vertical listening axis. It is just above the tweeter, which puts it at 38" or 39"—almost exactly ear-height in my listening seat.
Most loudspeakers cop out when it comes to full-range performance. They do the top end moderately well—there are plenty of good tweeters around these days. While the totally neutral midrange has yet to be achieved, a reasonable facsimile is not that hard to find. There are still some quackers out there, but they no longer corner the market. But when you get to the bottom end, you're often talking outboard subwoofers if you really want to plumb the depths.
Footnote 1: At the time of the review, Ken Kantor was VP of Research and Development for both NHT and the parent corporation. As of 2000, he became a consultant. As of 2004, NHT is now part of the Rockford Corporation.—John Atkinson
Footnote 2: Wife Acceptance Factor.—Thomas J. Norton
Footnote 3: I claim no expertise in determining marketing costs, but the figures I hear mentioned most often are that retail is 3–5 times parts costs. Throw in another $50 or more for premium, low-production-run crossover parts, and you've increased the retail price by up to $250.—Thomas J. Norton