New England Audio Resource NEAR-50M loudspeaker
I've said it before: Much of the coloration introduced by dynamic drivers is due to in-band resonances and their associated harmonic distortions. The frequency-response curve may look reasonably flat, with only a few wiggles here and there, but that in itself does not tell the whole story. When a cone breaks up, its various sections may move out of phase relative to one another so that their output cancels out and produces little sound radiation at the fundamental frequency. However, it does not follow that significant cancellation of the harmonics takes place. While the frequency-response wiggles may look benign, the audibility of the dissonant breakup harmonics may be significant.
Metal diaphragms have other advantages. Sensitivity to environmental effects is virtually eliminated. Because aluminum is non-hygroscopic, a metal cone cannot absorb moisture. Unless completely waterproofed, a paper cone's performance is affected as the cone is softened by moisture absorption; and the impact of many wet/dry cycles weakens the diaphragm. The promise of metal is repeatable sonic performance for many years. And, because metal is a good conductor of heat, the diaphragm can heatsink the voice-coil, thereby improving the thermal rating and, very likely, the driver's reliability.
On the debit side, metal cones tend to be much heavier than paper or plastic and thus typically result in less sensitive drivers. Metal also has very little internal damping: when metal cones break up, they do so screaming with a high-Q resonance. It's important, therefore, to use metal drivers well within their comfort zones. NEAR uses a fourth-order linear-phase network to roll off the woofer just below 300Hz. Such a network offers a slower descent into the stopband around the crossover point, thereby improving transient response at the cost of sacrificing some selectivity. A mix of first- and fourth-order networks (footnote 2) is used to blend the midrange and tweeterthis the well-known Vifa 1" metal-dome unitat about 4kHz (footnote 3).
Equipment reports occasionally undergo exceptionally long periods of gestation. Such was the case with the '50M. The review process unfolded in several chapters, and since all of these episodes are germane to this story, you shall be told about them. It is Stereophile's policy to describe all experiences with a particular product. For example, the fact that the first sample of a particular product blew up should be of great interest to you, reflecting as it does the product's reliability. You, of all people, have a right to know about it. In this case, it's important for you to know that this product continued to evolve during the review process, and that manufacturing mistakes were made. After all, you could have ended up with any of the three samples of the '50M that I auditioned.
The first samples arrived nearly two years ago, shortly after the 1991 Winter CES. A brief audition revealed serious problems with driver integration. Small changes in toe-in angle produced major alterations in tonal balance. I put the speakers aside for several months, and, as is often the case in the high-end business, NEAR called in May 1991 to say that the '50M had been upgraded to include new crossover parts, wire, and a generally more refined tuning of the system.
The second set of samples (SNs 014021 & 22), arriving in June 1991, caught me in the midst of moving out of Stereophile's dedicated listening room and setting up a whole new listening environment. I thus didn't commence my audition of them until fall 1991. By December I'd reached a set of final impressions, the gist of which could be titled "A Tale of Three Drivers."
The treble was quick and nicely controlled. The core of the midrange was very smooth and spatially convincing. The decay of hall ambient information was easy to resolve, thanks to a consistently transparent soundstage. What bass there was the woofer enunciated very cleanlythe level of bass definition was remarkable indeed.
But the transition region between drivers wasn't right. The upper bass was too lean, robbing the double bass of body. The upper registers of soprano voice were dulled, as were violin overtones. Not only were the sheen and brilliance that reside in the upper mids greatly compromised, the lower treble was rendered in a grainy manner all the more disconcerting in view of the smoothness of the midrange. The transition between the midrange driver and tweeter irritated me the most. The tonal-balance anemia through the orchestra's power range was no asset, to be sure, but certainly sounded no worse than that dished out by a host of minimonitors. However, the dry, dull upper mids, coupled with the gritty lower treble, were much more difficult to live with.
The set of frequency-response measurements I ran to get a technical handle on these speakers was a real eye-opener. The most significant finding was a large discontinuity between the midrange driver and tweeter. I think that you would agree that a 6dB suckout centered between 3 and 4kHz would be clearly audible. Such a major alteration of the harmonic envelope neatly explained my sonic impressions, but also raised the possibilities of a design error or manufacturing defect. Since both channels measured identically, the possibility of a design error was more credible. Still, I decided to give NEAR the benefit of the doubt, and requested another sample to evaluate.
I was told that some mistakes had been made in a batch of crossover networks manufactured around June 1991, and that while most of these rogue networks had been caught before they left the factory, "a couple snuck out": my review pair, and the pair sent to the The Abso!ute Sound (footnote 4).
Apparently, another major product revision, introduced at the Winter CES in January 1992, gave the '50M an "entirely new and more extended bass crossover." I waited for a current production sample for several months, but only after a discussion of the matter with NEAR's Lewis Athanas did the final sample (SNs E015263 & 64) finally arrive in late May 1992.
Yet another surprise. The first thing I did was to measure the '50M's frequency response to determine just how the "correct" crossover network integrated the drivers. To my surprise, both the old and new samples measured essentially alike at 1m. All of my measurements so far had been conducted with my Neutrik System 3300, but those measurements agreed with those performed with a recently acquired ATI Loudspeaker Measurement System (LMS). While I don't pretend to claim that it's possible to fully characterize loudspeaker performance with a simple figure of merit or a few response measurements, it has been demonstrated that the nearfield on-axis response correlates well with listening impressionsat least in well-treated rooms and in a listening configuration optimized for imaging precision.
Nor was there much sonic news to report. Lesley's vocalsas, for example, on "Jazz Me" from Lesley (ViTaL 011)lacked upper-register sweetness and smoothness. What should have been a gloriously sweet ride turned dry and grainy. Joni Mitchell (Blue, Reprise MS 2038) didn't have it either. Through the 50M, she sounded like she'd been gargling all day with mouthwash.
Soundstage transparency and focus were excellent to the point that massed voices were convincingly resolved. There was no difficulty in pinpointing individual voices during Walton's Belshazzar's Feast (EMI SAN-324). The chorus floated in space without becoming homogenized or smeared. The foundation of the LSO was sketched out with tight definition, but I was unsatisfied with the tonal balance; the upper bass and lower midrange still lacked adequate heft and body.
There was always lots of midrange and treble detail; the finesse detail so often masked or fuzzed over by cheap drivers was readily elucidated. Transients were also well-articulated, though they typically conveyed a slight metallic aftertaste. The artificial treble flavor may not necessarily indicate a time-domain deficiency (ie, ringing), but may simply be a perceptual effect caused by this speaker's frequency-response aberration. It was the BBC's D.E.L. Shorter who, some 35 years ago, called attention to this effect. (Very few folks ever knew more about loudspeakers than Mr. Shorter.) If a progressive decline in response with increasing frequency is followed by an increase (ie, a suckout), the treble will be heard to stand out in unnatural reliefeven though the treble response nowhere rises above the midband level. A loudspeaker that rolls off uniformly with increasing frequency mimics our everyday experience (eg, sound that has traveled around a corner); if anything, such a speaker ought to sound quite natural. There is, however, no natural mechanism whereby treble is selectively reinforced in the manner of the '50M.
To be fair, there were isolated records that sounded quite good. Muddy Waters, Folk Singer (Chess CH-9261), for example, sounded just fine, while Buddy Guy's guitar work was amazingly well resolved. The core of the midrange was illuminated with great clarity. Bass lines were lightweight, but otherwise there was much sonic virtue to celebrate here. Each driver appeared to excel over a particular bandwidth. However, the whole remained less than the sum of its parts.
If timbral accuracy and tonal-balance realism are low on your totem pole of sonic priorities, then the NEAR-50M may indeed be worth a listen. Its level of spatial and detail resolution may win you over. Nevertheless, the '50M strikes me as not yet quite ready for prime time.
Based on the above, I can only conclude that NEAR is guilty of suboptimal system engineering. That the NEAR metal drivers are capable of excellent sonic performance is quite evident. I hereby raise my glass in salute to NEAR's contribution to the art. Unfortunately, drivers alone do not an excellent loudspeaker make. Although they're essential ingredients by definition, sound engineering to assure proper driver integration is also part of the recipe for success.
Folks, there's gold in them thar aluminum drivers. The potential is there, still waiting to be fully realized.Dick Olsher
As this review went to press, NEAR informed us that the NEAR-50M had been revised yet again, with now a front baffle made from 1" rather than ¾" MDF, a more compliant midrange surround, a slightly different crossover, and higher-quality 5-way binding posts. Although we requested samples of this latest version of the '50 for a "Follow-Up" be sent, this never happened.John Atkinson
Footnote 1: Both drive-units are manufactured using machinery and molds that NEAR inherited from the original Bozak company.Dick Olsher
Footnote 2: I believe it was John Bau, of Spica, who first used a combination of a fourth-order low-pass filter with a first-order high-pass in the Spica TC-50. When optimally configured, this kind of asymmetric crossover gives very good time-domain performance.John Atkinson
Footnote 3: The crossover networks are complex. The bass low-pass filter, which is mounted on a printed circuit board on the cabinet's sidewall behind the woofer, consists of one air-cored and two ferrite-cored inductors, one resistor, and seven capacitors, including one electrolytic and three film caps in parallel. The midrange band-pass and tweeter high-pass filters are mounted on the base of the cabinet, and feature three air-core inductors, four resistors, and seven plastic-film capacitors, four of which are in parallel.Dick Olsher
Footnote 4: See TAS #77, pp.94100.Ed.