Floor Loudspeaker Reviews

Sidebar 1: Measurements

Looking first at the NEAR-50M's electrical impedance and phase (fig.1), the valley at 30Hz reveals the port tuning. Overall, the impedance is more like that of a 5 ohm than an 8 ohm speaker, with a precipitous drop almost to 2 ohms centered at 8kHz. (I assume the specified impedance minimum of 5 ohms is a typographical mistake.) The '50M will be a hard speaker to drive; I find it surprising that DO and other reviewers have used tube amplifiers with it, which will tend to reduce energy around this mid-treble minimum due to the interaction between the amplifier's output impedance and the impedance of the speaker.

Fig.2 shows the crossover between the woofer (measured at 45") and the midrange unit (measured in the nearfield). Ignore the suckout in this plot at 3kHz, which is due to interference from the close microphone position. The metal-cone woofer shows what is probably some breakup around 1kHz, but this is well-suppressed by the sensible choice of crossover frequency and slope. The midrange driver appears to be a honey, with a basically flat response in its bandpass, rolling out above 4kHz, as specified. Looking at the impulse responses of the two drivers reveals them to be connected with the same acoustic polarity.

The 4kHz crossover frequency might be thought quite high for a 4"-diameter drive-unit. Actually, the radiating diameter of the unit is 2.5". Nevertheless, I looked at how the 50M's response changes as the listener moves to the speaker's side. As can be seen from fig.3 (which just shows the changes—the on-axis response therefore appears as a straight line), the midrange unit does become quite directional at the top of its passband, leading to an off-axis mid-treble peak due to the tweeter basically behaving in a semi-omnidirectional manner at the bottom of its passband. The HF unit's directionality increases at the top of its passband, leading to the top-octave rolloff seen to the side of the main axis. Like Dick, I was bothered by changes in subjective balance as I moved slightly to one side while taking the measurements to generate fig.3. And on dual-mono pink noise, there was some "vertical Venetian blinding" noticeable (footnote 1).

In the vertical plane, there is evidence of the lack of tweeter/midrange integration mentioned by DO. The straight line in fig.4 represents the assumption that the response on the midrange axis (39" from the floor) is flat; the other lines are the changes as the listener moves up or down. Basically what this graph shows is that the sound changes drastically for listeners who move above or below that 39"-high axis. To the right of fig.5 is the quasi-anechoic response measured on that axis using a Mark Levinson No.27.5 solid-state amplifier; here is DO's suckout in the NEAR-50M's crossover region. Referring back to fig.4 reveals that this suckout is worse on the midrange axis and only fills in for standing listeners or those sitting with their ears 1' above the ground, both unrealistic listening axes. On the midrange axis, I found the oboe to lose its upper partials, sounding more like a small soprano saxophone, while the violin lost some its sheen.

The presence of this suckout could also be heard on the LEDR diagnostic tracks on the Chesky Test CD (JD37). While the lateral and depth imaging tests were well-reproduced, the Up and Over tracks were quite unstable, due both to the speaker's treble tonal-balance modification of these psychoacoustically processed sounds, and the direction of the main response lobe toward the listening-room ceiling to generate a stronger than normal reflection.

Looking at the electrical responses of the networks seems to indicate a quite large region of overlap for the tweeter and midrange unit, from about 2.5kHz to 6kHz, the exact region where there is a lack of on-axis energy in fig.5. The two units appear to be connected with the same polarity, but the phase shift due to the crossover filters presumably means that they are out of phase through the region where they overlap.

Fundamentally, despite NEAR's use of what appear to be an excellent midrange unit and tweeter, the crossover between them appears to be suboptimally engineered in that the main response lobes point up and down, not straight at the listener. The result will be a laid-back, over-mellow tonal balance that leaves the upper treble sounding isolated and dry on-axis. I found this suckout also lent pink noise a distinctly hollow-sounding character, while the sound was ultimately uninvolving, despite the speaker's excellent upper-midrange clarity, due to the lack of on-axis presence-region energy. Both speakers of the pair measured alike, and as two other reviewers of the NEAR-50M have commented on the presence of this mid-treble suckout (Michael Fremer in TAS #77, February/March '92, and Julian Hirsch in Stereo Review, September '92), it is not a sample-specific problem but a design flaw.

A simple fix for this problem might be to connect the tweeter in inverse polarity, but that is not an easy option for the speaker's owner. And it might compromise the speaker's time-domain performance. In addition, fig.1 reveals that using a typical tube amplifier, with its high output impedance, will exaggerate the depth of this on-axis suckout.

Looking again at fig.5 and ignoring the crossover suckout, the '50's basic balance is pretty flat. There are virtues in this design. To the left of fig.5 are shown the nearfield responses of the woofer and port, the latter's peak at 30Hz confirming the tuning implied in fig.1. In a typical listening room, this speaker will offer quite respectable bass extension, which was confirmed by the in-room response (fig.6). Note the excellent octave–octave balance in the midrange and treble of this graph—other than a peakiness in the low treble which will render the sound rather bright in live listening rooms. (This might also tie in with Dick's mention of grittiness in the low treble.) Despite the presence-region suckout on-axis, the room energy is well-balanced, suggesting that listeners with very large, reverberant rooms who sit a long way away from the speakers will get a much better impression than those who sit close to them in smallish, well-damped rooms.

Though the low-frequency peaks in fig.6 are room effects which haven't been minimized by the spatial averaging, the '50's low bass did sound to me rather "lumpy" and lacking in definition, despite its good subjective extension. DO was bothered by a leanness in the upper bass, a problem which in three-way designs tends to suggest crossover integration problems between the woofer and the midrange unit (see fig.2). This was also evident in my listening, though it isn't implied by the in-room curve in fig.6. However, I note that the woofer and midrange unit are physically quite far apart (though just still within one wavelength at the 250–300Hz crossover frequency), and it is possible that experimenting with the polarity of the woofer with respect to the midrange—something that is made possible by the bi-wirable terminals—will correct for this leanness. I briefly tried connecting the woofers in inverse polarity; orchestral recordings did acquire more body in the lower midrange.

In the time domain, the NEAR-50M offers a well-behaved impulse response (fig.7) without too much ultrasonic ringing from the tweeter. The step response on the same axis (fig.8) indicates that the drive-units are connected in inverted acoustic polarity.

The "waterfall" plot calculated from the impulse response (fig.9) shows a basically clean treble and upper midrange, marred only by the ubiquitous on-axis suckout. In resonance terms, these are very well-behaved drive-units which will contribute to the excellent sense of clarity in the midrange and above that both Dick and I noted in our auditioning. As expected from a quick knuckle-rap test, however, the large, relatively unbraced cabinet was quite lively (footnote 2). Fig.10 shows a waterfall plot of the front baffle's impulse response, taken with an accelerometer. Several strong resonant modes can be seen, the highest in level lying at 400Hz, and there is a powerful peak at 250Hz (the cursor position).

The presence of these modes would be expected to obscure the speaker's clarity in the lower midrange, something that has been noted in other reviews of this speaker. Not only could I hear an added woodiness to the sound of male spoken voice—J. Gordon Holt on Stereophile's first Test CD, for example, or Richard Lehnert on the second—but I found the resonances to be so strong that they added a sense of false reverberation to musical notes in their vicinity. Pausing the disc while the 250Hz warble-tone track from the second Stereophile Test CD was playing, for example, allowed a distinct woody-sounding "bong" to be heard in the ensuing silence. This is pretty poor cabinet performance in these days of low-cost instrumentation.

While I was carrying out these measurements (which involve driving the speakers with continuous tones at a few watts' level), the midrange drive-unit of the right speaker developed a rattle, implying a possible reliability problem.

All in all, my measurements tend to reinforce Dick's feelings about the NEAR-50M: that excellent drive-units alone do not an excellent loudspeaker make. Lewis Athanas and the rest of the NEAR design team need to devote more of their talents to system design and a less lively enclosure; there is a much better loudspeaker waiting to be unleashed from the same basic ingredients as the '50M, in my opinion.—John Atkinson

---

Footnote 1: This is an interference effect in which the tonal balance changes in a severe but regular manner as the listener moves to one side, giving the illusion that the sound is being heard through a series of vertical bars.

Footnote 2: NEAR claims in their literature that the '50M's cabinet is damped and stiffened, with two chambers (the second chamber presumably being for the midrange unit). When I removed the woofer to examine the crossover networks, however, although I could see a horizontal brace just above the woofer, large areas of both the front baffle and rear panel seemed otherwise unbraced. This would explain why the enclosure sounded so live.