Floor Loudspeaker Reviews

Because of the Alexx V's size and bulk, I measured its performance in Jim Austin's listening room. I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the Wilson Audio Alexx V's frequency response in the farfield, an Earthworks QTC-40 microphone for the nearfield and in-room responses, and Dayton Audio's DATS V2 system to measure the speaker's impedance.

Wilson's Peter McGrath had adjusted the positions of the Alexx V's tweeter and midrange modules so that their axes converged on the positions of JCA's ears, which, with him sitting in his IKEA chair, were 34" from the floor and 138" from the top of each Wilson's upper-midrange enclosure. I measured the quasi-anechoic response of the left-hand Alexx V, averaged across a 30° horizontal window centered on the position of his ear. However, the presence of early reflections from the room boundaries means that the FFT-calculated response was not valid below 1kHz or so. We therefore estimated what the microphone height from the floor would be at my usual measurement distance of 50". This was 42", which was level with the lower-midrange module. I took a second set of measurements with MLSSA at that microphone position.

Looking first at the Alexx V's voltage sensitivity, my estimate of 90.8dB(B)/2.83V/m was slightly lower than the specified 92dB. However, Wilson's specification is for 1W at 1m; with its specified impedance of 4 ohms the loudspeaker actually draws 2W from the amplifier with a 2.83V signal. As with the original Alexx, which Michael Fremer reviewed in May 2017, the Alexx V will play loudly with relatively few amplifier watts—or it would if its impedance were not low. My impedance measurement is shown in fig.1. The impedance magnitude (solid trace) remains between 2 and 4 ohms for almost the entire audioband, with a minimum value of 1.9 ohms between 242Hz and 270Hz. The electrical phase angle (dotted trace) is low from the upper bass to the low treble, but there is a combination of 3 ohms magnitude and a phase angle of –39° at 53Hz that exacerbate the drive difficulty.

Using a spreadsheet prepared by Jim Austin, I calculated the effective resistance across the audioband (EPDR, footnote 1) that results from the combination of magnitude and phase angle. The EPDR drops below 2 ohms over most of the midrange and mid-treble regions, with minimum values of 1 ohm between 50Hz and 64Hz, 1.15 ohms between 197Hz and 203Hz, and 1.05 ohms between 3kHz and 3.16kHz. Owners of the Alexx V need to match the speakers with amplifiers that are unfazed by very low impedances.

I listened to all the enclosure panels with a stethoscope while I played the MLSSA test signal. The woofer bin's heroic cabinet construction meant that I couldn't hear any panel resonances. However, some low-level liveliness was audible with the three upper-frequency enclosures—tapping the sidewalls with my knuckles I could hear a faint "plink" with the stethoscope. Investigating this behavior with a plastic-tape accelerometer, I found some modes close to 500Hz with all three enclosures (fig.2). However, the relatively high frequency and Q (Quality Factor) of these modes and the very small radiating areas mean that this behavior will not have audible consequences.

Both woofers cover the same passband and cross over to the lower-midrange unit around 150Hz. That unit, in turn, appears to be crossed over to the upper midrange unit at about 750Hz. The impedance graph suggests that the large port is tuned to 20Hz, and nearfield measurements of the woofers' outputs revealed that they each had a minimum-motion notch close to that frequency. The port's output, again measured in the nearfield, peaked between 15 and 30Hz, but didn't begin its upper-frequency rolloff until 65Hz. The port's output in the midrange was well suppressed, however.

The blue trace below 300Hz in fig.3 shows the complex sum of the nearfield port, woofer, and midrange responses, taking into account their radiating areas, amplitudes, and acoustic phase angles. The small peak in the midbass is probably an artifact of the nearfield measurement technique; the Alexx V's low frequencies don't quite extend to the port tuning frequency, which, like the original Alexx, suggests that the woofer alignment is somewhat overdamped. Higher in frequency, the blue trace shows the Wilson's farfield response averaged across a 30° horizontal window centered on the calculated listening axis at 50". The small excess of energy between 1.3kHz and 2kHz is lower in amplitude than with the Alexx's response (fig.2 in that review), but there is a similar suckout above that region. Again as with the original Alexx, the suckout tends to fill in at the listening position (red trace). Other than the upper-midrange peak, the speaker's balance is relatively even between 100Hz and 20kHz, with small peaks balanced by small dips.

It wasn't possible to lift the Alexx V onto a wheeled dolly to allow it to be rotated for me to examine the loudspeaker's horizontal dispersion in detail. However, the geometry of JCA's room meant that I could move the microphone by 5° intervals up to 45° to the side of the listening axis at 50". It appears that the Alexx V maintains its spectral balance below 15kHz up to ±15° to the sides of the measuring axis and that the on-axis suckout at 3kHz tends to fill in a little off-axis. In the vertical plane, there is a little more energy in this region as you move above the measurement axis, though the suckout deepens below that axis.

I averaged 20 1/10-octave–smoothed spectra, individually taken for the left and right speakers, in a rectangular grid 36" wide by 18" high and centered on the positions of JCA's ears to produce the Alexx V's spatially averaged response (fig.4, red trace). For reference, the blue trace shows the response taken under identical conditions of the Magico A5, which JCA reviewed in July 2021. (I have truncated the Magico's response below 35Hz, because the plot is contaminated by subsonic noise from JCA's apartment's heating/ventilation system, which could not be turned off on the March morning that I performed the A5 measurements.)

On the face of things, the two speakers behave similarly between 200Hz and 10kHz, though the Alexx V has a little less energy in the midrange and a little more low-treble energy. The Magicos have greater output in the upper bass, but both types of speaker excite the low-frequency room resonances to a similar degree. The response of each speaker gently slopes down above 5kHz, which will be due to their tweeters becoming more directional as the frequency increases and to the increased absorption of the room furnishings in the treble. (What you don't want to see with a spatially averaged in-room response is a flat treble output, which will sound excessively bright/shrill.)

In the time domain, the step response on the listening axis at 50" (fig.5) reveals that the tweeter and lower-midrange unit are connected in positive acoustic polarity, the upper-midrange output and woofers in negative polarity. (This is identical to the polarities of the drive-units of the original Alexx, and I confirmed this by examining the individual step responses of the Alexx V drivers.) If you look closely at this graph, you can see a slight discontinuity at 3.75ms, which is due to the upper-midrange unit's output arriving very slightly too late to smoothly blend with the negative-going decay of the tweeter's step. Similarly, the small discontinuity just before 4ms is due to the lower-midrange unit's step arriving slightly after the negative-going decay of the upper-midrange unit's step. But other than those admittedly minor issues, the Alexx V's step response is time-coherent on this axis (footnote 2), implying optimal crossover implementation.

I tried eliminating the first discontinuity by moving the microphone higher, but this moved the lower-midrange unit's output farther back in time. When I lowered the height of the microphone, this resulted in a better blend of the two midrange-unit outputs but worsened the discontinuity between the tweeter and upper-midrange steps.

To be fair, the geometry of the upper-frequency drive-unit outputs had not been optimized for this relatively close microphone distance. Fig.6, therefore, shows the Alexx V's step response at the position of JCA's ears. (Ignore the boundary reflections after 11.5ms in the trace.) The decay of the tweeter's step now blends smoothly with the negative-going upper-midrange step, though there is still a slight discontinuity between the two midrange steps. This suggests that the lower-midrange drive-unit should have been moved forward one notch in the gantry to give optimized time-domain performance.

Finally, the cumulative spectral-decay or waterfall plot at 50" on the listening axis (fig.7) shows a clean initial decay, but as with the original Alexx there is then some low-level hash present throughout the midrange and treble. I suspect that this behavior is due in part to early reflections from the complicated, upper-frequency drive-unit array.

Measuring a loudspeaker as large, as heavy, and as complicated as Wilson Audio's Alexx V without access to an anechoic chamber is problematic. But that smooth, relatively even in-room response is probably what is most important when characterizing its sound quality.—John Atkinson

Footnote 2: I should clear up some readers' confusion about my use of the terms "time-coherent" and "time-coincident" with multiway loudspeakers. The latter means that the outputs of the drive-units arrive at the nominal listening/microphone position at the same time. The step response is therefore a right triangle—a vertical rise from zero with then a slow decay to the timeline. This is very difficult to arrange—the only dynamic speakers I have measured that were truly time-coincident have been various Spicas, Thiels, Dunlavys, and Vandersteens. By "time-coherent," I mean that when the crossover's phase shift in the crossover region and the different distances of the acoustic centers of the drive-units from the listening/microphone position are taken into account, the result is a step response where the decay of each unit's step smoothly blends with the start of the step of the next lower in frequency. To the ear, the difference between perfect time-coincidence and perfect time coherence is relatively minor.