Wilson Audio Specialties CUB loudspeaker Measurements

Sidebar 3: Measurements

John Atkinson performed the measurements on the Wilson CUB. Because the loudspeakers were initially reviewed for Stereophile Guide to Home Theater, I was unavoidably familiar with these measurements when I performed the final listening tests for the present review, though not when I made the observations for SGHT, which I summarize here.

The Wilson's sensitivity measured 93dB/W/m (B-weighted), a very high figure. Its impedance is shown in fig.1. The enclosure is tuned to just above 50Hz, with the magnitude plot displaying the saddle shape typical of ported enclosures. The impedance drops down to around 3 ohms over the 200Hz–1kHz region. A good deal of the power in typical program material falls in this range, which suggests that the CUB will not be particularly easy to drive. While any competent high-end amplifier of the sort likely to be used with a loudspeaker such as this should have no problem, I still recommend listening to the CUB with the intended amplifier before purchase. The small ripples visible at 16kHz and 22kHz are due to the resonances in the metal-dome tweeter.

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

'Fig.2 (top curve at the 100Hz point) shows the CUB's FFT-derived response on the high-frequency axis and averaged across a 30º lateral window, combined with the complex sum of the nearfield responses of the woofer and port. (The low-frequency nearfield responses of the woofer alone and of the port alone are shown in the lower curves.) This full-range plot explains a great deal about the CUB's audible performance. First of all, the bass holds up very well to just below 60Hz, then drops like a stone. Hence the limited deep-bass response but the lack of any leanness—the low-frequency response holds up strongly to 50Hz, which covers the fundamentals of nearly all instruments. (Many loudspeakers with limited very-deep-bass response taper off gradually, starting at a much higher frequency.)

Fig.2 Wilson CUB anechoic response on tweeter axis at 50", averaged across 30º horizontal window and corrected for microphone response, with the nearfield woofer and port responses and their complex sum plotted below 300Hz, 1kHz, and 300Hz, respectively.

Further up the range, the entire upper-midrange/lower-treble region is prominent enough to produce much of the presence I heard—and the occasional edge or bite at high levels—but not enough to cause a serious skewing of the overall balance. The most notable aberrations occur above 4kHz. The response is down across most of this region, generally not more than 3dB, but notably more so in a narrow but steep dip centered on 6kHz. I suspect (though cannot verify) that this dip is caused by diffraction from the cabinet structure around the recessed tweeter. Taken in total, the lowered high-frequency response could well explain the lack of "air" I heard from the CUB, though the problem clearly extends down into the mid-treble, not just the high treble as I speculated in the auditioning comments. While there are indications that the 6kHz dip is less significant at normal listening distances than it might first appear (the measurements are made from just beyond 4'), I suspect it is not totally innocuous.

Finally, the peaks due to the metal tweeter resonances are visible above 15kHz. While they look nasty, and while one of them is in the "audible" range, in practice they are narrow (high Q), and high enough in frequency that they're unlikely to cause any clearly identifiable treble coloration. The overall response, while not as smooth as many we have measured, is within ±3dB over most of the audible range (if you accept that the dip at 6kHz will be more benign at the listening chair).

Fig.3 shows the CUB's vertical response family, plotted relative to the on-axis response. (The curves are normalized to the on-axis response, with the remaining curves showing the changes as the listener moves off-axis.) Note that the response becomes quite rough at points significantly off the main vertical axis. You should take Wilson's stand-height recommendation seriously. This response also verifies an observation I made during my home-theater testing: Used horizontally as a center channel, where the vertical dispersion in fig.3 then becomes the horizontal dispersion, the CUB's balance is erratic as one moves from the center to the sides of the listening area. In a home theater, the CUB is at its best in installations where the center channel can be positioned vertically.

Fig.3 Wilson CUB, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 45º–5º above-axis; reference response; differences in response 5º$45º below-axis.

Fig.4 shows the horizontal dispersion, again normalized to the tweeter-axis response. This is far smoother than the vertical response. There is a slight excess in off-axis energy at about 5kHz, not far from the bottom of the 6kHz dip in the averaged front response. Thus the overall room response in this region will probably be flatter than that shown in fig.2. A lot will depend on the loudspeaker position and the characteristics of the room. I generally like to see a loudspeaker that has a smooth response in the averaged frontal region and well off-axis. Such a loudspeaker should, in theory, be the least room- and placement-sensitive (at least above the bass region).

Fig.4 Wilson CUB, horizontal response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90º–5º off-axis; reference response; differences in response 5º–90º off-axis.

Fig.5 shows the CUB's step response on its tweeter axis, which indicates that the CUB is not time-coherent. This is not a serious shortcoming—very few loudspeakers are time-coherent, and then only at a specific point in space—but one then wonders about the need for the tweeter's step-back. The woofer also appears to be connected in opposite polarity to the tweeter.

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

The cumulative spectral-decay or "waterfall" plot is shown in fig.6. The response here is very clean, with little or no treble hash. There is a noticeable resonance around 3kHz—right where the ear is most sensitive—that might contribute to some of the minor colorations heard.

Fig.6 Wilson CUB, cumulative spectral-decay plot at 50" (0.15ms risetime).

Finally, the CUB's cabinet resonances, as measured on one of its side panels, are shown in fig.7. There is a single small resonance centered at 410Hz, and that is well damped. This is an excellent result, verifying the effectiveness of the Wilson's solid enclosure. A similar measurement taken on the top panel (not shown) is even more innocuous.

Fig.7 Wilson CUB, cumulative spectral-decay plot of accelerometer output fastened to cabinet side. (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz.)

Altogether, this is a good but not exceptional set of measurements. We still do not, of course, have all the answers when it comes to measurements vs listening, particularly with regard to loudspeakers. Nevertheless, the strengths and weaknesses I heard while auditioning the Wilson Audio CUB are explained by these measurements far more convincingly than usual.—Thomas J. Norton

Wilson Audio Specialties
2233 Mountain Vista Lane
Provo, UT 84606
(801) 377-2233