Wilson Audio Specialties X-1/Grand SLAMM loudspeaker system Measurements part 2

My inner engineer keeps whispering that flat curves must sound better; but it ain't necessarily so. Judgment of the semi-anechoic laboratory frequency response of a loudspeaker must be tempered by assessment of the performance over a range of forward axes. You need to judge the energy uniformity seen in the forward solid angle which encompasses the listener and the relevant reflective boundaries. The overall energy response as reflected in the Room Averaged Response is also most useful.

A loudspeaker's single axial response cannot, therefore, fully reflect its perceived "sound." Yet occasionally, with untoward combinations of source quality and matching system, the X-1 did, in fact, hint at these response features. Proof that these irregularities are rather less dominant than the axial curve suggests was given in the room-averaged response (see below), where only very mild tonal-balance variations are seen over the critical areas—in close agreement with the subjective results.

Looking at the axial response in detail, the upper midrange is lifted by 3.8dB at 1.8kHz, this classed as a medium "Q" bump, and associated with a broader lift of 1dB from 800Hz to 3kHz. Between 3.5kHz and 11kHz, the treble is subdued by 2dB or so relative to the mean, and then appears a bit peaky at 13kHz.

You might be tempted to suspect that the tweeter is responsible, but you'd be wrong. A spare high-frequency unit, lent to me by Wilson, checked out pretty flat from 3kHz to 25kHz. Those bumps are, in all probability, due to the time-aligned modular cabinet design. In particular, the top module in its normal state of alignment overhangs the baffle area of the tweeter; despite there being absorbing foam at the resulting ledge, some reflection occurs. If the raw response had been a true measure of the energy output, the variations would be considered fairly serious. In fact, it simply shows the effect of a moderate local reflection at this close measuring distance.

Fig.7 also shows the port response with its maximum output at 24Hz. There's also a strong return in output at around 220Hz which, in an ideal world, would be better controlled. The curve marked "B" shows the output of the bass drivers alone, revealing their gentle rolloff above 150Hz, which results in significant acoustic output to 1kHz. Interesting.

On the far right of fig.7 can be seen the response of the rear ambience tweeters, peaking in the top audible octave to counter their natural fall in directivity with rising frequency. They may be given some credit for smoothing the energy response in the room at high frequencies.

Off-axis in the vertical plane—the lower group of traces in fig.8—the central lobe was found to be nicely symmetric, and was accurately directed at the listener in Wilson's recommended state of "tune." Above or below this axis by 15 degrees, the response had notched symmetrically by 12dB or so near the 3kHz crossover frequency. The 7.5 degrees vertical off-axis trace (dotted) was just fine, however. As these curves show, the X-1's focal point is well-defined.

Fig.8 Wilson X-1/Grand SLAMM, horizontal (top) and vertical (bottom) response families at 45", normalized to response on tweeter axis. Dotted curves are -7.5 degrees vertical and 15 degrees horizontal; dashed curves are ±15 degrees vertical and 30 degrees, 45 degrees, and 60 degrees horizontal.

In the lateral plane—the upper group of traces in fig.8—the good conformity and moderate loss at large off-axis angles indicates quite low diffraction and good directivity. This speaker is intended to be auditioned at a glancing angle set by a sight-line along the inside edges of the midrange enclosures, and not on the median driver axis. This is significant: the dotted trace on this graph, 15 degrees off-axis and close to this sight-line, shows an improvement in output in the 6-9kHz treble region just where the axial curve in fig.7 had shown a loss in amplitude.

Room Responses: I took 1/3-octave-smoothed measurements in both of the rooms used for the listening sessions, with particular emphasis on the X-1/Grand SLAMM's interaction with the room. In my room (fig.9), two results are given, one for the theoretical best placement (solid line), and one for the position chosen for the auditioning (dashed). The wisdom of the preferred listener placement may be seen in the improved low-frequency distribution. Nevertheless, in this room the range below 25Hz did not fully reflect the loudspeaker's intrinsic low-frequency extension.

Fig.9 Wilson X-1/Grand SLAMM, spatially averaged 1/3-octave response in Martin Colloms's listening room.

This was where the second, larger room proved its worth (fig.10). While the overall response character can be seen to be similar—an overall response of 19Hz-16kHz ±3dB was obtained in the large room (a fine result)—the 20Hz level improved from -5dB to -2.5dB. This is a significant difference, given our near-linear hearing response in the bass, and was certainly clearly audible. However, what this graph cannot show is the greater sense of aural freedom and "air" that the larger room also engendered.

Fig.10 Wilson X-1/Grand SLAMM, spatially averaged 1/3-octave response in Ricardo Franassovici's listening room.

Time Domain: Looking at the impulse response (fig.11) with the woofers out of circuit due to problems with room proximity effects, the X-1 was reasonably time-coherent. The output was free from significant ringing within the 100Hz-25kHz measurement bandwidth. The merits of its reduced-diameter drive—a 0.75" coil on a 1" dome—can be seen in the fact that the metal-dome tweeter's "oil-can" resonance is placed well beyond the measuring bandwidth. It's also comparatively well-damped. The step response (fig.12), again with the woofers disconnected, reveals the tweeter to be connected in reverse polarity and leading the midrange units slightly in time.

Fig.11 Wilson X-1/Grand SLAMM, impulse response on tweeter axis at 45" (4ms time window, 30kHz bandwidth).

Fig.12 Wilson X-1/Grand SLAMM, step response on tweeter axis at 45" (5ms time window, 30kHz bandwidth).

Due to the inverted tweeter connection and the use of second-order crossover filters, the X-1's phase response is not linear in the treble, though it does approach this condition in the midrange and low treble. This can be seen in the plot of excess phase vs frequency (fig.13), which is "zeroed" in the midrange. The X-1's departure from a minimum-phase performance was mild (footnote 2), reaching -40 degrees at 5kHz, and smoothly advancing to a phase lead of 80 degrees at 20kHz, which is of no consequence.

Fig.13 Wilson X-1/Grand SLAMM, excess phase on tweeter axis at 45" (20 degrees/vertical div.).

The first waterfall plot (fig.14), with its 30dB amplitude window and a fast 0.1ms risetime for the measuring filter, is optimized to display decay rate. While the presence of the cabinet reflections mentioned earlier can be seen in the upper treble, the overall decay rate was clearly very good, judging by the large "white" area at the back and the clean "floor." This graph correlates very well with the X-1's superb perceived transient response—its amazing clarity on fast, complex percussion.

Fig.14 Wilson X-1/Grand SLAMM, cumulative spectral-decay plot at 45" (0.01ms risetime).

Analyzing the longer-term resonance behavior is the duty of the second waterfall (fig.15), which uses a wide 60dB amplitude window and a more frequency-selective 0.2ms filter setting. Again, the result for the X-1 was commendable. Yes, two treble ridges can be seen, at approximately 14kHz and 20kHz, but they last no more than 1.5ms. Very little long-term resonance hangover or "coloration" effects are present.

Fig.15 Wilson X-1/Grand SLAMM, cumulative spectral-decay plot at 45" (0.2ms risetime).

Enclosure: This excellence extended to the vibration performance of the Grand SLAMM's enclosure. I probed several locations on the three upper-range modules and the bass enclosure with my low-mass B&K accelerometer, but it proved very hard to find anything of significance. The worst part of the largest panel of one of the midrange modules still had a desirably smooth output.

Very little vibrational energy emanated from the bass enclosure over its nominal 20-500Hz working band (fig.16). There was some vibrational output above 900Hz, but this was merely a low-level signal derived from the point contact with the stack of upper modules.—Martin Colloms

Fig.16 Wilson X-1/Grand SLAMM, spectrum of accelerometer output fastened to low-frequency enclosure panel.

Footnote 2: For examples of very poor and very good excess-phase performance, see figs.10 and 11 in my Dunlavy SC-I review elsewhere in this issue.—John Atkinson
Wilson Audio Specialties
2233 Mountain Vista Lane
Provo, UT 84606
(801) 377-2233