Stand Loudspeaker Reviews

Sidebar 2: Measurements

I use a mixture of nearfield, in-room, and quasi-anechoic FFT measurement techniques (using primarily DRA Labs' MLSSA system with a B&K 4006 microphone calibrated to be flat (footnote 1) when used on-axis up to 30kHz—see "Follow-Up," October 1991, Vol.14 No.10—plus an Audio Control Industrial SA-3050A 1/3-octave spectrum analyzer with its own microphone) to investigate factors that might explain the sound heard. The impedance phase and amplitude of the speakers were measured using the magazine's Audio Precision System One.

The Westlake's impedance magnitude and phase can be seen in fig.1. The tuning of the ports is indicated by the minimum of 4.2 ohms at 39Hz. You can see that the impedance drops to a very low value of 2 ohms throughout the upper bass and lower midrange, rising above 4 ohms only between 1200Hz and 3700Hz. Note also that the phase angle is pretty serious in the upper bass, meaning that the current in this region significantly leads the voltage. As a result, the BBSM-6's effective impedance in this region acts as if it is even lower than 2 ohms. Speaker cables should be very short or of large gauge, therefore, and the BBSM-6F must be driven by an amplifier capable of driving many volts into very low impedances if its dynamics are not to be compromised. This is a shame, as the speaker's high sensitivity—92dB/2.83V/m at 1kHz—might suggest its use with relatively low-powered, high-quality tube amplifiers. However, even if it could source sufficient current, a typical tube amplifier's output impedance of 1 ohm will result in a broad 1.6dB depression between 80Hz and 1kHz, which will exacerbate the BBSM-6's rather lightweight balance.

Fig.1 Westlake BBSM-6F, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

The Westlake's impulse response on the tweeter axis at a measuring distance of 45" is shown in fig.2. Though the Westlake literature claims the speaker's crossover to be phase-compensated with a "coherent wavefront" as the result, this is not borne out by the step response in fig.3, which reveals a highish-order crossover. This is not to say that there won't be a position somewhere in front of the speaker where the outputs of the drivers sum in a time-coincident manner, but it certainly isn't on the tweeter axis. Though it is hard to see on the scale this graph has been printed in the magazine, there is also a succession of reflections of the impulse all the way along the tail (though that just after the 7ms mark is the first room reflection, that from the ceiling). These could possibly be from the edges of the wide baffle.

Fig.2 Westlake BBSM-6F, impulse response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.3 Westlake BBSM-6F, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

To investigate a speaker's frequency-response behavior, I take five impulse responses like that in fig.2 across a 30° horizontal window. Transforming the reflection-free portions of these responses to the frequency domain and averaging gives a good idea of the basic frequency response free from mike-position–dependent interference artifacts. Carrying out this process for the BBSM-6 (with the grille removed) gives the trace shown to the right of fig.4. Generally smooth, there are a few bumps and dips in the lower treble but nothing significant. The soft-dome tweeter starts to roll off above the range of my hearing, being 3dB down at 18kHz.

Fig.4 Westlake BBSM-6F, anechoic response on tweeter axis at 45", averaged across 30° horizontal window and corrected for microphone response.

To the left of fig.5 are shown the nearfield responses of the woofers and ports. (The level matching between each of these and the quasi-anechoic curve above 200Hz can only be approximate.) The maximum output from the twin ports occurs at 40Hz, as expected from fig.1, with the minimum woofer output a little offset from this at 34Hz. All things being equal, which they never are, the Westlake should be capable of giving useful bass output down to 40Hz, the lowest note of the double bass—which is basically what I found to be the case from my listening, though the bass region was shelved down overall.

Fig.5 Westlake BBSM-6F, acoustic crossover on tweeter axis at 45", corrected for microphone response, with the nearfield responses of the woofers and port, plotted in the ratios of the square roots of their radiating areas below 200Hz and 800Hz, respectively.

The individual responses of the BBSM-6's woofer and mid/treble sections are also shown in fig.5. The midrange rolls off in classic 24dB/octave manner below 600Hz. The woofer, however, shows a little bit of character around 500Hz before it rolls out. Also, for curiosity's sake, I looked at the tweeter-axis response with and without the grille. The difference the grille makes is shown in fig.6, which confirms Westlake's suggestion that the grille be left off for serious auditioning.

Fig.6 Westlake BBSM-6F, differnece made to anechoic response on tweeter axis at 45" made by grille.

When a speaker has horizontally mounted drive-units, the off-axis behavior can be critical. Shown in fig.7, therefore, are the differences in response to be expected as the listener moves to the side of the BBSM-6's tweeter/midrange line. The response actually changes only a little up to 15° off-axis, suggesting that the speaker will have quite a wide "sweet spot." Though this finding would contradict the "vertical venetian blind" effect that I noticed during the auditioning, it is possible, of course, that the horizontal lobing is too fine-structured to be picked up by these relatively coarse measurements. At 30° off-axis—the listening position with the speakers firing straight ahead—the response gets lumpy in the lower crossover region and peaky throughout the low treble. This is most probably connected with the fact that the high 6kHz crossover frequency for the midrange/HF transition means that the midrange unit is being used in a range where its size is larger than the wavelength of the emitted sound: the wavelength of a 6kHz tone is around 2"; the unit's cone is 3" in diameter. This will result in severe beaming for the top octave or so of the midrange-unit output, which probably leads to the "vertical venetian blind" effect heard.

Fig.7 Westlake BBSM-6F, lateral response family at 45", normalized to response on tweeter axis, from back to front: differences in response 30–7.5° off axis, reference response, differences in response 7.5–90° off axis.

In the vertical plane, the close drive-unit spacing and the drive-unit phasing chosen by Westlake result in little change in response between the tweeter and midrange-unit axes (fig.8). Sit so that you can see the top of the cabinet, however, and a severe suckout develops in the upper crossover region, leading to my subjective feeling of "hollowness" to the sound. Though I didn't look at the response 15° below the tweeter axis, the symmetrical 24dB/octave nature of the crossover should give a similar suckout on that axis. If you can't find stands tall enough to enable you to sit somewhere between the tweeter and midrange axes, you should experiment with tilting the speakers upward so that the speaker's main response lobe will be aimed up at your listening seat. (Why would Westlake engineer the speaker's crossover to give such a small vertical window in the first place? Remember the BBSM-6's professional lineage: the usual place to put a pair of nearfield monitors is just above the mixing console's meter bridge, so that engineer's ears will be on or just below the midrange axis.)

Fig.8 Westlake BBSM-6F, vertical response family at 45", normalized to response on tweeter axis, from back to front: differences in response 15–7.5° above axis, reference response, differences in response 7.5° below axis.

Carrying out a spatially averaged 1/3-octave spectrum analysis in-room gives the curve shown in fig.9, with a broad peak centered on 500Hz. The rolloff in the treble is smooth, that below 500Hz is broken up by residual room effects and extends to 40Hz or so. That the BBSM-6 sounded more lightweight than this curve would indicate is explained by the ear taking the exaggerated midrange level as its reference.

Fig.9 Westlake BBSM-6F, spatially averaged, 1/3-octave response in JA's Santa Fe listening room.

Finally, looking at how the Westlake's response changes as the sound decays gives the cumulative spectral-decay or "waterfall" plot (fig.10). There is some inconsequential hash from the soft-dome tweeter between 7 and 12kHz, which would add a very slight degree of "thuffiness" to the sound of treble transients; otherwise, the decay in the mid- and high-treble is clean. (The black streak just below 16kHz is the computer monitor's scanning frequency and should be ignored.) Note, however, the ridge parallel to the time axis at the cursor position, 2575Hz. This represents a major resonance, perhaps due to a breakup mode of the midrange cone, which would be expected both to lend the sound some hardness in the low treble and pop some piano notes forward in the soundstage, as noted during the auditioning. There is also some complicated behavior noticeable below 1200Hz or so, which might correspond to reflections of the impulse from the baffle edges.—John Atkinson

Fig.10 Westlake BBSM-6F, cumulative spectral-decay plot on tweeter axis at 45" (0.15ms risetime).

---

Footnote 1: The B&K microphone's own deviation from a flat response was subtracted from all the published curves apart from the three-dimensional cumulative spectral-decay plots. Some readers have pointed out that for me to imply that the amplitude response of the microphone can be corrected in this manner is incorrect, the mike's response changing with both angle of incidence and its distance from the sound source. In absolute terms they're correct, of course, but as I always use the same measuring distance with the microphone oriented so that its capsule faces the sound source, applying this correction seems justifiable. However, it certainly would not be appropriate for reverberant-field measurements, such as those made using a spectrum analyzer.