Floor Loudspeaker Reviews

Appropriately for an inexpensive loudspeaker, the Vandersteen 1B is an easy load for an amplifier to drive, as can be seen from fig.1 (its plot of impedance amplitude and phase against frequency). The former stays above 6 ohms across the band. The trace dropping to 6.15 ohms at 10kHz (the cursor position) is with the HF control set to its maximum "+1dB" position. With it set to "–3dB," the impedance stays above 8 ohms throughout the treble. The specified sensitivity is a high 90dB/W/m, again implying a speaker that is easy to drive. From my measurements, I estimate the B-weighted figure to be somewhat lower, at 87dB/W/m, which is average.

Fig.1 Vandersteen 1B, electrical impedance (solid) and phase (dashed) with HF control set to maximum and minimum positions. (2 ohms/vertical div.)

To the left of the rather complex-looking fig.2 are shown the nearfield responses of the rear-facing rectangular port—the bandpass response peaking between 30 and 50Hz—and the woofer (the trace rolling out below 100Hz to reach a minimum excursion point at 32Hz). Also shown is the complex sum of the two responses, indicating the anechoic, half-space low-frequency response. This reaches its –6dB point at 38Hz, implying that the 1B will basically reproduce all but the very lowest notes of the double bass and four-string bass guitar in full measure. Note that the combined response of the woofer and port is less than that of the port alone below 45Hz. This is because below the woofer's minimum-excursion frequency, its output is out-of-phase with that of the port, canceling rather than adding to it.

Fig.2 Vandersteen 1B, HF control set to Flat, anechoic response on tweeter axis at 45", averaged across 30° horizontal window and corrected for microphone response, with the nearfield responses of the woofer (red) and port (blue), along with their complex sum, plotted below 300Hz.

In a conventional reflex-loaded design, this results in an overall 24dB/octave LF rollout, but in the Vandersteen 1B, the slope is more gentle due to the misalignment of what would be a normal reflex tuning to some sort of hybrid. However, to state that the 1B is a transmission line is not strictly true, both the impedance plot (fig.1) and fig.2 indicating that the port does play a significant role in extending the speaker's response downward in frequency.

To the right of fig.2 is the 1B's quasi-anechoic response measured close to the tweeter axis at a distance of 48" and averaged across a 30° horizontal window. (The HF control was set to "Flat" for this measurement.) The tweeter's inaudible, ultrasonic resonance can be seen at 25.4kHz. Some minor bumps and dips are noticeable throughout the midrange and treble, superimposed on a response that gently falls across the audio band, the level at 20kHz being 6dB lower than that at 1kHz. This in itself will lend the 1B a dull tonal quality, though the peakiness in the upper midrange might add some brightness. Nevertheless, Corey remarked on the 1B treble's sounding "shut-in and airless." This was even with the HF control cranked to its max.

Fig.3, which shows just the differences made to the tweeter-axis response by this control, indicates that at its maximum, it boosts the mid-treble by 2.9dB (compared with a maximum reduction in the same range of 4dB). This boost might be thought to significantly alleviate the 1B's dull balance, but as CG found that it didn't, some other factor must also be at work here.

Fig.3 Vandersteen 1B, differences in tweeter-axis response made by setitng HF control to maximum and monimum positions.

Fig.4 shows the manner in which the 1B's response changes as the listener moves to its side. (Just the differences are shown in this graph, meaning that the on-axis response appears as a straight line.) Though the speaker has wide dispersion in the mid-treble, much of its top-octave region does quickly become depressed off-axis. In all but very lively rooms, this will also tend to make the speaker sound rather shut-in. Note that a severe suckout in the low-treble/upper midrange develops at extreme off-axis angles, which might contribute to a rather laid-back, possibly uninvolving in-room balance. The general unevenness of the off-axis changes shown in fig.4 correlates with the "vertical-venetian-blind" effect CG noticed in his auditioning.

Fig.4 Vandersteen 1B, lateral response family at 45", normalized to response on tweeter axis, from back to front: differences in response 90–5° off axis, reference response, differences in response 5–90° off axis.

Vandersteen's excellent handbook for the 1B goes into considerable detail on how the listener can arrange the speaker's stands and tilt-back to ensure that he or she is on the optimal axis. The speaker's behavior in the vertical plane is shown in fig.5. (Again, just the differences in measured response are shown.) To get the most even balance from the 1B, the listener must sit on or just below the tweeter axis, which, without tilt-back, is low—30", even with the speaker sitting on its stand. Even 4" above the tweeter axis, the mid-treble drops significantly, with an even more alarming lack of treble energy apparent on an axis level with the top panel (the one with the appealing tummy hole cut in it) supported by the four dowels. Again, this might make the balance rather uninvolving, even though the mid-treble comes back up again in level for standing listeners, as shown by the rear two traces in fig.5.

Fig.5 Vandersteen 1B, vertical response family at 45", normalized to response on tweeter axis, from back to front: differences in response 20–5° above axis, reference response, differences in response 5–15° below axis.

It is in the time domain, however, where this inexpensive speaker is an overachiever, due to its use of a crossover with first-order electrical slopes and a baffle geometry that really does time-align the drive-units on the listening axis. (To say that because a speaker has a sloped-back baffle necessarily means that its driver outputs are aligned in time is one of the big audio lies.) The 1B's impulse response on its tweeter axis (fig.6) reveals that the beginning of the woofer's output (the gentle rise at 3.6ms) precedes that of the tweeter. The step response calculated from this impulse (fig.7) reveals that this is what is required on this axis to produce a generally excellent right-triangle wave shape. The only other speakers I have encountered that produce this good a step response are models from Thiel and Spica (see this issue's "Follow-Up" on the Spica Angelus, for example). The decay of the 1B's step response back to the time axis, however, is rather uneven, this partly due to what in fig.6 appears to be a number of early reflections of the impulse, presumably from the dowel speaker-frame supports and the top plate. Such early reflections, in my experience, do tend to smear imaging precision and reduce soundstage depth.

Fig.6 Vandersteen 1B, impulse response on tweeter axis at 45" (5ms time window, 30kHz bandwidth).

Fig.7 Vandersteen 1B, step response on tweeter axis at 45" (5ms time window, 30kHz bandwidth).

The waterfall plot calculated from the impulse response (fig.8) also reveals that part of the step response decay's irregularity is due to a resonance mode centered on the 1kHz peak noticed in fig.2. This will tend to throw off the midrange balance, adding a forward character if not an outright coloration. Fig.8 also reveals some minor resonant modes in the low- and mid-treble, probably due to woofer-cone misbehavior above its passband. This will probably be innocuous at low levels, but my experience has been that such treble anomalies rapidly add hardness or strain to a speaker's sound at high playback levels.

Fig.8 Vandersteen 1B, cumulative spectral-decay plot at 45" (0.15ms risetime).

The port trace in fig.2 shows some anomalous behavior in the 250–350Hz and 1kHz regions, the former coinciding both with slight peaks in the overall response and with some impedance irregularities. This pattern is usually a sign that the speaker's cabinet has some sort of resonant problem at these frequencies. Remember that CG was bothered by persistent congestion in the lower-midrange. I couldn't get reliable vibration measurements with an accelerometer, due to the cloth "sock" covering the speaker. However, listening to the speaker's enclosure walls with a stethoscope while I swept a signal generator up and and down in frequency did reveal some serious panel-vibration problems in the entire 110–450Hz region. While the sidewalls were relatively well-behaved, the front and rear walls had significant resonant modes at 115Hz, 140Hz, 175Hz, 300Hz, 315Hz, and 415Hz. Even the corner dowels and top panel vibrated strongly at several frequencies between 275Hz and 360Hz.

Admittedly, the 1B is an inexpensive speaker. But I suspect that if bracing had been used to concentrate this resonant behavior into fewer frequency regions (as with the Linn Keildih reviewed elsewhere in this issue), these cabinet problems would have had far less of an effect on the 1B's character. However, the fact that the front panel of the enclosure both has a reasonably large surface area and faces the listener must make its broad-band vibrational misbehavior a contributing factor to the 1B's lack of clarity in the lower-midrange.

Overall, the fact that the 1B has an excellent time-domain performance, very extended bass (especially considering its affordable price), and a smooth if down-tilted balance, must be put against its closed-in top octave and rather colored, congested midrange. That can only be an individual decision.—John Atkinson