Floor Loudspeaker Reviews

Sidebar 3: Measurements

The MartinLogan Montage is of just-above-average voltage sensitivity, as an estimated 88.2dB(B)/2.83V/m, which is below the specified 90dB figure. Its impedance plot (fig.1) also indicates that it approaches a 4 ohm load in the midrange and rarely goes above 8 ohms in the treble. The speaker will work best with amplifiers and receivers rated to drive a 4 ohm load.

Kal's primary criticism of the Montage concerned its lack of bass definition at high volumes, which he associated with the cabinet behavior: "If I turned up the volume . . . the port volume increased a bit but the cabinet vibrations rose in alarming proportion." There are two very slight wrinkles in the impedance traces in the 200–400Hz octave, which might be due to resonant modes in the enclosure walls. I therefore fastened a simple piezoelectric plastic-tape accelerometer to various places on the cabinet while I drove the speaker with high-level MLS noise bandwidth-limited to 2kHz.

I found two major resonances on all the enclosure surfaces in this region, at 277Hz and 360Hz, as well as a lower-level mode at around 500Hz. Fig.2, for example, is a waterfall plot calculated from the accelerometer's output while it was fastened to the sidewall level with the lower of the two woofers. The sound file demonstrates the sound of this plot.

It clearly shows all three modes as ridges of delayed energy, with the mode at 277Hz highest in level. It is possible that this behavior correlates with Kal's dissatisfaction with this aspect of the Montage's sound. The resonant behavior remains masked by the music at low levels, but increasingly intrudes as the playback volume increases.

Turning to the frequency domain, the saddle in the impedance magnitude plot (fig.1) indicates that the rear-facing port is tuned to the quite low frequency of 32Hz, which implies good bass extension. The blue and red traces in fig.3 are the nearfield responses of the lower and upper woofer, respectively; both traces show the expected reflex notch in their outputs at the port tuning frequency. (At this tuning frequency, the back pressure from the port resonance effectively holds the woofer cones motionless.) However, the port's response (green trace, scaled in proportion to its radiating diameter) covers a wider bandpass than usual, not rolling off until above 80Hz. The port's output doesn't show any trace of the cabinet resonances, but both woofers show a slight discontinuity in their responses at 277Hz, the frequency of the more powerful resonance in fig.2.

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

Fig.2 MartinLogan Montage, cumulative spectral-decay plot calculated from the output of an accelerometer fastened to the cabinet's side panel level with the woofer (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz).

The lower woofer appears to roll off quite steeply above 500Hz, though the upper woofer maintains its upper-midrange output up to the crossover point to the tweeter. The black trace below 300Hz in fig.3 is the complex sum of the three nearfield low-frequency outputs, taking into account acoustic phase and distance from the nominal farfield point. The apparent upper-bass boost in its output is due in part to the nearfield measurement technique, which assumes a 2pi acoustic environment. (See fig.6 in Keith Howard's article on measuring low-frequency loudspeaker response, on p.161 of the April issue.) But this trace does indicate a slightly elevated bass response, to give the illusion that the Montage is larger than it actually is—a common design strategy with relatively small speakers.

Fig.3 MartinLogan Montage, anechoic response on tweeter axis at 50", averaged across 30º horizontal window and corrected for microphone response, with the complex sum of the nearfield upper-woofer (red), lower-woofer (blue), and port (green) responses, taking into account acoustic phase and distance from the nominal farfield point, plotted below 300Hz.

Moving higher in frequency, the MartinLogan's overall balance is flat in fig.3, though disturbed by some small peaks and dips. The suckout at 600Hz doesn't occur in the behavior of the woofers when measured nearfield. It could well be the result of some partial cancellation of their outputs at the microphone position due to their vertical spacing resulting in different arrival times, in which case it might not be a significant factor in-room.

I originally assumed that the suckout just below 4kHz in this graph was due to the crossover between the upper woofer and the ribbon tweeter. All else being equal, the contrast between this lack of energy and the slight excess in the region immediately below it might explain Kal's feeling that the speaker was "just a bit more forward" than he was used to. But as the bottom trace in the plot of the Montage's vertical dispersion reveals (fig.4), the actual crossover frequency is slightly lower than both 4kHz and the specified 2.5kHz, at close to 2.2kHz. (This plot also reveals that the lower-frequency suckout worsens above the tweeter axis but fills in below it, suggesting that it is indeed due to interference between the vertically separated woofers.) However, this graph indicates that, in the treble, the Montage is relatively tolerant of listener ear height, as long as the listener sits on or above the 35"-high tweeter axis.

Fig.4 MartinLogan Montage, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 15–5º above axis, reference response, differences in response 5–15º below axis.

In the horizontal plane (fig.5), the Montage offers wide dispersion in the midrange and low treble. There is a lack of off-axis energy in the region of the on-axis treble suckout, however, which will add to the audibility of the energy excess immediately below that region. The tweeter behaves as a dipole above 10kHz, which will make the speaker sound a little mellow in large rooms. Peculiarly, there is a significant off-axis flare between 7kHz and 9kHz, which might also add to the feeling that the speaker has a "forward" balance.

Fig.5 MartinLogan Montage, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90–5º off–axis, reference response, differences in response 5–90º off-axis.

There are no surprises in the time domain, the Montage's step response (fig.6) revealing that all three drive-units are connected in positive acoustic polarity. The tweeter's output can be seen to arrive first at the microphone, followed by that of the two woofers. The different distances of the woofers from the microphone results in the slight undulations seen in the "tail" of their step, which also manifests itself as the ridge of delayed energy at 600Hz in the cumulative spectral-decay plot (fig.7). The treble region in this graph is commendably clean, however. Kal's positive opinion of this tweeter is well-founded.

Fig.6 MartinLogan Montage, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.7 MartinLogan Montage, cumulative spectral-decay plot at 50" (0.15ms risetime).

Taking its affordable price into consideration, the Montage is a well-balanced design, and its appearance is very family-friendly. However, that lively cabinet works against a strong recommendation. I will be writing more on this promising speaker in a "Follow-Up" next month.—John Atkinson