Focal-JMlab Nova Utopia Be loudspeaker Measurements
Because of the size and bulk of the Nova Utopia Be—260 lbs apiece—it made both practical and economic sense for me to fly to Minnesota to measure the speakers in Paul Bolin's listening room rather than have them shipped to my Brooklyn lab. The upside of this was that I could spend some time listening to the speakers in Paul's system; the downside was twofold, in that a) I couldn't examine the Nova's dispersion in any direct manner, and b) the measurements were affected by reflections from the inevitable room boundaries. (In Brooklyn I do my acoustic measurements outdoors with the speaker raised high off the ground, which eliminates the effect of reflections—see my article on measuring speakers.) The microphone used was the Joe D'Appolito-designed Mitey Mike II rather than my usual calibrated DPA 4006. However, the MMII is extraordinarily flat.
The big JMlab was very sensitive, at an estimated 90dB(B)/2.83V/m, measured on the tweeter axis. However, despite its having a nominal value of 8 ohms, the Nova's impedance varied considerably, remaining below 6 ohms for much of the midrange and above 9 ohms for much of the treble (fig.1, solid trace). If the Nova Utopia is used with an amplifier having a highish source impedance—a classic tube design, for example—the treble region will shelve up by a couple of dB, changing the perceived balance considerably. However, it is probably not a good idea to use a tube amplifier with this speaker, as it is a demanding load in the bass region. Not only does the impedance drop almost to 3 ohms in the midbass and the middle of the midrange, there is a punishing combination of 4.1 ohms magnitude and -50 degrees electrical phase angle at 62Hz. A good 4-ohm-capable solid-state amplifier will work best with this speaker.
Fig.1 Focal-JMlab Nova Utopia Be, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)
There is a small glitch just below 200Hz in the impedance traces, which would imply the presence of some kind of cabinet resonance at that frequency. I hadn't taken an accelerometer to Minnesota, but I did listen to the half-step-spaced tonebursts on our Editor's Choice CD (Stereophile STPH016-2) while listening to the walls of the Nova Utopia's four individual sub-enclosures with a stethoscope. The tweeter cabinet was absolutely inert, but I could detect some high-Q behavior around the F below Middle C (175Hz) coming from the sides of the two midrange enclosures. This was so low in level, however, and restricted to such small radiating areas, that it should be subjectively benign. Certainly I heard nothing in the hours Paul and I spent listening to music that suggested the presence of cabinet-induced coloration.
The saddle at 31Hz in the impedance magnitude trace suggests that this is the tuning frequency of the rectangular port at the foot of the baffle. This was confirmed by the nearfield measurements of the woofer's response (fig.2, red trace), which has the usual minimum-motion notch at the same frequency, and that of the port (fig.2, green trace), which peaks in this region. Both woofer and port outputs are commendably free from any out-of-band effects. The twin midrange units start to roll off gradually below 400Hz (fig.2, blue trace), but are still putting out significant output at 80Hz.
Fig.2 Focal-JMlab Nova Utopia Be, quasi-anechoic response on tweeter axis at 36", averaged across 30 degrees horizontal window and corrected for microphone response, with the nearfield responses of the midrange unit (blue), woofer (red), and port (green), and their complex sum (black), taking into account acoustic phase and distance from the nominal farfield point, plotted below 300Hz, 450Hz, 650Hz, and 360Hz, respectively.
As a result of this overlap, the JMlab's nominal low-frequency response in the farfield (fig.2, black trace)—calculated by adding the outputs of the individual sources in the ratio of the square roots of their radiating areas and taking acoustic phase into account—is boosted in the upper bass. This is so, even when it is taken into account that the nearfield measurement technique assumes a hemispherical acoustic environment for the drive-units. Certainly, in my own auditioning of the Nova Utopias in PB's room, the bass region sounded a little elevated, though very tight, very extended, and very clean overall.
Higher in frequency, the JMlab's response, averaged across a 30 degrees horizontal window on the tweeter axis, was very flat through the midrange and treble, though with a slight loss of energy in both the presence region and the top octave. This response shape will be due partly to the fact that I measured at a 36" microphone distance rather than the 50" I have standardized for my outdoor measurements, which is still to some extent in the speaker's nearfield. However, it will also be due partly to the fact that the Nova Utopia's top octave does start to roll off more than 10 degrees to the sides, presumably due to the tweeter's dispersion being modified by the wide baffle. (On-axis, the tweeter's output extended to above my measurement limit of 30kHz.) I did feel that the JMlab's balance sounded mellow, even though this was not at the expense of the presentation of high-frequency detail.
The Nova's 1/3-octave-smoothed in-room balance is shown in fig.3. To get this graph, I averaged ten response curves each for left and right speakers individually, taken in a 30" wide by 18" high grid centered on the position of PB's ears when he was sitting in his listening chair. This spatial averaging both minimizes the effects of room modes on the measured response and results in a reasonably good correlation between what is measured and what is heard. The elevated bass region is apparent in this graph, as is the top-octave rolloff, while the mid-treble is very slightly raised in level. Otherwise, the JMlab's in-room balance is superbly flat.
Fig.3 Focal-JMlab Nova Utopia Be, spatially averaged, 1/3-octave-smoothed response in PB's listening room.
In the time domain, the JMlab Nova Utopia's step response (fig.4) indicates that both the tweeter and midrange units are connected with the same acoustic polarity, while the woofer is connected in the opposite polarity. The speaker is not time-aligned, but each drive-unit's step feeds smoothly into that of the next lower in frequency, which inherently suggests excellent frequency-domain integration. Ignore the discontinuities after the 6 millisecond mark in this graph, by the way, which are due to reflections of the speaker's direct sound from the walls and ceiling of PB's room arriving at the microphone.
Fig.4 Focal-JMlab Nova Utopia Be, step response on tweeter axis at 36" (5ms time window, 30kHz bandwidth).
Unfortunately, the presence of these reflections compromises the speaker's cumulative spectral-decay plot unless the impulse response is drastically windowed, which it was to produce fig.5. The windowing results in invalid data in this graph past the 2ms mark, hence the "dotting out" seen in the graph's floor. A couple of ridges of delayed energy can be seen in the high treble; these are probably low enough in level to have no subjective consequences.
Fig.5 Focal-JMlab Nova Utopia Be, cumulative spectral-decay plot at 36" (0.15ms risetime).
Measuring large, bulky speakers can be an exercise in frustration, even under ideal circumstances, and attempting to do the work in a room further darkens the glass. But through that dark glass of measurements it can be seen that the Nova Utopia Be offers some excellent speaker engineering.—John Atkinson