Sonus Faber Cremona loudspeaker Measurements

Sidebar 2: Measurements

The Cremona was significantly more sensitive than average, at an estimated 91dB(B)/2.83V/m. It should therefore work well with moderately powered amplifiers. However, that amplifier must still be capable of delivering a goodly amount of current. Even though the Sonus Faber's impedance magnitude remains reasonably uniform at higher frequencies, it drops to 2.8 ohms through the upper bass, with a cruel combination of 4.5 ohms and a -53 degrees electrical phase angle evident at 73Hz (fig.1).

Fig.1 Sonus Faber Cremona, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

The saddle at 30Hz in the magnitude trace indicates the tuning of the port that reflex-loads the twin woofers. However, both traces in this graph are also perturbed by discontinuities at around 120Hz and 900Hz, which might indicate the presence of problematic resonances. The solidly built cabinet was generally dead to a knuckle-rap test, but did seem a little lively in a few places. Investigating the cabinet's vibrational behavior with a plastic-tape accelerometer, I found modes present at 387Hz, 477Hz, and 539Hz (fig.2). It's possible that the presence of these panel resonances correlated with the trace of coloration I heard with the clarinet. What might be more possible is a strong resonance at 900Hz in the output of the upper port (not shown). This port doesn't appear to do much of anything, but the resonance indicates that something is amiss at this frequency.

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

The colored traces in fig.3 are the responses of the midrange unit (blue), woofers (red), and lower port (green), all measured in the nearfield and plotted with their levels in the ratio of the square root of the radiating areas. The midrange unit takes over from the woofers above 300Hz or so, with what appears to be 6dB/octave filter slopes. The woofers have the expected notch in their output at the reflex port tuning frequency of 30Hz. The port itself has a rather broader peak than the norm, and is significantly down in level. However, this behavior correlates rather nicely with the tight, unboomy character of the Cremona's low frequencies. What is less commendable, however, is the notch in the output of the woofers just above 100Hz, which coincides both with a peak in the port output in the same region and with one of the glitches in the impedance graph. There is obviously some sort of internal acoustic resonance present in this region.

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

Higher in frequency, the presence region is slightly depressed compared with the levels in the upper midrange and the top octave. However, it should be remembered that these measurements were taken on the tweeter axis, which is a highish 38.5" from the floor. With the tiltback provided by the spikes, the actual listening axis will be a little below the tweeter, which has a significant effect on the response. This is shown in fig.4, which shows the changes in response as the microphone moved above and below the tweeter axis. At extreme off-axis angles, a severe suckout develops just below 3kHz, which appears to be the crossover frequency between the tweeter and midrange unit. But there is more presence-region energy 5 degrees below the tweeter axis. The intended tiltback of the speaker does more to even up the perceived treble balance than it does to confer true time alignment.

Fig.4 Sonus Faber Cremona, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 15 degrees-5 degrees above axis, reference response, differences in response 5 degrees-15 degrees below axis.

The lateral dispersion (fig.5) is wide and uniform through the mid-treble, correlating with the stable, well-defined stereo imaging I noted in my auditioning. However, the speaker's output drops rapidly to the sides above 8kHz, which will work against the slight on-axis peak in this same region. In a large room, the Cremona's sound will lack top-octave "air," while in a small, lively room, the flare in the radiation pattern in the bottom octave of the tweeter's passband might make the speaker sound a little bright. In my own room, with the Sonus Fabers in the positions arrived at by Sumiko's John Hunter, the frequency balance throughout the treble is impressively even (fig.6). The lower-midrange and midbass regions are a little suppressed, but this was the tradeoff Hunter had arrived at in getting the best balance between upper-bass bloom and low-bass extension. And in-room, with the expected low-frequency boundary reinforcement, the Cremonas' bass extends down to the 25Hz 1/3-octave band.

Fig.5 Sonus Faber Cremona, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90 degrees-5 degrees off-axis, reference response, differences in response 5 degrees-90 degrees off-axis.

Fig.6 Sonus Faber Cremona, spatially averaged, 1/3-octave response in JA's listening room.

Fig.7 shows the effect of the "vertical string" grille on the Cremona's quasi-anechoic response. A regular series of peaks and dips can be seen, which will have a small effect on the speaker's balance, but these are surprisingly small in amplitude compared with the gross effect introduced by the similar-looking grille of the Krell LAT-1 (see fig.4 in my measurements accompanying Larry Greenhill's review in August 2001). I did most of my serious auditioning with the grilles in place.

Fig.7 Sonus Faber Cremona, effect of the grille on the anechoic response on the tweeter axis (1dB/vertical div.).

In the time domain, the Cremona's step response (fig.8) indicates that, even with its first-order crossover, the speaker is not time-coincident. The tweeter's sound arrives first at the ear, and is connected in inverted acoustic polarity, followed then by the correct-polarity midrange unit, then the inverted woofers. However, the step response of each drive-unit smoothly hands over to the next lower in frequency. This indicates good frequency-domain integration, suggesting that this is the meaning of Sonus Faber's statement that the crossover is "optimized for acoustic phase response." Note that on the tweeter axis, which is where this step response was taken, the tweeter's output is very slightly ahead of where it needs to be for the best integration of its step with that of the midrange unit. Tilting the speaker back slightly, as recommended by Sonus Faber, will bring the tweeter's output into the correct alignment.

Fig.8 Sonus Faber Cremona, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Finally, the Cremona's farfield waterfall plot on the tweeter axis (fig.9) is clean throughout the treble. However, there is some delayed energy in the 1kHz region, which I imagine correlates with the internal acoustic mode mentioned above. All things considered, this is excellent measured performance that suggests a careful balancing of factors on the part of the Cremona's designer.—John Atkinson

Fig.9 Sonus Faber Cremona, cumulative spectral-decay plot at 50" (0.15ms risetime).