Sonus Faber Guarneri Homage loudspeaker Measurements
The Guarneri's reflex-port tuning was set at 52Hz—a region where the power handling was enhanced for heavy midbass inputs on typical program material. Deep bass was beyond the compass of this system, and the -6dB bass rolloff point was at 49Hz, giving the potential for satisfactory in-room bass reproduction down to 42Hz—the lowest note of both the orchestral double-bass and 4-string Fender bass.
Fig.1 shows the overall low-frequency output in the nearfield (upper trace); to achieve the moderate bass extension, there is no need for the low-frequency response to be tapered or overdamped, and the output is maximally flat down to 54Hz. Fig.1 also shows the nearfield response for the port (lower trace). This showed a couple of minor duct resonances, at 680Hz and 1.3kHz, but at 15dB below the primary port output and rear-directed, these are not considered significant.
Fig.1 Sonus Faber Guarneri, overall nearfield response (top) and nearfield response of port (bottom) (5dB/vertical div.).
On-axis at Stereophile's customary 45" microphone distance (fig.2), the Guarneri gave an impressively uniform output of ±3dB from 55Hz-20kHz, either in the 1/3-octave presentation shown, or in normal high-resolution analysis mode. There were mild variations, but the golden rule, which asks for the "ups" to balance the "downs," is followed here. There were none of the usual tonal-balance aberrations, with the midrange correctly set against the bass and, likewise, the treble correctly leveled against the mid. A particularly flat response was seen in the midrange—a remarkable ±1dB from 200Hz-1kHz.
Fig.2 Sonus Faber Guarneri, quasi-anechoic response on listening axis at 45" (top, 5dB/vertical div.), with difference between the two samples (bottom, 1dB/vertical div.).
The lower, dotted trace in fig.2 shows the pair matching—the difference between the two samples—with an expanded 1dB/division scaling. A minor irregularity can be seen near the crossover point, but the overall result was within ±0.3dB limits—an excellent result.
While the unusual grille design passed muster in the listening tests, and the reference response in fig.2 was taken with the grille in place, I couldn't resist finding out the exact contribution it made. I don't exactly hate grilles, but in most cases I would have preferred a speaker's designer to have omitted them altogether. The upper trace in fig.3 indicates the change in response that occurs when the grille is fitted. The lower trace shows the effect of just the vertical threads. Remembering that the acoustic half-wavelength at 10kHz is little more than ?" it was extraordinary to find that the array of grille-threads was, in fact, a relatively powerful acoustic device that dominated the result. Those regular ripples between 5kHz and 10kHz are an interference or diffraction-grating effect. Careful comparison with the reference axial response suggests that most of the response ripple was grille-induced. In one sense, this is encouraging, since the fundamental response was so good and the mild grille variations would, in practice, average out over a range of listening axes.
Fig.3 Sonus Faber Guarneri, differences in on-axis response made by complete grille (top) and by grille threads alone (bottom) (1dB/vertical div.).
Fig.4 shows the response on an expanded vertical scale, showing the individual driver responses (ignore the premature rolloff below 250Hz, due to FFT windowing error). The dotted response shows the acoustic crossovers. This nominally blends at 2.5kHz, with initial 6dB/octave (20dB/decade) slopes. The midrange was the most successful at this, holding a gentle slope from 1-6kHz. A 2-6kHz first-order range is achieved for the tweeter, though both drive-units fall off more rapidly than the specified 6dB/octave slopes (dashed lines) more than one or two octaves away from crossover.
Fig.4 Sonus Faber Guarneri, quasi-anechoic response at 0.75m on listening axis, with individual responses of tweeter and woofer (2dB/vertical div.).
If the coloration is low, a monitor-like truthfulness will result—this trend was indeed maintained over a range of forward axes. Looking at the variations in the vertical plane (fig.5, footnote 1), the output was skewed in favor of the preferred below-axis response, which was quite uniform. Above the "reference" axis (actually near the tweeter axis), the output dipped in the crossover region by 17dB, while the treble output was lifted by 1.5dB, unbalancing the overall result. Graphed for 30 degrees, 45 degrees, 60 degrees, and 90 degrees off-axis angles in the lateral plane (fig.6), the Guarneri's output was quite uniform, showing very good lateral directivity and low diffraction—this undoubtedly contributing to the sharp stereo focus heard in the auditioning.
Fig.5 Sonus Faber Guarneri, vertical-response family at 45", normalized to response on listening axis (solid line); difference in response 15 degrees above listening axis (dotted); difference 15 degrees below listening axis (dashed) (5dB/vertical div.).
Fig.6 Sonus Faber Guarneri, horizontal-response family at 45", normalized to response on listening axis, from top to bottom: reference response (solid); differences in response (dotted) 15 degrees, 30 degrees, 45 degrees, 60 degrees, and 90 degrees off-axis (5dB/vertical div.).
Footnote 1: For clarity, the axial outputs of the vertical and lateral responses have been referenced or normalized to the on-axis response. The latter therefore appears as a straight line.