Sidebar 3: Measurements
Because of the YGA Sonja 1.3's bulk and mass, I performed the farfield measurements, using DRA Labs' MLSSA system and a calibrated DPA 4006 microphone, only on the Sonja 1.1 module. For the nearfield and spatially averaged room responses of all three modules, I used an Earthworks QTC-40 microphone. The Sonja's voltage sensitivity is specified as 88dB/2.83V/m. However, my estimate was significantly less than that, at 85dB(B)/2.83V/m. The Sonja 1.3 needs a big amplifier to reach the appropriately high SPLs of which it is capable.
The 1.3's electrical impedance is specified as 4 ohms, with a minimum value of 3 ohms. Fig.1 show the measured impedance. It ranges between 3 and 6 ohms most of the time, with slightly greater values in the bass and in the upper midrange. However, not only does the magnitude remain between 3 and 4 ohms throughout the midrange, with a minimum magnitude of 2.73 ohms at 281Hz, but the electrical phase angle is significantly inductive in the midrange, with a combination of 3.8 ohms and +45° phase angle at 814Hz. This speaker needs to be used with an amplifier capable of delivering both amps and volts in quantity—I suspect that the hardness I noted at sustained SPLs above 100dB was simply due to the amplifiers clipping into the demanding impedance. Fig.2 shows the impedance and electrical phase of the Sonja 1.1 module alone. As the input signal is high-pass filtered with a series array of capacitors, this results in the Sonja 1.1 becoming a highly capacitive load in the upper bass and below. Yes, the speaker's output is rolling off by the time you reach the combination of 4.2 ohms and –60° at 70Hz, but music has a lot of energy in this region. The Sonja 1.1 is thus an even more difficult load than the Sonja 1.3, and needs to be used with an amplifier capable of delivering large currents without flinching.1
Fig.5 shows how the Sonja 1.1's individual responses sum in the farfield, averaged across a 30° horizontal window centered on the tweeter axis (blue trace), to which is added the combined response of the Sonja 1.2 and 1.3 modules (red trace). Despite their different enclosure shapes, the 1.2 and 1.3 behaved identically; the summed nearfield outputs of the 1.2 and 1.3 roll off below 30Hz and above 50Hz. With the spatial averaging evening out the tweeter's behavior in its passband, the Sonja 1.1 offers an astonishingly flat farfield response.
Footnote 1: The 1.1 is also available without the high-pass filter.
Fig.1 YGA Sonja 1.3, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
Fig.2 YGA Sonja 1.1, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
The traces in figs. 1 and 2 are free from the small discontinuities that would indicate the presence of enclosure resonances; investigating the vibrational behavior of the Sonja 1.1's cabinet with a plastic-tape accelerometer, I found nothing untoward, other than a very mild mode at 815Hz (not shown). However, the 1.2 and 1.3 cabinets emitted faint, metallic-sounding bonks when tapped, and with pink noise I could hear faint whistles from the woofer enclosures when I listened with a stethoscope. The accelerometer revealed the 1.2 cabinet to have relatively strong resonances at 580 and 795Hz, the 1.3 cabinet at 560 and 795Hz (fig.3). Although these frequencies are well above the passbands of the 1.2 and 1.3 woofer enclosures—you would not, therefore, expect them to be excited with music playing—I could faintly hear the midrange resonances when I used a stethoscope to listen to the woofer enclosures while the speaker played pink noise. This is presumably due to the very tight mechanical coupling between the three enclosures. YGA's Yoav Geva does say that these panel vibrations above the woofer crossover frequency are due to the speakers being sited on a nonrigid floor, which hinders efficient draining of the vibrations to ground.
Fig.3 YGA Sonja 1.3, cumulative spectral-decay plot calculated from output of accelerometer fastened to center of 1.3 woofer unit side panel (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz).
Turning to the YGA's frequency response, the Sonja 1.1's acoustic crossover in the farfield is shown in fig.4. The two midrange units (red trace) offer a superbly flat output in their passband, handing over to the tweeter (blue trace) just below 2kHz. The tweeter's output is basically flat, but with some small peaks and dips evident. (A response graph supplied me by YGA, taken in the NRC's anechoic chamber in Ottawa, was identical in this respect.) The acoustic filter slopes appear to be close to fourth-order low-pass for the midrange drivers, third-order high-pass for the tweeter. The midrange units roll out below 100Hz, reaching –6dB at 65Hz, as specified.
Fig.4 YGA Sonja 1.1, acoustic crossover on HF axis at 50", corrected for microphone response, with summed nearfield responses of midrange units (red) plotted below 400Hz.
Fig.5 YGA Sonja 1.3, anechoic response on HF axis at 50", averaged across 30° horizontal window and corrected for microphone response (blue), with nearfield responses of Sonja 1.1 module (blue) and the sum of the 1.2 and 1.3 modules (red) plotted below 400 and 200Hz, respectively.
Whether or not a flat on-axis response corresponds to a neutral balance in-room depends on the speaker's radiation pattern. The Sonja 1.1's lateral dispersion, normalized to the tweeter-axis response, is shown in fig.6. Other than in the top octave, the YGA's output generally drops smoothly and evenly to its sides, but there is a slight off-axis flare evident at 5.5kHz. However, this is mainly due to a small suckout in the on-axis response filling in to the sides. In the vertical plane (fig.7, plotted across the full ±45° window rather than the ±15° I sometimes use for large, cumbersome speakers), the vertical dispersion is disturbed by a lack of energy developing more than 10° above and below the tweeter axis. This will be due to destructive interference between the spaced midrange units, and confirms my listening impression that the Sonja's tweeter needs to be aimed at the listening position.
Fig.6 YGA Sonja 1.1, lateral response family at 50", normalized to response on HF axis, from back to front: differences in response 90–5° off axis, reference response, differences in response 5–90° off axis.
Fig.7 YGA Sonja 1.1, vertical response family at 50", normalized to response on HF axis, from back to front: differences in response 45–5° above axis, reference response, differences in response 5–45° below axis.
The red trace in fig.8 shows the Sonja 1.3s' spatially averaged response in my listening room, with the response of the Vandersteen Treos I reviewed in March shown in blue. (Both traces were generated by averaging 20 1/6-octave–smoothed spectra, taken for the left and right speakers individually using SMUGSoftware's FuzzMeasure 3.0 program and a 96kHz sample rate, in a vertical rectangular grid 36" wide by 18" high and centered on the positions of my ears. This eliminates the room acoustic's effects, and integrates the direct sound of the speakers with the in-room energy to give a curve that I have found correlates reasonably well with a speaker's perceived tonal balance.) Both speakers benefit from the 32Hz diagonal mode in my room, but whereas the Treos rolled off below that mode, the Sonja 1.3s extended at full level to well below 20Hz. There is still a lack of energy between 70 and 120Hz, which leaves the YGAs' lows sounding slightly disconnected from their midrange, but the upper-frequency regions are generally even. Compared with the Vandersteens, which are no slouches in the treble, the YGAs have a slight excess of energy in-room between 5 and 15kHz, which may well correlate with the somewhat analytic-sounding balance.
Fig.8 YGA Sonja 1.3, spatially averaged, 1/6-octave response in JA's listening room (red); and of Vandersteen Treo (blue).
Turning to the time domain, the Sonja 1.1's step response on the tweeter axis (fig.9) indicates that the tweeter and midrange units are all connected in positive acoustic polarity. The decay of the tweeter's step smoothly blends with the start of the midrange step, correlating with the excellent frequency-domain integration of their outputs seen in fig.4. The two woofers' step responses (not shown) confirm that these, too, are connected in positive acoustic polarity. The Sonja 1.1's cumulative spectral-decay plot on the tweeter axis (fig.10) is superbly clean.
Fig.9 YGA Sonja 1.1, step response on HF axis at 50" (5ms time window, 30kHz bandwidth).
Fig.10 YGA Sonja 1.1, cumulative spectral-decay plot on MF axis at 50" (0.15ms risetime).
Finally, I very rarely examine a speaker's distortion because of the difficulty of doing so with absolute accuracy outside of an anechoic chamber. (Ambient noise and the room's acoustics are both interfering variables.) Out of curiosity, however, I did look at the Sonja 1.1's linearity. This is a low-distortion speaker. A 500Hz tone at a high continuous SPL of 95dB at 24" generated just 0.1% of third harmonic (fig.11), while the distortion harmonics were even lower than that with a 1kHz tone at the same SPL (fig.12). In this graph, the second and fourth harmonics are the highest in level, at –66dB each (0.05%). With a stimulus higher in frequency, the tweeter produced about 0.2% of second harmonic at this same high SPL.
Fig.11 YGA Sonja 1.1, spectrum of output on HF axis at 24", 500Hz at 95dB SPL (10dB/vertical div., linear frequency scale).
Fig.12 YGA Sonja 1.1, spectrum of output on HF axis at 24", 1kHz at 95dB SPL (10dB/vertical div., linear frequency scale).
The YGA Sonja 1.3 offered the excellent measured performance you should expect at its price.— John Atkinson
Footnote 1: The 1.1 is also available without the high-pass filter.
