Epos Elan 10 loudspeaker Measurements
I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the Epos Elan 10's frequency response in the farfield, and an Earthworks QTC-40 for the nearfield responses. All the measurements were performed with the clothless "audiophile" grille fitted to the front baffle, which gives a smoothly curved contour in the vicinity of the drivers. The Elan 10's voltage sensitivity is specified as 89dB/2.83V/m; my estimate was very close, at 88.6dB(B)/2.83V/m, which is within the margin of experimental error. The impedance is specified as 4 ohms; my measurement (fig.1) indicates that the Elan 10 drops to that value only in the lower midrange, and that its impedance remains above 6 ohms for most of the audioband. This speaker will be an easy load for the partnering amplifier.
There is a small discontinuity in the impedance traces between 400 and 500Hz, and a smaller one just above 1kHz. Looking at the panels' vibrational behavior with a plastic-tape accelerometer (similar to a piezoelectric guitar pickup) revealed resonant modes at 465 and 1110Hz on the sidewalls and top panel (fig.2). However, these modes are low in level and high in Q and frequency, all factors that will work against their having any audible effects.
Fig.3 indicates that the crossover between the tweeter (green trace) and woofer (blue) occurs slightly lower in frequency than the specified 3.2kHz, with what looks like an 18dB/octave acoustic slope for the tweeter feed. Both units are relatively flat within their respective passbands, though the woofer has a slightly elevated region between 600Hz and 2kHz. The minimum-motion notch in the woofer's nearfield output, which is where the cone is held stationary by the back pressure from the reflex-port resonance, occurs relatively high in frequency, at 57Hz. The output of the port (red trace) peaks in textbook fashion at this frequency, and though its otherwise smooth upper-frequency rolloff is broken by some resonant modes in the upper midrange, these are well down in level.
The Elan 10's response, averaged across a 30° horizontal window centered on the tweeter axis (fig.4), is basically flat, though with a slight lack of midrange energy. There is very little sign of the usual upper-bass boost that results from the nearfield measurement protocol, which suggest that the low-frequency alignment is somewhat overdamped; ie, tuned for clarity rather than ultimate weight. As BJR noted, "the midbass through the upper bass was completely clean, crisp, and natural, with no trace of overhang." The low frequencies are down by 6dB at 50Hz, which suggests only modest bass extension.
The slight excess of top-octave energy apparent in fig.4 is compensated for by the tweeter's increasing directionality in this region, as revealed by the plot of the Epos speaker's lateral dispersion (fig.5). The Elan 10's output above 6kHz falls off rapidly more than 30° to the speaker's sides. In the vertical plane (fig.6), a deep suckout at the crossover frequency appears more than 5° above and 15° below the tweeter axis. The stands used with the Elan 10s should be tall enough that a seated listener's ears are on or just below the tweeter axis.
The Elan 10 is specified as using a crossover with second-order electrical slopes for both the high- and low-pass filters. Usually, to compensate for the phase shift in the crossover region that you get with symmetrical second-order filters, the woofer is connected in inverse polarity, which gives a flat response in the crossover region. However, the Elan 10's step response on the tweeter axis (fig.7) reveals that both drivers are connected with positive acoustic polarity. The decay of the tweeter's narrow step doesn't quite blend smoothly with the rise of the woofer's step, which implies that the optimal listening axis will be just below the tweeter. The speaker's cumulative spectral-decay plot (fig.8) is clean overall, but with a very slight amount of delayed energy evident just above 1.2kHz.
Overall, this is impressive measured performance for a speaker that costs $1000/pair.John Atkinson