Epos ES 14 loudspeaker Measurements

Sidebar 3: Measurements

Almost all the measurements accompanying this review were made with the DRA Labs MLSSA system, a full-length card that fits into a PC. This generates a pseudo-random noise signal which is picked up by a B&K 4006 microphone, calibrated to be flat on-axis at the approximate measuring distance I use. The host computer then compares the microphone's output with the drive signal it sent out to the loudspeaker and, by performing a cross-correlation operation, calculates the speaker's impulse response. By windowing out reflections of the speaker's output from room boundaries, a good estimate of its anechoic amplitude in the midrange and above can be made by performing a Fast Fourier Transform operation on that impulse response. To assess low-frequency performance, I repeat the procedure with the microphone almost touching the woofer and port opening; this, too, allows the room to be almost entirely removed from the equation. To look at the speaker's horizontal and vertical dispersion, I take a series of measurements, rotating the speaker a predetermined angle between each reading using the Italian Outline computer-controlled speaker stand/turntable.

The impedance magnitude and phase were measured using the Audio Precision System 1, while for the in-room spectral analyses, I average six measurements at each of 10 separate microphone positions for left and right speakers individually, giving a total of 120 original spectra. These are then averaged to give a curve which in my room has proved to give a good correlation with a loudspeaker's perceived balance. I use an Audio Control Industrial SA-3050A spectrum analyzer with its own microphone, which acts as a check on the MLSSA measurements made with the B&K mike. I also used the Goldline DSP-30 automated spectrum analyzer, which is currently under review.

Finally, I examine the vibrational behavior of a loudspeaker's cabinet by taping a simple PVDF-film accelerometer to the panels and driving the loudspeaker with the MLSSA signal at a high level (around 7.6V RMS). The computer can then calculate the impulse response of the panel from the output of the accelerometer, which is amplified/buffered by a small battery-powered preamplifier.

The ES 14 has a reasonably high sensitivity, 2.83V of B-weighted pink noise raising an estimated 87dB/W/m. Coupled with this highish sensitivity, it is very easy to drive, as shown by the plots of impedance magnitude and phase (figs.1 and 2). It drops below 8 ohms only in the very top octave, and then to just 7.2 ohms. I would have said that this combination of sensitivity and easy impedance characteristic would make the ES 14 a prime candidate for being driven by a classic tube amplifier. Note, however, the large swing in impedance value, particularly in the bass and upper midrange. This will mean that the speaker's sound will become exaggerated in these frequency regions when used with a tube amplifier with a high output impedance. The resultant tonal balance may depart too far from strict neutrality to be acceptable, given that the ES 14 is already a little forward-balanced in the upper midrange.

Fig.1 Epos ES 14, electrical impedance (solid) and phase (dashed) without foam plug in reflex port (2 ohms/vertical div.).

Fig.2 Epos ES 14, electrical impedance (solid) and phase (dashed) with foam plug in reflex port (2 ohms/vertical div.).

Fig.1 shows the impedance measured with the ES 14's reflex port in operation. The saddle at 43Hz reveals the port's tuning. For comparison, the impedance traces in fig.2 were taken with the foam plug in the port. The single peak in the bass indicates that this effectively transforms the ES 14 into a traditional sealed-box design, tuned to 52Hz.

I'm not sure what the wriggles in the trace at 200Hz signify. They don't look like a resonance—the blip at 25kHz is due to the metal-dome tweeter's "oil-can" resonance, for example. Perhaps there's some sort of mechanical equalization present in the woofer's construction. Although you won't be able to see it at the scale these graphs are reproduced in the magazine, there is a very slight wrinkle around 350Hz which does correlate with a cabinet resonance at this frequency (see later).

Fig.3 shows the individual responses of the port, woofer, and tweeter. The port's output is the bandpass centered on 45Hz, this frequency the same as the minimum-motion point of the woofer. Though there is a significant port resonant mode apparent at 870Hz, this is about 15dB from the speaker's nominal output at that frequency. This, in conjunction with the fact that the port is on the speaker's rear panel, facing away from the listener, should minimize this resonant mode's audibility. The woofer's output rises a little toward the top of its passband before rolling out steeply above 3kHz. This is the natural response of the drive-unit, there not being any low-pass filter present; but note the series of cone breakup modes visible between 4kHz and 22kHz. The tweeter comes in relatively slowly, but suffers from a series of peaks and dips before its output zooms up by 20dB at its ultrasonic resonant frequency.

Fig.3 Epos ES 14, acoustic crossover on tweeter axis at 45" corrected for microphone response, with nearfield woofer and port responses (without foam plug) plotted below 300Hz and 1kHz, respectively.

How these responses sum on the tweeter axis can be seen in fig.4. Both sealed-box and reflex low-frequency responses are shown in the bass. The former is over-damped, gently sloping down below 100Hz, while the reflex alignment is pretty much maximally flat down to 50Hz, rolling off steeply below that frequency. The upper midrange is a little forward, while the lower midrange is a little recessed; which is perceived by the listener will depend largely on the program material. Apart from the ultrasonic and therefore inaudible peak, the tweeter has a significant response spike apparent at 9.5kHz. This almost certainly is responsible for the occasionally "sniffy" quality I heard.

Fig.4 Epos ES 14, anechoic response on tweeter axis at 45" averaged across 30 degrees horizontal window and corrected for microphone response, with complex sum of nearfield woofer and port responses below 300Hz (top) and nearfield response of woofer with port blocked by foam plug (bottom).

A peak of energy appears to the speaker's sides (fig.5) just above 5kHz, filling in the on-axis notch in fig.4. The tonal balance doesn't change significantly, however, until the listener is more than 15 degrees off-axis. Vertically (fig.6), the ES 14 needs to be used on tall stands—a significant suckout in the crossover region appears if the listener can see the top of the enclosure. However, considering the use of a first-order crossover with its significant degree of overlap between the drive-units, the vertical dispersion is much better controlled than I would have expected. In my listening room, the spatially averaged response (fig.7) is impressively flat through the high midrange and treble, with useful bass power available down to the 32Hz band, the room reinforcing the speaker's bass response for half an octave below its nominal 50Hz cutoff.

Fig.5 Epos ES 14, horizontal response family at 45", 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 Epos ES 14, vertical response family at 45", normalized to response on tweeter axis, from back to front: differences in response 15 degrees-5 degrees above tweeter axis; reference response; differences in response 5 degrees-15 degrees below tweeter axis.

Fig.7 Epos ES 14, spatially averaged 1/3-octave response in JA's listening room.