Audiostatic ES-100 loudspeaker Measurements
Sidebar 1: Measurements
The Audiostatic's impedance magnitude (fig.1) only drops below 10 ohms in the mid-treble and above, reaching a minimum value of 5.7 ohms at 14kHz. It lies above 20 ohms in the midrange and bassnote the expanded scale on this graphimplying that the speaker is a very easy amplifier load, even considering the high capacitive phase angle below 100Hz. Its sensitivity, however, as DO suspected, is very low. Its B-weighted output on noise lies some 2.5dB below that of the already chronically low LS3/5a, implying a sensitivity of 80dB/W/m (footnote 1). Small amplifiers will be turned away at the door.
Fig.2 shows the ES-100's overall response on an axis midway up the panel (42.5" from the floor), averaged across a 30° horizontal window at a distance of 45". This is spliced to the nearfield panel output below 300Hz. The trend drops throughout the treble, due to the speaker's very limited horizontal dispersion; on-axis, the output is maintained into the high treble. In the bass, the output is dominated by the panel's "drumhead" resonance centered between 40Hz and 50Hz. There's effectively no bass output below 40Hz, while the notches between 50Hz and 60Hz and at 225Hz are perhaps due to some kind of anti-resonant behavior.
Measuring the on-axis response of physically large speakers is far from trivial, due to the proximity effect present when the microphone distance is not significantly larger than the largest dimension of the speaker. This will tend to add a clockwise tilt to the response. To investigate this aspect of the ES-100's performance, I measured the on-axis response at microphone distances of my standard 45" and 90", the farthest I could easily get away from the speaker and still get a reasonable time window before the first reflection of the speaker's sound from a room boundary arrived. Fig.3 shows these responses, plotted from 200Hz to 20kHz. The cancellation suckout at 1700Hz is deeper at the closer distance, but other than that and a few other minor peaks and dips, there's no significant difference in the response shape. (The difference in apparent bass rolloff is not real, and is due to the difference in time windows used to derive the frequency response from the windowed impulse response.)
Note that doubling the measuring distance only results in an average 3dB drop in sound-pressure level, a clear indication that the speaker acts as a line source. (With a point-source speaker, doubling the distance drops the level by 6dB.)
Laterally (fig.4), the speaker's top two treble octaves drop off significantly with each 5° increase in off-axis angle. At extreme angles, both midrange and treble are significantly depressednot surprising, given the ES-100's dipole nature. Note, however, that even to the side of the panel, there's an obvious peak at 1770Hz. Vertically, the Audiostatic's response changes little for listening positions between 30" and 50" from the groundagain, line-source behavior.
In the time domain, the Audiostatic's step response (fig.5) has an excellent, if inverted, shape, though some HF ringing can be seen. The cumulative spectral-decay, or waterfall, plot (fig.6) is very clean for a panel speakerthis presumably associated with the very low diaphragm mass. Though some low-level and probably inconsequential hash can be seen in the very high treble, this clean decay is the reason for DO's positive comments on the speaker's treble. The suckout just below 2kHz, noticeable in figs.2 & 3, which presumably ties in with the off-axis peak at the same frequency, is associated with some delayed energy.John Atkinson
Footnote 1: Despite this low voltage sensitivity, the Audiostatic is actually quite efficient in the midrange and below, due to its very high impedance. Driven with 2.83V, which will result in 1W power dissipation in an 8 ohm resistor, the ES-100 actually draws much less than 1W from the amplifier to achieve its 80dB spl.John Atkinson