Mirage M-3si loudspeaker Measurements
The impedance curve (fig.1) shows the double hump at low frequencies typical of a reflex design; the "saddle" between these two humps at about 29Hz (and just over 5 ohms) is the tuned frequency of the M-3si's ports. The overall impedance indicates a fairly easy load to drive, dipping just below 4 ohms only once, at just under 100Hz. The sharp peak just above 20kHz probably indicates the presence of a notch filter to tame a tweeter resonance.
Fig.1 Mirage M-3si, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
The impulse response of the M-3si, taken on the tweeter axis, is shown in fig.2. The ringing typically present with a metal-dome tweeter is minimal. The step response (fig.3) indicates that the drive-units are connected with positive acoustic polarity.
Fig.2 Mirage M-3si, impulse response on tweeter axis at 45" (5ms time window, 30kHz bandwidth).
Fig.3 Mirage M-3si, step response on tweeter axis at 45" (5ms time window, 30kHz bandwidth).
Taking the transform of the impulse response, suitably windowed to eliminate (above about 200Hz) the effects of the room, yields the frequency response, this averaged in fig.4 across a 30° window on the plane of the tweeter. The complex sum of the nearfield low-frequency responses of the woofer and the two ports. Fig.5 indicates the separate responses of the midrange/tweeter and the woofer, indicating an acoustical crossover at about 400Hz. On the left of this graph are the nearfield woofer and port curves.
Fig.4 Mirage M-3si, anechoic response on tweeter axis at 45", averaged across 30° horizontal window and corrected for microphone response, with complex sum of nearfield woofer and port responses plotted below 300Hz.
Fig.5 Mirage M-3si, acoustic crossover on tweeter axis at 45", corrected for microphone response, with nearfield woofer and port responses plotted below 300Hz and 2.5kHz, respectively.
As noted above, the response curves shown in figs.4 and 5 were taken on the tweeter plane. The M-3si's tweeter sits quite high; a seated listener's ears will not be on the tweeter axis. This is probably just as well, as the response at this elevation shows a somewhat recessed midband (with a definite lumpiness at and an octave or so above the woofer-midrange crossover) and a rising top—though the latter is most evident above 15kHz.
Fig.6 shows the response of the M-3si on the woofer axis—the axis which yields the flattest response (this axis is at 31.5" above the ground, somewhat below a normal seating height). Except for some response irregularities above about 8kHz—a very slight brightening just below 10kHz followed by a softening just above that, and a final rise in a range of interest primarily to bats and small children—the response is much smoother on this than on the tweeter axis. Since the tweeter is 48.5" above the base of the loudspeaker, few seated listeners are likely to listen at this height. But others might want to experiment with their seating height or the height of the loudspeakers, the latter altered either by elevating the speakers on short custom bases or, more practically perhaps, tilting them back slightly to "lower" the effective seating height. (If you try the latter, be careful of the speakers' balance and stability.)
Fig.6 Mirage M-3si, anechoic response on woofer axis at 45", corrected for microphone response, with complex sum of nearfield woofer and port responses plotted below 300Hz.
The lateral dispersion of the M-3si is shown in fig.7, with the tweeter-side changes shown to the front. (Note that the tweeter and midrange of the mirror-imaged M-3sis are offset slightly from center.) The loudspeakers are typically set up with the tweeters mounted inboard, toward the center of the soundstage. Note in fig.6 that the top curve is the on-axis curve, which appears as a flat response—there are no changes. The normalized curves below that—actually forward of it, if we view this as a three-dimensional graph—are each an additional 15° off-axis. Note some off-axis peaking at around 800Hz and a considerable amount of irregularity beyond 45° off-axis—presumably due to the bipolar radiation pattern. Keep in mind that our measurements, taken in the front 180° hemisphere of the loudspeaker's circular radiation pattern, cannot account for the significant subjective and objective contributions to the response of the rear-radiating midrange unit and tweeter.
Fig.7 Mirage M-3si, horizontal response family at 45", normalized to response on tweeter axis, from back to front: differences in response 90°–5° off-axis on midrange side of baffle; reference response; differences in response 5°–90° off-axis on tweeter side of baffle.
Fig.8 shows the vertical response family, the normalized on-axis response shown third from the top. Note the rolloff as we move below the tweeter, but also remember that the response on the woofer axis (fig.6) is actually the flattest in absolute terms.
Fig.8 Mirage M-3si, vertical response family at 45", normalized to response on tweeter axis, from back to front: differences in response 10°–5° above tweeter axis; reference response; differences in response 5°–15° below tweeter axis.
Finally, fig.9, the "waterfall" response on the woofer axis—which shows how the speaker's response changes as an impulse response decays, and can therefore illustrate delayed resonances in the response—shows a very clean HF character with just a small amount of hash in the lower treble. Some low-level, not at all unusual, resonances show up around the cursor position at 6kHz. Note that the usual ultrasonic peak typical of a metal-dome tweeter has been suppressed, though a residual ridge remains just below the notch frequency. An interesting phenomenon just below 1kHz is a "valley" in the delayed response visible just behind the "foothills" at about 3 milliseconds. Note that the response decays into this valley, then rises again to form part of the "ridgeline" visible in the lower lefthand corner of the plot. Something is causing the midrange to "take off" again without additional stimulus in this region, though the effect remains quite low in level. The audible consequences of this behavior are likely to be inconsequential.—Thomas J. Norton
Fig.9 Mirage M-3si, cumulative spectral-decay plot at 45" (0.15ms risetime).