Floor Loudspeaker Reviews

A highlight of the two audio shows I attended this past spring, SSI in Montreal and Axpona in Atlanta, was listening to the Voxativ by Schimmel Pianos Ampeggio loudspeaker. At first sight it looked like yet another rear-loaded horn using a Lowther driver, albeit with a superb level of finish—but on first listen it was obviously not a Lowther-based design, having none of the Lowther "quack." So when Art Dudley told me that he had a pair of Ampeggios in residence to review for this issue's "Listening" column, I resolved to drive upstate, give a listen in his room, and take some measurements.

Because my DRA Labs MLSSA system is installed in an antique Pentium desktop computer (running MS-DOS!) that is not very portable, I measured the Voxativs using Fuzzmeasure 2.0 running on my MacBook laptop in conjunction with an Emu 0404 USB audio interface and a calibrated Earthworks QTC-40 microphone. I brought along one of my 1978-vintage Rogers LS3/5a speakers for reference, and, for consistency, drove the Voxativs with the Musical Fidelity MA-250 solid-state monoblock amplifier that I use for all my speaker measurements, rather than AD's tubed Shindo amplifiers.

Depending on frequency, the Ampeggio was between 12 and 20dB more sensitive on axis at 1m than the LS3/5a, which would confirm its 98dB/W/m specification. However, the impedance is significantly higher than the usual 8 ohms, with minimum magnitudes of 11 ohms in the midrange, and 12 ohms and higher throughout the treble (fig.1; because this graph was taken with Fuzzmeasure, I haven't plotted the electrical phase angle, and the scaling is different from our usual Audio Precision plot). I confess to not knowing as much about horn-speaker theory as I should, so the presence of multiple impedance peaks in the bass in this graph must remain unexplained—though I suspect they relate to the length of the horn and the wavelength of the sound. (Comments are welcome at Stletters@sorc.com.)

Art helped me lift one of the Ampeggios onto a box so that its AC-3X drive-unit was midway between the floor and ceiling of his listening room. Even so, I measured the quasi-anechoic response at 1m rather than my usual 50", in order to maximize the time difference between the arrival of the direct sound and the reflected sound at the microphone. The response at 1m, averaged across a 30° horizontal window, is plotted as the blue trace above 400Hz in fig.2. The treble is a little elevated, with various peaks and dips evident, and drops off above 9kHz or so, though not as rapidly as I was expecting, given the absence of a tweeter. The midrange is commendably flat, and the drive-unit's nearfield response, plotted below 400Hz, rolls off below 200Hz, though with peaks evident at the same frequencies as the low-frequency impedance peaks. Looking at the responses at 5° intervals (not shown), the treble peakiness is at its most extreme on axis; the Ampeggio's high-frequency output is very much smoother as you move off axis. Listening 10° off axis—ie, not toeing the speakers all the way in to the listening seat—gives the most neutral tonal balance in the top audio octaves.

The red trace in fig.2 shows the output at the horn mouth, measured in the nearfield and scaled in the ratio of the square root of the radiating areas of the woofer and horn mouth. It reinforces the output of the AC-3X drive-unit below 200Hz in a somewhat peaky manner, and rolls off below 70Hz, suggesting only modest bass extension. This can be seen in the Ampeggio's spatially averaged response, taken in a grid centered on the listening position in Art's room (fig.3, red trace), and with the speakers positioned where he had auditioned them. However, what is notable about this in-room response is how flat it is. Other than a slight suckout at 3kHz, it falls within an astonishing ±1dB window from 330Hz to 9kHz! For reference, the blue trace in fig.3 shows the spatially averaged response of AD's reference Audio Note speakers placed in his room's corners. While these have very much more bass, their midrange is less flat, and as a result does sound more colored, I feel. The Audio Notes' high frequencies also roll off more rapidly, at least below 13kHz.

In the time domain, the Ampeggio's step response (fig.4) is dominated by a sharp spike correlating with the elevated treble on axis, but multiple arrivals are evident. These give rise to a rather hashy-looking treble in the Ampeggio's cumulative spectral-decay plot (fig.5, plotted with MLSSA using the impulse-response data from Fuzzmeasure). The midrange decay is relatively clean, however.

At first sight, the Voxativ Ampeggio's measured performance doesn't look impressive. But I keep returning to that superbly flat in-room response as being most indicative of what I heard at the audio shows and in Art's room. And that genuine 98dB sensitivity means that the Ampeggio will give satisfyingly high sound-pressure levels even with flea-watt tube amplifiers.—John Atkinson