Stand Loudspeaker Reviews

The integration of the bass octaves with the midrange proved to be another strength. You may recall the famous MartinLogan Sequel II upper-bass hole, which squeezed the testosterone from male vocalists. The Aria will do none of that. It does not slight any one particular bass register. The reproduction of male voice, cello, and double bass was tonally quite authentic. The upper-bass/lower-midrange anemia that more often than not afflicts so many minimonitors was blissfully absent.

There wasn't a prodigious quantity of deep bass, which meant that orchestral foundation lacked authority. The Aria could not deliver even a hint of bass lines around 40Hz, which isn't surprising in view of its chosen bass alignment. There's extension to 60Hz, a little below the lowest note of the cello, but that's all you get. The midbass was tight and nicely detailed, but the upper bass was a bit too thickly textured and congested, as if a dab of Heinz ketchup was allowed to permeate the fabric of the music. What this means, in plain English, is that the timbres of Peclard and Pidoux's cellos (Offenbach, Suites pour deux violoncelles, Harmonia Mundi HMC-901043), while quite convincing in terms of pitch definition, lacked sufficient air and body to bloom to their full potential.

As expected of a vented design with a highish box resonance, the Aria had trouble with subsonic energy during LP playback. There was considerable cone pumping with accompanying loss of midrange smoothness. Use of the Pawel Acoustics Subsonic Filter greatly reduced the sense of strain due to distortion products that had crept into the midrange. Clearly, a good subsonic filter would be a wise investment for this loudspeaker—especially if you plan to listen to a lot of vinyl.

When it was not beset with subsonic challenges, the Aria's midrange was consistently smooth and clean without a hint of any response glitches around the crossover frequency. There was plenty of midrange detail. Laudate! (Proprius 7800), one of my favorite choral music selections, features a heavily layered backdrop of male and female voices. The Aria breezed through without any problem in resolving massed voices.

I've deliberately left the bad news for last, principally to underscore the frustrations inherent in this business. The Aria showed such finesse in the areas of soundstaging, lower-octave integration, midrange tonality and detail, that in view of its asking price I was itching to embrace it with wide-open arms. But, alas, there were nasty things going on above 4kHz.

Take a look at fig.3. (This, as well as the other MLSSA measurements, were performed by John Atkinson, who has taken to this system like a fish to water.) What is shown is the impulse response of the Aria measured on the tweeter axis at a 48" distance. (Fig.4 shows the step response, which reveals that all three drive-units are connected in positive acoustic polarity.) As you can see, the Aria rings for over two milliseconds, which usually means trouble in tweeter land.

The treble sizzle of the Aria 5 hits the listener squarely in the face. Sibilants were overemphasized and brass was too edgy in tone. Treble transients were spitty. Nylon strings sounded too hard and steely. Violin overtones were grainy and lacking in sheen. Massed strings managed to sound strident. The upper registers of female voice were somewhat rough during quiet passages, and quite obnoxious when things got loud. Of course, I could find recordings or particular cuts that were deficient in the treble and which were rendered acceptably well through the upper registers. But, on the whole, even with one of the best tube amps money can buy, the treble continued to gnaw at me.

Naturally, I experimented with varying the listening axis in the horizontal plane. That this would help was evident from a set of in-room measurements I made with the Neutrik System 3100. (A 1/3-octave warble tone was used in all of these measurements.) At 4' on-axis with the tweeter (ignore the bass response at this distance), the rising response of the tweeter above 3kHz is quite evident in fig.5. Note that there is an over-response in the 4–8kHz and 15–17kHz regions. There is a useful reduction in treble energy 30° off-axis—especially in the extreme treble—as can be seen in fig.6. The Focal literature shows a well-behaved frequency response, with only small glitches around 7 and 17kHz.

It is not clear, however, just how the published data were derived. On the basis of what had been promised, I had assumed that the tweeter was equalized to remove the on-axis rise above 4kHz. So finding the bright balance of the apparently raw tweeter intact was surprising indeed. Take a look at the righthand side of fig.7. Shown here is the Aria's FFT-derived response on the tweeter axis averaged across a 30° lateral window. This is supposedly the promised land, the sweet spot within which the design should have provided a stable polar response that does not shift with frequency. The average response is, in fact, markedly elevated above 4kHz because the on-axis response is so hot.

Ultimately, the best means of achieving an off-axis listening environment turned out to be extreme toe-in so that the tweeter axes crossed over well in front of the listening seat. With no toe-in, so that the listener was given a more neutral treble balance, the imaging suffered, outlines becoming diffuse, which undermined one of the Arias' strengths.

Let me forestall any suggestions of achieving the same end by toeing-out the speakers. In my opinion this would be a stupid move, resulting only in the creation of excessive lateral reflections by splashing a lot of energy at the side walls. Such an approach would negate the entire design philosophy of the Aria 5. If you're going to do that, then you might as well consider an omnidirectional design.

If the treble problem were completely amenable to solution by the simple expedient of turning away from the hot axis, we would be home free. Unfortunately, the tweeter continued to ring.