Focal Aria 5 loudspeaker kit Page 3

The left-hand side of fig.7 shows the Aria's bass response, measured in the nearfield of both the lower woofer and the port. The port (red trace) can be seen to handle the range below 85Hz or so, its tuning revealed by the minimum in the woofer output to lie at 51Hz. This was confirmed by measuring the impedance of the speaker, the minimum value between the two peaks in the bass also lying at 51Hz (fig.8). Overall, this measurement confirms the Aria's 4 ohm rating, with slight dips just below that figure apparent in the upper bass and the presence region.

Fig.8 Focal Aria 5, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).

The tale of the tweeter
Just exactly where the T90K tweeter rings may be gleaned from the MLSSA "waterfall" plot shown in fig.9. The activity around 15kHz is simply the sound of Kevlar breaking up. Kevlar, while rigid, is not as rigid as a good metal dome, meaning that its first breakup mode lies below 20kHz. There are also minor resonances in the range from 4–8kHz; as you can see, energy continues to flow for nearly 2.5ms. While the intensity of the ringing diminished off-axis, it remained audible—especially in the lower treble. There was no escaping the grainy and raw character in this region, even with extreme toe-in. The situation was not as critical as that presented by the on-axis or direct-fire geometry; the speakers would not drive me away immediately. But after a while the brash personality of the tweeter began to assert itself. Then, when the tweeter got socked with a really heavy load of treble energy, I would start searching for the cotton balls.

Fig.9 Focal Aria 5, cumulative spectral-decay plot on tweeter axis at 50" (0.15ms risetime).

The RC network to the rescue
After Kimon Bellas of Focal America heard that I was unhappy with the treble of the Aria 5, he called to inform me of an option that provides a gentle rolloff in the treble starting around 4.5kHz, and a cut of 3dB at 18kHz. This RC network is available to kit purchasers who feel the Aria 5's treble balance to be a bit on the bright side of reality. The network, consisting of a series half-ohm resistor with then a 3 ohm resistor in series with a 1.5µF capacitor shunting the excess highs to ground, may easily be inserted into the tweeter crossover leg, as shown in fig.10. (The impedance of the apeaker with the revised crossover is shown in fig.11.) I deemed any prospect of improving the treble worth investigating, and agreed to hold off continuing with the review until the parts arrived.

Fig.10 Focal Aria 5, tweeter crossover modification.

Fig.11 Focal Aria 5, original sample with HF shelf filter, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).

Fig.11 shows the frequency response of the Aria at 1m on-axis with the RC network installed. The impact of the network is obvious from a comparison of figs.5 and 12. There is clearly less treble output with the RC network in place. But the basic character of the treble remained unchanged: the impulse response of the modified speaker on the tweeter axis is shown in fig.13, the step response in fig.14, and its waterfall plot in fig.15.

Fig.12 Focal Aria 5, original sample with HF shelf filter, 1/3-octave in-room frequency response on HF axis at 39" (5dB/vertical div.).

Fig.13 Focal Aria 5, original sample with HF shelf filter, impulse response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.14 Focal Aria 5, original sample with HF shelf filter, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.15 Focal Aria 5, original sample with HF shelf filter, cumulative spectral-decay plot on tweeter axis at 50" (0.15ms risetime).

Another listening session followed. I tried a number of listening axes. I switched amps and speaker cable. The treble failed to smooth out enough to make long-term listening enjoyable for me.

As a matter of fact, some performance aspects had deteriorated. The network veiled the soundstage so that the dramatic sense of transparency I experienced before was considerably diminished. Neither were image outlines as crisply defined as before. So, in the end, the network introduced as many problems as it solved.

An interesting finding, uncovered by JA, is documented in fig.16. Shown here is the vertical response family of the Aria 5 at a 48" distance. The middle curve is on-axis with the tweeter, while the curves immediately above and below it are on-axis with the woofers. (The top and bottom curves are the response 15° above the cabinet top and 15° below the cabinet base.) Surprisingly, the flattest axis is that level with the upper woofer; again, something one would not expect from the theory of the D'Appolito configuration. But since these measurements were made on the modified speaker, this anomaly could be due to the presence of the RC network.

Fig.16 Focal Aria 5, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 15–5° above axis, reference response, differences in response 5–15° below axis.

Chapter Two
After perusing the pre-publication copy of this review—manufacturers are sent such a copy of reviews in order that they can respond in "Manufacturers' Comments" in the same issue—Kimon Bellas immediately raised a red flag, strongly suggesting that a faulty sample had been provided for review. Focal's measurements and listening tests, he said, were at odds with Stereophile's findings. It is not that unusual for a manufacturer to disagree with findings that cast a negative light on certain performance aspects of his product. Neither is it unusual for a manufacturer's measurements to deviate from our own. I'm not suggesting that Focal is guilty of this—in hindsight they clearly are not—but acoustic measurements can easily be manipulated by judicious choice of measurement locations to produce what appears to be exemplary performance.

I should point out that it is Stereophile's policy to request a second sample if we suspect that the originals are defective. Other than bad sound in the treble, the Aria did not appear to be malfunctioning. Still, the fact that the first sample had not come directly from Focal persuaded JA to accept a second pair of speakers for evaluation.

Joe D'Appolito's pair of Aria 5s were shipped to us. These are the firstborn pair and were used in generating Focal's own measurements.

There were some clearly visible physical differences between the first and second samples. The cabinet was now solid maple; quite pleasing to the eye, but more significantly from a sonic perspective, the enclosure was much better damped than its MDF counterpart. The knuckle-rap test produced nothing more than a satisfyingly dull thud. A beautiful set of binding posts adorned the back baffle. No plastic here! Another nice touch was that all of the drivers were rebated into the front baffle.

Driven by a bridged pair of Classé Audio DR-8s, the new Arias easily surpassed the sonic performance of the original samples. The general impression was of a stronger affinity to the essence of the music. Specific improvements were apparent in three areas.

First, bass lines became much more distinct. The thick and muddled midbass textures that I complained about earlier evaporated. The bass registers of cello and double bass were clearly detailed and much tighter sounding. It became possible to follow bassline nuances with the greatest of ease. This served to catalyze the bass registers and the lower mids into a smooth, organic whole. The power range of the orchestra greatly benefited from the Aria's robust textural tapestry, now amply sprinkled with detail. Despite its considerable resolving powers, the Aria did not descend to the level of many minimonitors in that it was emphatically other than anemic, threadbare, or analytical sounding. Many of these minimonitors remind me of toy tanks, zipping along the soundstage and pelting the listener with hails of detail. But because all of this detail is no longer camouflaged by an adequately fleshed-out lower range, their sound strikes me as artificial—hi-fi–ish rather than musical.

The Aria's full-bodied tonal balance and considerable dynamic range demonstrate that it is not necessary to sacrifice either of these orchestral elements to achieve imaging excellence. Not only was the music's dynamic ebb and flow on Kol Nidrei (Bruch: Collected Works for Cello and Orchestra, EBS 6060) easily accommodated with tonal authority, but the soundstage was sketched out with convincing spaciousness and transparency. I've earlier described the imaging capabilities of the Aria 5. If anything, these aspects of its performance matured to the point of placing me in more intimate contact with the music. I was finally beginning to understand why Kimon Bellas was so enthusiastic about this speaker. The Aria 5 can definitely sing!

Focal America, Inc.