Avalon Eclipse loudspeaker Page 3

It was in the presentation of the recorded soundstage that the Eclipse excelled. My usual test for image specificity is the imaging test tracks on the Chesky Test CD (JD37). The Eclipses performed better than any speaker I've heard on this test apart from the big Thiels. Lateral images were tightly defined in space from left of left to right of right. (It was interesting that the reverberation excited by Bob Enders's voice remained pooled in the center-rear of the soundstage no matter where the direct sound came from.) On the LEDR tests, the signals could be heard quite unambiguously to rise in the air above the loudspeaker, almost reaching ceiling height. This, again, is better than any speaker I've tried these tests with, apart from the CS5s.

Soundstage depth, too, was astonishingly deep. In fact, one friend who visited, impressed by the width and depth of the soundstage thrown by a pair of Eclipses, said that he felt the speakers exaggerated the sense of space, almost as if one of the tweeters had been wired incorrectly. As measurement was later to show, this was not the case; the space reproduced by the Eclipse is genyooine! These Avalons join Thiel's CS5 in doing the "disappearing-speaker" thing better than any other speaker I have heard. I've always thought the time-coherency of a speaker's performance an important factor behind this ability. Yet though the CS5 is time-coherent and the Eclipse not, due to its high-order crossover, both are equally excellent.

The imaging excellence offered by the Eclipses seemed to be a function both of their superb image specificity in every plane and also of an astonishing lack of midrange and treble grundge that allows every reverberant morsel of sound to be savored. The title track on Michael Hedges's Aerial Boundaries album (Windham Hill WD-1032) has an upfront balance, but via the Eclipses the guitarist could be heard illuminating a huge dome of ambience with his hammered bass notes, with the occasionally excited flutter echoes presented well to the outside edges of the loudspeakers.

Well-recorded classical music benefited in spades from this ability of the Eclipses to retrieve even the finest amount of ambience from recordings. My listening notes repeatedly featured the words "delicate" or "exquisite" in this regard. Elsewhere in this issue, Gordon Holt recommends the Bruno Walter performance of Beethoven's "Pastoral" Symphony. Surely he jests, thought I, recommending this 30-year-old John McClure-produced CBS recording. After all, wasn't it Mr. McClure who was responsible for those bright, thin Columbia classical recordings of Stravinsky and Bernstein? So I bought the CD release (MYK 36720), and do you know, Gordon was right. Tonal problems apart, the sense of space on this disc as reproduced by the Cardas-linked, tube-driven Eclipses sent a shiver down my spine.

This was as true for mono recordings. A much-played set of records in the Atkinson household is the Smithsonian Virtuoso collection of 78 and early tape transcriptions. On Artur Rubinstein's 1938 performances of piano works by Fauré, Poulenc, de Falla, and Chopin, the Eclipses threw a superb sense of image depth, the piano being distinctly set behind the loudspeakers, with a natural tonality. Even the live Glasnost/Mondial CD, which even the most enthusiastic member of the party-goer audience would not say has been recorded with audiophile sound quality (footnote 3), reproduced via the Eclipses with a terrific sense of "there" there, sonic warts'n'all.

So given that its intrinsic balance in-room will be on the bright side, making it unkind to electronics other than of the thermionic kind, its bass is reticent though clean, and that it throws a superbly defined, wide, deep, deep soundstage, what else is there to say about the Eclipse? Only that its dynamics were limited in a way that I found surprising, given the speaker's superbly clean presentation.

My positive comments were all noted when the playback level was held within strict bounds, in which case the music's ebb and flow were superbly presented. However, although massed soprano voices, such as those on the Telarc Duruflé disc above, had a very true tonality at moderate playback levels, they took on a distinctly strident rattle when the going got loud, causing me to leap for the volume control. Couple that quality with the speaker's brightish balance with solid-state amplifiers, and I eventually just wanted to turn the sound off with less than audiophile-quality recordings.

Take the "My Man's Gone Now" track from the aforementioned Miles Davis album: if I set the playback level so that the opening bass and drum sounded true, allowing me to be enveloped in the sense of space and sonic communication, when Miles's trumpet (which admittedly has been recorded on the bright side) entered, it tended to rip my head open every time he ventured to the top of the treble staff and above unless I turned down the volume to the point where serious listening was not really possible.

Sonic heaven with the right ancillaries and with spls ranging between 75 and 95dB, with a feeling of effortlessness to its sound, the Eclipse acquires an unmusically hard quality to its sound at 96dB or above. This is all the more disturbing in that, due to the speaker's exceptional transparency, the listener is given no forewarning sense of increasing strain.

Avalon pointed out that this threshold could possibly be due to the onset of amplifier clipping. Although the average level of the drive signal to the Eclipses at this level was around 11.75V RMS, supposedly well within the capability of either the Audio Research or VTL amplifiers, capturing the loudest waveforms with a storage 'scope did reveal the peak voltages at these spls to reach ±33V (Classic 60) and ±37V (VTL) (implying that the choral signal had a peak/mean ratio of 10dB or so). These voltages are uncomfortably close to the amplifiers' measured clipping voltages at 1kHz into the Eclipse of ±34.8V and ±47V, respectively. Substituting a 500W VTL amplifier gave a 4dB or so increase in acceptable level from the pair of Eclipses before the onset of hardness on the Duruflé Requiem, suggesting that Avalon did have a point. However, as the more powerful amplifier was nowhere near clipping at 100dB levels in-room, the threshold of audible hardness would now be due to the speakers running out of dynamic range.

Footnote 3: Featuring a high-end band of 30-something magazine writers and manufacturers (including yours truly on Fender bass), this recording was made with a single-point microphone and the Colossus digital recording system. It's yours for a $15 donation to MADD-LA. Write to Bainbridge Records at P.O. Box 8248, Van Nuys, CA 91409-8248.—John Atkinson