Celestion 3000 loudspeaker Page 4

Levels of coloration in the low treble and midrange were low. My piano recording on the Stereophile Test CD reveals midrange problems on conventional two-way moving-coil speakers by some notes jumping out of the soundstage. The Celestion 3000 was remarkably clean in this respect, the Steinway tonality emerging unscathed. The flute and piano tracks on the Stereophile Test CD also reproduced as being extremely true to the sound of the original instruments, a slight accentuation of flute breath noise notwithstanding.

At higher frequencies there was some liveliness apparent in the low treble which, though adding a sense of life to the recorded sound, could easily become fatiguing. This was program-dependent in that it related to how much energy in this region was present in the recording. The sweet-sounding, modern-instruments Four Seasons from the Boston Symphony and Seiji Ozawa (Telarc CD-80070) acquired a little bit of wiriness when violins ventured above the treble staff, but not objectionably so; the original-instruments Four Seasons from the Taverner Players (Denon 38C37-7283), which already has a very astringent violin tone, became unlistenable.

Voices also acquired a slight lisp, which became accentuated at high replay levels, and the breathy edge to the sound of trumpet was also rather exaggerated. Though the second samples of the 3000 did not have the distortion problems that rendered the sound of the earlier speakers unacceptably hard at high levels, they still tended toward an overall bright balance. I found the effects of different cables to be easily audible through the 3000s, as well as the small changes in sound due to my altering cartridge VTA. The differences between the Chesky 128x-oversampling A/D converter and the industry-standard Sony PCM1630 on the Stereophile Test CD were also much more audible than via the Celestion SL700s. I assume that this is due to the 3000's accentuation of frequencies where the ear is already at its most sensitive.

Pink noise revealed that the listening axis I ended up with flattened the mid-treble balance at the expense of depressing the top octave a little when compared with the balance on the ribbon axis. This was with my ears level with the ribbon midpoint, which is where the listening chair placed them. Sitting more upright so that my ears were level with a point two-thirds of the way up the ribbon gave a more neutral mid-treble balance, though with even lower top-octave energy, which made the BSO Four Seasons sound too mellow. This sensitivity to listening axis means that getting the optimal treble balance will be very much a matter of experiment, though the basically wide treble dispersion of the ribbon does mean that those sitting to the sides of the central seat will still get a good share of the music's highs.

Whereas I had found the original samples to have rather disconnected lows even with the Mark Levinson amplifiers driving them, the second samples seemed more overdamped by comparison, being very much tighter in the bass. The transition to the lower midrange was very much better managed, though the tradeoff was a rather more reticent low-bass register. This, to my way of thinking, is the right way to go in musical terms. The first track on Simple Minds' The Amsterdam EP (Virgin SMXCD6) is a superb, if over-rich production by Trevor Horn and Stephen Lipson (footnote 5). Spectoresque in the way it floats tambourine and acoustic guitar over a thick goop of synthesizer and sampled drums to produce a majestic rhythmic sound, there is so much going on in the bass that it is hard for a speaker with any kind of undamped LF tuning to keep the musical threads separate. The Celestion passes this test in a satisfying manner.

Classical music fared equally well, timpani featuring a clean impact with a minimum of overhang. Percussion warhorse it may be, but Copland's Fanfare for the Common Man—my favorite recording is Antal Dorati with the Detroit Symphony, London 414 273-2—is a good test of a speaker's ability to retain control in the bass. The Celestion 3000 had less low-frequency definition than the SL700 with this recording, this perhaps a function of the near-wall placement, but it usefully had more overall bass weight than the smaller speaker.

To be charitable, I'll assume that the first pair of 3000s had either suffered on their long voyage to Santa Fe or were a "Friday afternoon" pair. It worries me, however, that if the latter is indeed the case, how could they have escaped the factory? The last time I visited the Celestion facility I was impressed by the intensive quality-control procedures inherent in their production lines. Perhaps the first pair had been manufactured while the company was still relatively low down on the inevitable production learning curve that accompanies the introduction of any novel technology.

This leaves me with the awkward job of reaching a valid value judgment based on the sound quality of the second pair. These were undoubtedly a much better-sounding pair of loudspeakers than the first. Yes, their treble was still on the bright side of neutral, but the overall sound was much more evenly balanced. Not only was the soundstaging more palpable, the mid-treble was better integrated with the low treble and the low frequencies were both faster and better defined. With the first samples, putting the too-high distortion to one side, the various aspects of the sound just didn't meld except, paradoxically, at the transition between the cone unit and the ribbon, whereas the second samples were effectively seamless throughout the audio band.

Beautifully made and incorporating a well-thought-out ribbon driver, the Celestion 3000 is an impressive design. Its seamless, coloration-free midrange, tight low frequencies, and exceptional transparency are excellent by any standard and are matched by a good sense of dynamics up to spls of around 100dB or so. Its unique design will also ensure that more than one listener will be able to hear both a well-balanced sound and a reasonably well-delineated soundstage. With the right kind of music and optimal electronics, a pair of Celestion 3000s will produce a sound that is vividly detailed, uncolored throughout the midband, and has excellent image depth.

Against these positive attributes have to be set its tendency to mid-treble brightness, which both leads to a sensitivity when it comes to choosing suitable electronics—the Audio Research combination worked much better than the Mark Levinson, for example—and makes the speaker rather unkind to anything but the best-produced recordings; and its intended near-wall placement, which makes optimal setting-up tricky. With recordings that are already bright, the wrong ancillary equipment, and without enough attention paid to placing the speakers in the room, the sound will be fatiguing, even unpleasant.

At $2198/pair including the essential stands, we are talking about an expensive loudspeaker. As with any imported loudspeaker, the price inevitably suffers from the economic facts of life, the domestic competition always having a large price advantage. The Celestion 3000 comes under strong competition from American models using traditional moving-coil drivers, the Vandersteen 2Ci, for example ($1455/pair including the Sound Anchor stands), and the Spica Angelus ($1275/pair), both of which are far more forgiving of poor recordings and ancillary electronics. And Magnepan's superb Magneplanar MGIIIa, which also features a ribbon HF unit, sells for almost exactly the same price and will play considerably louder. The 3000 also comes under strong fire from Celestion's own SL600Si, not significantly different in price when coupled with a pair of Celestion SLSi stands, and I must admit that the diminutive '600 (reviewed in Vol.12 No.5, May 1989) produces a sound that is more to my taste in the treble.

But if you value the importance of retrieving the maximum of detail and reverberant information from recordings and find the SL600Si and Vandersteen 2Ci to be too laid-back in balance and the Magnepan to be too colored in the lower midrange, the Celestion 3000 will be worth auditioning.

Footnote 5: Recommended to me by Martin Colloms, this superbly produced 3" single is worth looking out for. It also contains a cyberpunk arrangement of Sir Hubert Parry's pseudo-anthem Jerusalem that will stir mixed emotions in the blood of any true-blooded Englishman raised on rock'n'roll.