Celestion DL8 Bookshelf speakers
The DL8 loudspeaker makes some of the advanced technology used in Celestion's very expensive SL6 and SL600 speakers available in a much more affordable price range.
One of the most distinctive features of Celestion's SL600 speaker system, which we tested last year, is its thin copper-foil tweeter, whose dome and voice-coil former are made as a single, integral unit. Celestion used laser interferometry and computer techniques to study the modes of diaphragm vibration, and the combined voice-coil and dome structure was a result of that investigation. As compared with conventional tweeter construction, this design has reduced mass (for extended high-frequency response) and improved heat conduction (for higher power-handling ability). However, it is a costly driver to manufacture, requiring precision hand operations to mount the tweeter's integrated dome and voice coil in its extremely narrow magnetic gap.
The DL8 uses a similar tweeter, although instead of copper it is made of aluminum alloy, which actually reduces its moving mass still further. At 2,500 Hz, there is a crossover from the 1-1/4-inch tweeter to an 8-inch cone woofer operating in a sealed enclosure. The system is rated for a nominal frequency response of 50 to 20,000 Hz and a sensitivity of 89 dB sound-pressure level (SPL) for a 1-watt input, and it is recommended for use with amplifiers rated to deliver from 10 to 150 watts per channel.
The Celestion DL8, housed in a walnut-finish wooden cabinet, measures 19-3/4 inches high, 10-3/4 inches wide, and 10-3/4 inches deep. Its acoustically transparent dark-brown cloth grille is stretched on a plastic frame that snaps into retainers on the front panel of the cabinet. The heavy-duty metal binding-post terminals are recessed into the rear of the cabinet. Although they accept the stripped ends of wire leads, they cannot be used with banana plugs or similar connectors. Each speaker weighs about 22 pounds. Price: $500 per pair.
Lab Tests
Celestion provided us with a frequency-response chart made on a DL8 system (not one of the units supplied to us). To our surprise, our own measurements, both of room response and FFT quasi-anechoic response, correlated extremely closely with the Celestion curve. In fact, we have never before encountered such close agreement between our measurements of a speaker and those made by its manufacturer, even when the measurements were made on the same test sample.
For measurements and listening, we installed the DL8's on stands about 25 inches high, with the rear of the speaker cabinets about 18 inches from the wall, in approximately the position Celestion suggests.
The room response was among the smoothest we have ever measured, and the close-miked woofer response was an exact match to the curve supplied by Celestion. The two curves spliced smoothly and unambiguously, resulting in a composite frequency response of ±1.5 dB from 600 to 20,000 Hz. The woofer's maximum output was at 200 Hz, where it was about 5 dB above the average midrange and high-frequency level. From 70 to 20,000 Hz the composite response variation was only ±3.5 dB. Below 200 Hz, the woofer's output fell at 6 dB per octave from its maximal + 5-dB level and a steeper rolloff was introduced by the crossover design below 45 or 50 Hz.
Our FFT response measurements, made on axis at a 1-meter distance, closely matched the manufacturer's graph, showing a 5- or 6-dB overall variation from 200 to 15,000 Hz, a distinct dip to - 6 dB (relative to average midrange levels) at 16,000 Hz, and a rising output above that frequency. The high-frequency rise was due to the tweeter's resonance peak at 24,000 Hz, where the output was more than 10 dB higher than the midrange level. Because of the high "Q" of the relatively undamped resonance, the output at 20,000 Hz was essentially at midrange level.
Response measurements on axis and 45 degrees off axis showed moderate directivity in the horizontal plane that changed very little with frequency. The two measurements differed typically by 3 to 6 dB at most frequencies above 1,000 Hz, but the divergence was never more than 10 dB up to 20,000 Hz. The phase linearity of the DL8 was very good, with less than 0.8 millisecond of group-delay variation from about 300 Hz to over 20,000 Hz.
The speaker's impedance was 7.5 ohms at 200 Hz and about 7 ohms from 10,000 to 20,000 Hz. It reached maxima of 32 ohms at 80 Hz and 52 ohms at 1,800 Hz. Its sensitivity was exactly as rated, 89 dB SPL at 1 meter with a 2.8 3-volt drive level. Woofer distortion was measured with a 3.2-volt drive, corresponding to a 90-dB midrange SPL. The distortion was extremely low (0.3 to 0.4 percent) at 90 to 100 Hz, and it increased to 1.9 percent at 60 Hz and 5 percent at 40 Hz.
In power-handling tests with 1-cycle bursts at frequencies of 100, 1,000, and 10,000 Hz, the woofer cone "bottomed" at 100 Hz with a drive level of 25 watts, producing an audible raspy sound that was clearly due to excessive physical cone excursion. At 1,000 Hz, still handled by the woofer, the amplifier clipped at 280 watts (into the 28-ohm impedance of the speaker at that frequency) before the output of the speaker showed significant waveform distortion. Finally, at 10,000 Hz the tweeter showed its true mettle: our amplifier clipped at 1,100 watts without any obvious signs of distress from the speaker!
Comments
By any standard, the Celestion DL8 proved itself to be a very fine speaker. Measurements of a speaker, of course, give only a few clues to its sound quality. Nevertheless, in this case all our test data were so good, and so closely in agreement with Celestion's claims, that we must confess to being impressed. Listening tests tended to confirm our impressions.
Its sound was superbly balanced, with a solid bass that belied its small size, yet without undue upper-bass emphasis or boxiness. The middles and highs were seamless and so well integrated that we simply heard the total sound and gave up trying to hear the minor response variations that our microphone had detected. After all, we don't go to concerts to perform an acoustic analysis on the music, do we? Perhaps that is the best way to describe the quality of this speaker-thoroughly musical, with an insidious tendency to be so listenable as to discourage technical criticism. We made a special effort to be aware of the DL8's spatial properties ("imaging," if you prefer). Like any conventional front-firing box system, its sound stage is essentially confined to the space between the speakers, but it manages to fill that part of the room with a well-defined sound stage, mostly in the listening room but sometimes including the space immediately behind the wall. We also found that these speakers respond very well to spatial enhancers such as "sonic holography" and similar systems.
Finally, encouraged by our measurements of the speaker's power-handling ability, we put that to a practical test. Playing a record with very well-recorded percussion sound, we turned up the volume until the sound was about as loud as we felt comfortable with in a modest-sized room (it had considerable physical impact at that level). The speakers sounded beautiful, with no detectable compression of dynamics and a surprisingly visceral effect from the larger drums. The amplifier power during this time averaged perhaps 10 watts per channel, with frequent peaks to well over 350 watts! This is something we would not recommend with any small speaker, including these, but it certainly says something about their construction and quality.
There are many good speakers selling at the price of the DL8. Without attempting to establish which are "better" than others (if only that could be done!) let me say that I cannot recall hearing a speaker comparable in size or price to the Celestion DL8 that I would rather listen to for extended periods.
