Columns Retired Columns & Blogs |
with similar colorations and similar stellar imaging attributes. I loved my TC-50s and regretted selling them. What they did well, they did REALLY well.
Such was the case with the Delaware Acoustics DELAC S10, which costs $629/pair. Only sold factory-direct, this would therefore have been low on Stereophile's priority list for review if it weren't for two things: first, the fact that the S10 was designed by one Ralph Gonzalez, a name that should be familiar to readers of Speaker Builder magazine for having written a very useful speaker design and analysis program; second, as implied in the first paragraph, the S10 is one of the weirdest speakers I have ever laid ears on.
A floor-standing model 42" tall, each enclosure is just 5" wide and a minuscule 3" deep. In fact, as can be seen from the diagram of the S10's cross-section (fig.1), the enclosure is only just large enough to accommodate the drive-units, these sourced from the Norwegian SEAS company. A 0.75" ferrofluid-cooled, plastic-dome tweeter crosses over below 4500Hz to two 4.5" bass/midrange units, these appearing to be doped paper-cone units featuring cast magnesium chassis. The use of large 15oz magnets in this enclosure size results in a fundamentally overdamped alignment, which should give both good time-domain performance and a lightweight balance. The S10 therefore needs to be used with a passive line-level equalizer, the E10, which fits either in a preamplifier tape loop or between pre- and power amplifiers, to extend the bass to a useful degree. (Mr. Gonzalez claims the response to extend smoothly to below 40Hz under normal listening conditions.)
As the E10 is wholly passive, the boost in the bass is actually achieved by cutting the level of the midrange and treble, resulting in a considerable insertion loss, up to 10dB, which may be unacceptable if your system has limited overall gain (as would be the case if you use a passive control center and low-sensitivity power amplifiers, for example). As well as in/out sockets, the E10 has a tape loop with switchable source/tape monitoring to replace the one it occupies on the preamp. It also has the degree of bass EQ adjustable with a front-panel switch, the "50Hz" position giving a greater degree of effective LF boost, thus greater insertion loss, than the "80Hz" position. Being passive, the actual degree of tonal modification provided by the E10 will depend to an extent on the source and load impedances of the components with which it is used. Delaware Acoustics therefore provides two versions: "L" has an input impedance of 4k ohms, and is for use with preamps having output impedances of less than 1200 ohms and power amplifiers with input impedances of more than 12k ohms; "H" has an input impedance of 40k ohms, and is for use with preamplifiers with output impedances of less than 12k ohms and power amplifiers with input impedances of 100k ohms or more.
Construction of the S10 and E10 appears to be to a very high standard, both drive-units on the front of the real-wood veneered enclosure being rebated. The cabinet is filled with polyester fiber and its top is sloped back to bring the acoustic centers of the drivers into approximate time alignment. This, in conjunction with the computer-optimized, first-order crossover slopes employed, should give a performance that reproduces wave shapes without significant distortion; ie, harmonics should arrive at the listening position in the correct time relationship with the fundamental. The two bass/midrange units are driven with the same polarity. Mr. Gonzalez says that this should give a figure-eight directivity pattern, but without the rear lobe being in inverse phase, as is the case with typical panel speakers.
I mused over this statement for a while, concluding that, yes, the response would be a figure-eight, but only when the wavelength of the sound becomes less than the enclosure width, which will be at the top of the midrange unit's passband, above 2kHz or so. Below that frequency, there will be increasingly more output to the sides due to diffractionthe lower the frequency, the more the drive-units radiate in an omnidirectional manner. It might be thought that when the distance between the two cones is an odd number of half-wavelengths, the outputs from the front and rear drivers will cancel: a cabinet depth of 3.5" would result in a first cancellation frequency of 1935Hz. As the drivers are becoming directional at this frequency, I would not have thought it to be a major problem. In any case, what would be the next cancellation frequency, 5780Hz, is both above the woofer's passband and in a region where its output will be very directional.
An advantage derived from the use of two woofers in this manner, however, will be the fact that the cabinet will be stressed by opposite and equal reaction forces in the bass and midrange, thus removing any smear of the sound due to cabinet wall motion. The bottom of the enclosure is filled with sand, and a 10" by 10" wooden base provides stability. This may be fitted with carpet-piercing spikes or with square-head screws (both supplied).
The crossover, which uses a single air-cored inductor in the woofer feed and three parallelled polypropylene capacitors in series with the tweeter, is mounted on a glass-fiber circuit board, with thick and wide traces, attached to the cabinet rear. High- and low-pass legs are electrically separate and two sets of binding posts on the outside allow single- or bi-wiring, and bi-amping. The crossover circuitry is completed with an RC Zobel network for each leg. The E10 contains a single pcb attached to the in/out socketry and also features high-quality components, including polypropylene caps. The well-written instruction book, which covers all aspects of installation and use, shows how to reduce the tweeter level by 3.5dB if the sound is found to be too lively, as well as covering bi-wiring and bi-amplification using the internal crossover. Unusually, it also goes into complete detail on how to bypass the S10's crossover and replace it with a line-level electronic crossover for maximum performance when bi-amped (though the two-year warranty will then be null and void).
About the only aspect of the S10 I found unappealing were the foam grilles over the front-mounted drivers. Though about as acoustically transparent as grilles can be, aesthetically they look like an afterthought. I left them off for the auditioning.
Sound
The DELAC handbook recommends placing the S10s 18" or more from the rear wall, angled so that their axes cross behind the listener. This, therefore, was how I carried out the auditioning, and it turns out that listening slightly to the side of the main axis does give the smoothest response. I use pink noise to reveal general balance idiosyncrasies: the S10 didn't fare as well as I had expected with this test signal. With my ears on a level with the cabinet top, a strong rise in the mid-treble could be heard. This reduced in level as the listening axis dropped, though on the natural axis given by the speaker's height, level with the baffle angle, the sound had a distinct "double-humped" character, coupled with a strong "aww" coloration. The best integration between the drive-units, with the fullest low treble, could be heard below the woofer axis, but this represents an unrealistically low listening position.
Turning to spoken male voice, this was refreshingly free from any kind of chestiness, though the upper mids were characterized by some "cupped-hands" coloration. But the precision of the vocal image stunned me. With a monaural vocal recording, the image of the voice approached the paradigm, a dimensionless point midway between the two speakers. Recorded depth was also well decoded, the S10s throwing a wide, deep soundstage. On naturally miked orchestral recordings, woodwind choirs could be heard to be set back behind the violins but in front of the horn and brass sections. And when Mike Skeet announced that he was standing outside his infamous garage door on the HFN/RR Test CD, he was outside, behind the rear wall of my listening room. The sound of the S10s was consistently spacious, even close-miked recordings taking on an attractive bloom. Listening to my recording of Anna-Maria Stanczyk playing Chopin on the HFN/RR CD, however, I started to feel that this aspect of the speaker's imaging performance was, if anything, a little exaggerated. The piano image was set farther back in the soundstage than I had heard before, the recording more reverberant than I remembered from the time of committing it to tape. I suspect that the delayed reflections from the rear-firing woofer/midrange unit are contributing to this always enjoyable effect, but, strictly, it departs from ultimate accuracy.
Tonally, the bass was lightweight, even with the equalizer set to "50," though it was very well defined at moderate levels. At SPLs over 88dB or so, orchestral recordings became rather congested. Dynamics, too, were good only up to a point, though below that point, there was a refreshing immediacy to the sound. Dull these speakers aren't. (Dull is here meant in musical terms, not the opposite of brightthe S10's highs are smooth and extended.) There did seem to be a slight lack of energy in the presence region, but the transition between the drivers appeared to be well-managed.
But I kept returning to the degree of coloration in the midrange, audible as a woodiness to piano sound, a cupped-hands edge to voice, and a rendering of violin tone more like that of a viola. I admit that I am less bothered by minor tonal imbalances than, say, J. Gordon Holt (in whose ears we trust), but the S10 was too flawed in this respect for me to want to continue listening.
Stereophile's measuring techniques are not yet sophisticated enough to suggest a direct reason for the S10 to sound imbalanced in this mannerit does have slightly too much midrange energy on-axis, but not any more than the less-colored-sounding but similarly measuring Rauna Freja which I reviewed in January. However, I do have a vague memory of some other speakers featuring a similarly narrow enclosure that also suffered from this kind of midrange problem. I don't know if there is any research to reinforce this idea, but I have a gut feeling that once a speaker's baffle gets narrower than the width of the human headafter all, it is the radiation pattern from a head-sized baffle that we are most familiar witha degree of suckout in the midrange becomes necessary to prevent the otherwise overly wide dispersion making the room sound too midrange-oriented.
Conclusion
I have to say that overall I had expected more from the DELAC S10, considering the authority with which its designer writes about loudspeakers, its excellent build quality, and the superbly comprehensive handbook. Yes, it does have a superbly spacious presentation, with beautifully precise soundstaging, both concerning depth and lateral imaging, but I just couldn't come to terms with the speaker's tonal balance and the lightweight nature of the bass. I think that the task Mr. Gonzalez set himself, to produce a full-range speaker using small, fast midrange/woofers in such a small enclosure, was too extreme to be ultimately successful in every area.
While I can't recommend the S10, therefore, it does show sufficient design promise that I suspect that a more conventional design from Mr. Gonzalez would prove to be a significant contender.
with similar colorations and similar stellar imaging attributes. I loved my TC-50s and regretted selling them. What they did well, they did REALLY well.