B&W CDM2SE.

Manufacturer: B&W Loudspeakers of America, 54 Concord St., North Reading MA 08164-2699; 508/664-2870

Price: $800/pair

Source: Manufacturer loan

Reviewer: David R. Moran

Two-way loudspeakers have an undeservedly high opinion among audiophiles. The reason has to do with a misinformed "less is more" ideology: fewer drivers; one, allegedly simpler crossover; less "phase" and "timing" distortion (to which we are demonstrably deaf); greater overall simplicity; and so on. All of which ignores the real problem: asking two drivers to do too much over too wide a range.

The result is never enough bass unless the woofer is large; never enough airiness unless the tweeter is small; and a lumpy stitch where they join -- unless the woofer is small and the tweeter is large. But a smaller woofer doesn't reach down low enough (at least not with enough output), and one large enough to go low also begins to beam as you take it up higher to where the tweeter will flare in if the tweeter itself can reach down low enough to meet the woofer, except that if it can do so easily, it will be too large to have good dispersion in the top octave or two, whereas if it does have good dispersion it will get cooked by upper-midrange energy as its small self reaches down low to meet that woofer -- and so on and yada yada: you get the picture.

In other words, with a two-way, by the nature of the beast, you simply do not get adequate low bass and seamless splicing and airy highs. (This is why God made multiway designs.)

Nonetheless, it has been fascinating over the last decade to see how designers have juggled these few intractable parameters, and it's been gratifying to see how two-ways have gotten better and better within their limitations. The very best now actually have good LF performance, good highs, and fairly smooth sound and imaging at the stitch. And the still-pretty-good designs, like the B&W CDM2SE under review here, have solid overall performance, performance one could certainly live with given compromises, like there being no other choices.

The reason this newcomer to these respected pages wanted to look at what B&W achieved in a conventional two-way is that this English company's fancier implementations have always been worthy of scrutiny. No, not so much that shiny kilobuck snail thing, and not that big totem pole thing, but the classics: the B&W 801 and especially their 802 series.

The 802 designs, while never voiced or balanced quite right, nonetheless have a most beautiful horizontal radiation pattern: smooth, virtually seamless, resulting in altogether focused imaging. (In my experience, only the KEF UniQ designs and their knockoffs, and an occasional Genelec, do as well among multiway systems.) The chief reason is those signature, uniquely small polyhedral cabinets for midrange and tweeter. While the 802 may have needed EQ, its soundstage is at once floaty and consistent. In subsequent B&W lines, including the bigger siblings in this new CDM line and most of the new expanded Nautilus line, the company has kept some of the virtues of proper "ball"-sized baffling for its otherwise conventional 1" tweeter.

The CDM2SE has none of this, though. It's a rectangular box, with woofer and tweeter on the same forward-facing surface just like almost every other two-way. The B&W literature does mention steep crossover filters, plus a nose-cone plug-type woofer voice coil dust cap that's claimed to enhance dispersion at the crossover point, plus the usual stuff about their famous yellow woven Kevlar 6.5" driver. I was eager to unpack the pair and fire them up.

The vented cabinets, in beautifully finished red or black ash, are quite hefty. You can slightly change the bass tuning with a foam cylinder supplied to plug into the rear port hole, and the terminals are biwirable for those who believe in that.

I set them up rather high on stands and well away from the front and side walls, staggering those three near-boundary distances so as to reduce any lower-midrange suckout, and listened happily for a month or two. (In a first for loudspeaker companies in the past decade or two, the B&W literature also mentions bookshelf usage, which I did not try since nobody does it anymore, even though it always produces flatter response than on-stand placement. Note that rear-vented systems used in bookcases should not be jammed in all the way up against the back of the bookcase or wall.)

The supporting equipment I used in my listening sessions included the dbx BX-1, Audio Dynamics B-200, and Sony 3200 power amps connected to the speakers by Radio Shack 12-gauge wire; dbx CX-1 preamp and AR SRC remote/switcher; and dbx DX-5 and Radio Shack Optimus CD players.

My auditioning notes are to the point: good job, nice sound. Quite good sound for a two-way, in fact, although not great absolutely; ditto bass, ditto smoothness, ditto detail, ditto "resolution," ditto imaging. Somewhat airless highs, as from all 1" tweeters on a woofer-width baffle. Solid upper bass and lower mids. I liked what I heard as I delved with pleasure through a range of repertory, including Brahms's Four Symphonies in the new Telarc recording by Sir Charles Mackerras and the Scottish Chamber Orchestra, The Celtic Album by Keith Lockhart and the Boston Pops Orchestra (BMG/RCA Victor), Eric Clapton's Pilgrim (Reprise), Days of the New (Outpost), plus a range of classical FM radio selections, including live broadcasts of orchestral concerts, e.g., the BSO.

Unfortunately, the previous three two-ways I have auditioned and measured in identical circumstances have all been superior performers. The now-defunct RA Labs' MicroRef had a smaller tweeter for airier highs, plus ruler-flat response overall. The NHT SuperOne ($350/pair) is most like the B&W (6" and 1" drivers) but is marginally better in horizontal radiation pattern and hence imaging. And the Genelec 1030A ($2,098/pair) is a stupefyingly smooth and wideband performer featuring crossover work, plus decisions as to driver width and baffle smoothness, that gives its radiation pattern an evenness and congruence right up at the B&W 802 or KEF UniQ level.

Overall, the B&W is not quite equal to those guys, and it's expensive to boot (at least, that is, compared to the NHT SuperOne): $800/pair. But it does a good job. Let's look at the measurements:

In-room Power Response (Figure 1): My protocol is to feed the speakers two channels of uncorrelated ("stereo") pink noise and measure their frequency response with a precision, steep-filtered, continuously averaging 1/3-octave real-time analyzer (RTA). A 0.25" B&K microphone is moved slowly throughout a typical listening position: 8-10 feet away at seated ear height. For the duration of the measurement, the microphone moves about a space of approximately 1.5 cubic feet, side to side, front and back, up and down, continuously averaging within the listener's head area. I repeat this in a couple of listening rooms (typical domestic American) at favorite and other likely listening positions.

[GRAPH OMITTED]

The resultant response measurements are themselves averaged. I compare these averages with the individual measurements to ensure fair and accurate representation of the speaker's frequency response performance. This in-room kind of averaged power measurement is accepted as closely correlating with listener perception of overall timbral or spectral balance and of tonal accuracy.

Each separate room and listening position measurement was quite close to this average shown in Figure 1, including measurements of the individual left and right speakers (speaker quality control is another thing that has really improved over the last 15 years). I would conclude that the response of the B&W CDM2SE pair is 60 Hz-20 kHz +/-3.5 dB, which is really nice, only a hair looser than the company's own published specs. (All these graphs' vertical divisions are 5 dB.)

Note particularly the substantially smooth quality of the mid and upper bass, the gratifyingly slight magnitude of the stand-caused lower-midrange dip (250 Hz), the odd small peak around 600 Hz (which showed up consistently), and the tininess of the post-crossover peak at 4 kHz. Impressive, and it sounded it.

Horizontal Radiation Measurements by Angle (Figures 2 and 3): To measure the speaker anechoically by angle, I take it outside and put it on a stand that places it at seated ear height. This environment is completely reflection-free across the entire audio band except for having a "floor." Pink noise is again used, and my backyard test acreage is suitably quiet. (I once told one of my former audio engineering colleagues about this protocol of mine, and he replied that when the gang got together next, he would be sure to tell everyone that, as a reviewer, I was out standing in my field).

[GRAPHS OMITTED]

Standing 8-9 feet away and holding (waving, actually) the mike 32-40 inches above the ground (seated ear height) and 1-2 feet to and fro, I take a measurement at 0 degrees (on axis), continuously averaging across +/-15 degrees horizontally to either side of the exact "slice" point. I repeat this measurement at 30, 60, 90, 135, and 180 degrees off-axis. Sometimes I take a measurement 45 degrees up to look at vertical radiation (not shown for the B&W; there were no particular anomalies).

In a room, the off-axis responses, especially the 30, 60, and 90-degree ones, become lateral reflections that form, or at least help form, the stereo images. And not only does the horizontal radiation pattern correlate with imaging properties, it also reveals starkly the decisions and tradeoffs the designer had to make about driver size, baffle width/smoothness, crossover point(s), and filter implementations.

Figure 2 pretends that the axis response is ruler-flat ("normalized"). It therefore shows the difference or deviation in horizontal-plane output as a function of off-axis angle, and may be considered the "imaging" curve set.

What is wanted here is even spacing and uniform, congruent striation, which help produce more constant and consistent imaging. The sidewise responses can dive early (starting in the midrange; very hard to achieve) and stream along in nearly parallel lines, or, more typically, they can arc, one after the other, each starting downward at a higher frequency. There are various philosophies about this, and each style has its own sonic identity, its own kind of beaminess, and its own way of casting images.

Note that some of the desired evenness is in fact present: the distances between angles are regular, and while a bit rough the responses are not all that rough, and have more dips than peaks -- flare-ins that can pull the imaging this way and that. Altogether this pretty decent off-axis output is what accounts for the B&W's fairly steady stereo imaging, I would suggest.

Superior two-ways such as the aforementioned models are less lumpy, though. With superior multiways, sometimes, you can hardly tell there are crossovers, or they show up mostly on the response curves taken toward the back of the unit or directly behind it, at what I label -135 or -180 degrees. Note that the CDM2SE crossover, where the tweeter comes in, appears to be more like 4 kHz than the 3 kHz specified; I find it surprising that B&W designers would not take a 1" tweeter an octave or two lower.

Raw Horizontal Radiation Measurement (Figure 3): Figure 3 is the raw (not normalized) measurements of the B&W outdoors on a stand, taken where an audiophile would sit. The on-axis response, second from the top, shows a tweeter flare-in peak centered on 4 kHz and a dip across the next octave: if it were true, as many "experts" allege and publish, that on-axis loudspeaker response counts for much to our integrating ears, the CDM2SE would sound intolerably "presency," which it doesn't. If you do sit awfully close and direct sound does therefore play a proportionally larger role in overall timbre, note how utterly flat the -30 degree response is. Toe these babies in toward each other and move up!

The lower-midrange notch in the 250-300 Hz range is caused by cancellation by the reflection from the ground. This will get filled in (and/or it typically moves down an octave or so) whenever the speakers are in a room and placed at staggered or unequal distances from the three nearest boundaries: floor, side wall, front wall. Such notching (called the Allison effect) is problematic only when the speakers are placed closer to equidistant from those three near boundaries. Realize that stand usage invites the problem, however.

Observe finally in both figures 2 and 3 the sizable drop in treble output as you move merely from -30 to -60 degrees off-axis, even more than is usual for 1" tweeters. This drop gives a pronounced forward orientation to the treble, putting much less of it into the reverberant field, with the result being tighter treble imaging, sure, but hand in hand with much less spaciousness and air in the harmonics -- a certain dryness I do not much care for.

Other Performance Aspects: A speaker's performance can be completely characterized by its amplitude response, both at a given angular slice and as its total output over a range of angles; its phase response; and its distortion as a function of input. How we hear these things, and how much less significant the last two are (when above threshold) compared with the first, are also well-known. When pertinent, I do look at bass distortion (not in this case, though), but since no one can show that a modern loudspeaker's "time-domain/"phase" performance is important to the sound -- quite the opposite, in fact -- I make no measurements of group delay or the like, nor of impedance and reactance nor sensitivity, since dealing with load issues is the amp's job. The B&W impedance spec is 4.5 ohms minimum, 8 ohms nominal, so there should be no problem. As for power-handling, I did not really crank it unduly, lest I cook it, but it went plenty loud in a typical 24' x 13' x 8' living room with one archway into the rest of the house and a second into a study. It did similarly well in an 18' x 20' cathedral-ceilinged family room.

Concluding Thoughts: It can be frustrating for readers to read a review that lamely ends with something along the lines of "If you are shopping for models along these lines, be sure to audition the (whatevers)" -- and this review will be no exception. The bottom line with the B&W CDM2SEs is that if you were rounding out a B&W home theater system, for example, these certainly should be considered for the side and rear units (or for all five!). Or if you very much wanted a smallish (albeit hefty), good-sounding conventional on-stand design in this very handsome finish, they're a fine choice. And if you were shopping and wanted to compare similar models from NHT (e.g., the SuperOne -- indeed, the B&Ws are so similar to the less-expensive NHT SuperOnes that I would be very interested to hear from any readers who have A/B'ed the two models in a controlled comparison, if only to see whether they agreed with my assessment of the speakers' relative imaging characteristics), Paradigm, PSB, Boston Acoustics, Polk, or the myriad other manufacturers of straightforward two-ways, then do be sure to include the B&W CDM2 Special Edition on your audition list. -- DRM