Thiel CS2 loudspeaker
So, like every other reviewer, I cannot claim to have surveyed all of the vast assortment of speakers in a given price category. I rely on word-of-mouth to select speakers that really merit reviewing on a high-end basis. This may not be fair, but it reduces an otherwise unmanageable list of manufacturers to about 40—even allowing for the annual burst of speaker manufacturers that don't go beyond producing a few prototypes, or who stay in business only a year or two.
I wouldn't want to arrange that list in order of importance, but I suspect that virtually all the other members of the list would include Jim Thiel as one of the manufacturers/designers who rank high in the "top forty." This makes any new Thiel speaker of more interest than most, and the new Thiel CS2 particularly so. Jim Thiel's try at a state-of-the-art cone speaker at the relatively affordable price of $1350 a pair, the CS2 needs no special drive electronics, bi-amping, or fussy placement.
In fact, if it is at all within your price range, the Thiel CS2 is one of those speakers which clearly merits a personal listening session.
The Thiel CS2 use three drivers: a 28-mm soft-dome tweeter which extends smoothly beyond 35kHz, a 100mm cone midrange with a bandwidth of over seven octaves, and a 210mm woofer with a long-throw voice-coil and very heavy magnet. All use modern cone and surround materials, and appear very well made. The keys to the Thiel's outstanding performance lie, however, in the cabinet and crossover.
The cabinet is heavy and rigid, with 30mm thick walls, a sloping front to time-align the drivers, and a molded baffle to minimize diffraction effects. Many speakers have somewhat similar features, including Thiel's own CS3. The modeled front mounting board for the CS2's drivers, however, succeeds to an amazing degree.
The resulting combination of driver characteristics and driver mounting results in coherence of midrange dispersion that rivals the Quad ESL-63, and the CS2 notably outperforms the Quad in dispersion over the top three octaves. The midrange to tweeter integration is particularly outstanding, sounding better than Thiel's own CS3; that makes it better than any cone system I've tested.
As for the crossover, it uses polypropylene and polystyrene capacitors and air-core inductors in fashioning a 6dB/octave slope. This first-order crossover provides superior time and phase alignment. As has become all too clear with the advent of digital sound, phase distortion is audible. Minimizing such distortion requires both time and phase alignment in order to provide optimal coherence, particularly in terms of transients, imaging, and harmonics.
Most speakers use at least a second-order crossover (12dB/octave), which inherently involves 180° of phase shift at the crossover point (the midrange and tweeter, for instance, are in opposite phase) if the crossover is to have flat frequency response. Some speakers use crossovers of fourth-order and above, and a fourth-order produces at least 360° of phase shift (footnote 1).
This lack of phase alignment is one reason I develop a nervous tic every time I hear the term "absolute phase." Even a first-order crossover relies on the cancellation of phase shifts of up to 90°. A phase reversal switch may produce better "average phase," but never "absolute phase"—at least by the time it reaches the human ear.
In any case, the Thiel CS2 is a cone speaker that really does have good time and phase alignment. Even most speakers which claim to be time- and phase-aligned rarely approach the phase alignment of a first-order crossover; most use crossovers with far higher amounts of phase shift. Worst of all are those who claim time alignment without mentioning phase alignment. The technical merits of "time alignment" without phase alignment are tenuous at best, and my experience to date has been that any design that time-aligns or "steps," its drivers, without making claims to phase alignment, is making visual changes rather than sonic ones.
Thiel naturally claims that all this technology produces exceptional detail and imaging, flat frequency response, and outstanding step response. I can't confirm all of these technical claims, but the sound lives up to them; musically, the CS2 is outstanding. Let me give you a few comparisons and insights:
•Side-by-side comparison with the Quad ESL-63 shows that the two speakers are very similar in their ability to reproduce detail without exaggerating timbre. Both are very transparent. The Quad has a bit more resolution, but the Thiel has more power-handling capability, flatter bass, and much better dispersed upper octave response.
•The Thiel CS2's bass power response is very good to excellent down to about 42Hz. This may not sound unusually deep to audiophiles who do not routinely measure speakers, but a one-third octave warble tone generator and a 125–20Hz sweep show that the Thiel produces flatter and more powerful bass than most of its competitors.
Further, 42–45Hz is far deeper than most audiophiles think. I should also note that the CS2's bass does not fall off a cliff at 42Hz (footnote 2). The Thiel provides some feeling of deep bass even if its power response is not suited for the far depths of the contra bassoon (29Hz), piano (27.5Hz), or organ (16.3Hz). Response that low, I regret to say, still costs at least $1000 more, even by means of subwoofers.
•The CS2 is unusually free from bad room interactions or bass standing waves, and in a wide range of locations. It lacks a bass "bump," and this may initially make it sound bass-shy to those used to less flat designs. If the signal has bass above 40Hz, however, the speaker will do an outstanding job of reproducing it.
•The transition from midbass to mid-midrange is exceptional, with no feeling of warmth or leanness. Timbres are correct, the midbass as tight as the low bass, and the midrange is remarkably detailed without drawing attention to particular instruments or vocal characteristics. There is no audible transition between bass and midrange. This is normally a special strength of electrostatics, but the Thiel CS2 shares it. It is a pleasure to listen to a speaker which, in this region, produces so few sonic surprises from record to record.
•This same "seamlessness" extends from the midrange to beyond audibility. The CS2 also has less apparent treble energy than previous Thiels (which had a little too much); there is no trace of hardness, yet no loss of detail. The highs in the Thiel CS2 lack the full "life" and "air" of the best ribbons, electrostatics, and Infinity's EMIM/EMITs, but they are very natural indeed.
•The imaging and depth are coherent, tightly focused, and exceptional. The Thiels justify inch-by-inch experimentation with placement. They float the soundstage with even halfway reasonable placement, but every change in toe-in, adjustment of distance from adjacent walls, and change in distance between speakers will have sonic results. They work very nicely six to eight feet apart, three feet from the side and rear walls, and slightly toed-in. They also benefit from spiked feet (provided), although they are sufficiently heavy and well-built to lessen the benefit of such spikes.
•The overall size of the sound image does not match that of bipolar, planar, or line source speakers, but the CS2s do behave as a relatively large, for a cone speaker system, apparent point-source.
•Dynamics and transients are also rendered very well; in fact, the CS2 is so good in these respects that it is even "analog ready"! Unlike some early Thiels, the CS2's excellent handling of low level transients and dynamics extends solidly up to major orchestral and sane rock levels. The CS2 does not have the rock power of the JBLs, and the Infinities, Apogees, and several other high-end speakers do better in terms of dynamic speed and detail. Nevertheless, the Thiel CS2s do very well.
•The lack of the electronic equalizer used in some previous Thiels seems to completely free the CS2s from the touch of transistor dryness that still appears in the CS-3s. For this reason I prefer the CS2 to the CS-3, in spite of the latter's notably better bass.
•The CS2 will definitely benefit from a high-powered amp, simply because it will then produce better dynamics. It did very well, however, with any amp I used. A PS Audio 70-watter was more than adequately powerful, and excellent in sound quality. Be careful, however: the CS2 will reveal the qualities of your electronics and cartridge. If the highs that go in are hard, they will come out just that way. No "Boston sound" here—this is a flat upper-octave power response.
•Finally, the CS2 is an exceptional speaker for audiophiles who still like to listen to natural-sounding solo instruments and voices. Far too many speakers show a distinct preference for one group of instruments or voice over the other. The Thiel shows no such partiality.
Summa Speakerconnundrum Scholastisolicism
Having justly praised the Thiel CS2, I come now to the little matter of how it compares with the competition. Though, as I said before, no reviewer can authoritatively rank all the speakers on the market, I can readily state that the CS2 ranks with anything I have heard in its price range.
The CS2 is directly competitive to the Quad ESL-63. It outperforms any of the comparably priced British monitors I have heard recently, although competition is stiff from firms like Proac and Spendor. In fact, the term "monitor" is unusually applicable to the Thiel CS2, as it provides enough accuracy for serious recording or playback. It is good enough to be a full-range monitor speaker in every area but deep bass.
If you like the sound character and coherence of the Thiel, compare it with the Vandersteen 2C and Spica TC-50 with subwoofer. The latter two speakers share some of the same general design concepts, but offer significant differences in sound character. Beyond that, the spectrum of good candidates extends from Acoustat to VMPS, and even I am not going to lengthen my reviews to Infinity in order to detail all those comparisons.—Anthony H. Cordesman
Footnote 1: The above statements are correct with respect to Butterworth filters, the simplest and most commonly used filter design. Other forms of filters have different phase and frequency response characteristics. A Bessel fourth-order, such as the one used in the Spica TC-50, is a minimum-phase (linear phase characteristic) filter, and can be used in combination with a Butterworth to yield virtually perfect phase linearity, as in the Spica. An analysis of all the different filter possibilities and their characteristics would require a hefty book.—Larry Archibald
Foootnote 2: Just a minor note for rock fans: It is terribly easy to confuse bass power with bass frequency. The lowest bass note common in rock music is the bottom of the double bass, which occurs at 42Hz. While I notice an increasing use of very–low-frequency synthesizer notes—largely on rock and pop drawn from movie soundtracks—most rock music exhibits very little true low bass; many pressings seem to have been mastered with a high pass filter around 40–60Hz. Don't expect even the best high-end subwoofers to give you sound like a rock concert—the power response from rock speakers is usually equalized to provide dramatic peaks well above 40Hz. It is the pure loudness of the resulting midbass that gives rock its usual punch.—Anthony H. Cordesman