Introduction. I originally requested the loan of an SACD player so that, at long last, I could hear the mythical DSD layer of my Telarc hybrid discs. I already knew the superb quality of the PCM layer but I had never heard DSD sound. Finally, I was going to hear for myself what all the SACD fuss was about. In the process I learned a little about design decisions, sigma-delta conversion, component voicing, and so forth, but as in life generally, what I ultimately discovered was not quite what I had anticipated. Living with the Classé Omega SACD 2 turned out to be more about hardware and circuit design, that is to say, about sound, than about one format versus another.

I own what is widely acclaimed one of the finest CD transports ever made, the Accuphase DP90. The DP90 originally sold for $8,500 and a dealer once told me that there are probably only some dozens of these units in the whole of the United States. This forty-five pound tank feeds a Bel Canto DAC2 converter, a small, unpretentious-looking black box I know intimately, having spent hours over many months chatting with the gentleman who designed its circuit board and layout. A company with a different design philosophy than Bel Canto might have put the same circuit into a rack-size, extruded aluminum chassis, anodized midnight blue, with colorful lights and digital readouts and switches and lots of connectors on the back—and charged three times the price. The DAC2 has a deservedly fine reputation, particularly at its price point. (There is a recent review of the DAC2 in StereoTimes.) It took time, patience and debt to acquire such a front end, and guests, some audiophiles, some not, have agreed with me that I have put together a great sounding stereo. I was content.

Then the Classé SACD 2 arrived. I burnt it in for a solid hundred hours, determined to wait before listening. However, occasionally curiosity got the best of me and I would turn on the preamp and listen to the music, albeit casually. And it seemed to me right off that I was hearing something special, and it didn’t seem to matter whether the source was an SACD or a CD, a CD with extraordinary, or merely average, sound engineering. But I held off coming to conclusions until after burn in. Subsequent listening confirmed my initial impression. This was the finest stereo sound I had ever heard in my living room. Things, of course, were not actually that simple, that black and white. It would be more accurate to say the SACD 2 was the most enjoyable and intimate stereo sound I had ever heard in my living room.

The Machine. The SACD 2 is part of Classé’s top-drawer source and amplification products, the Omega series. It’s big brother, the Omega SACD, a machine highly lauded in the press uses the Sony SCD-1 transport mechanism. (You may remember this drive from the first Sony SACD machine, the one that sold for the staggering price of $5,000. The SCD-1 moves the disc across a stationary read head, rather than the other way around. This is a reasonable design decision, but it always reminds me of a suggestion my friend Daniel once made: to make car wheels of cement and the roads of inflated rubber.) The Omega SACD has an external power supply, flexibility of input and output, and weighs 50 pounds. Oh, and it costs a bit more than the SACD 2: $12,000 vs. $8,000.

The SACD 2 uses the Sony 555ES transport, a unit that bears an uncanny structural resemblance to the drive in my Accuphase DP90. It has an internal power supply employing two 55VA custom toroidal transformers, 84,500uF of filtering capacitors and, presumably, multiple stages of voltage regulation. The SACD 2 chassis is made up a recessed sheet metal bottom section, front and back of machined aluminum, top and sides of 3/8” thick aluminum panels. The front panel is convex with a concave mid-section. This mid-section is black anodized and the black anodizing runs back across the top of the unit. The rest of the chassis is anodized silver, except the bottom, which is painted black. The readouts are classic LED red, easily legible at a distance. The remote control allows the option of turning the display off. The LEDs go extra bright momentarily when the unit receives a command from the remote. Front panel controls are Stop, Play, Track Forward, Track Reverse, CD/SACD, and Load. I miss the presence of a Pause control, for which you must use the remote control. The main power switch (located on the rear) is intended to be left on. After a few idle minutes, the display goes blank, and a second or two later a single, small LED lights. This stays lit to indicate that machine is in standby mode. (I do not know what other circuits, besides the display, may go into standby.)

[I can categorically state that, played on the Classé Omega SACD 2, the SACD layer of these Telarc discs sound better than the vast majority of CDs I own.]

The remote control is made of black anodized, extruded aluminum. It is well designed, attractive (it has a large “Ω” engraved on the bottom half), robust and has a nice heft. The SACD 2 weighs 30 pounds, is compact, and is very pleasing to the eye. For me, pride of ownership extends to what’s inside the box as well, and the internal component and construction quality of the SACD 2 is superb. Classé point out in the technical description of the SACD 2 that only the highest quality components are employed in critical circuits. When one considers just how critical are the analog drivers, or the SACD and CD filters, one readily appreciates Classé’s design decision.

In fact, throughout the technical description there is an obvious devotion to signal purity and sonic excellence. The SACD 2 uses the same stereo sigma-delta DACs as the Omega SACD, the Crystal CS4397; three of them, two for AES/EBU (balanced) output and one for RCA (unbalanced) output. Totally separate signal paths are maintained up to the DAC stages. CD signals from the transport mechanism are routed through an upsampling filter. This circuit is user-configurable, via the remote control, to sample rates of 44.1kHz, 48kHz, 88.2kHz, 96kHz or 192kHz, all interpolated to 24-bits. More importantly, this stage discards the S/PDIF clock signal, so any jitter arriving from the 555ES mechanism is irrelevant (the Bel Canto DAC2 converter does the same thing). Of the filter used for CD signals Classé writes:

It performs up to 8-times interpolation with moderately slow roll-off characteristics. The result is a finely tuned balance between frequency response and time domain response, giving optimum sonic performance. A separate processor is used to decode HDCD encoded discs. Although it can also perform 8-times interpolated filtering, we route HDCD decoded signals to filter/DAC combination for optimum performance.

A note about PCM digital output: the S/PDIF signal is null unless the SACD 2 is set to a 44.1kHz sampling rate. Presumably the word length is 16-bits for compatibility with all external DACs. I don’t know if this 16/44 S/PDIF signal is re-clocked in the upsampling filter, or simply bypasses that filter (which seems more likely). Presumably the signal fed to the internal DAC at this sample rate is re-clocked and interpolated to 24 bits.

SACD (Direct Stream Digital) signals, which are essentially Pulse Density Modulated data, can be filtered a number of different ways. Classé’s choice is to strike a balance between measured and sonic performance. Note that a second order passive (analog) filter is about the simplest practical topology to remove the carrier (2.82MHz for DSD data) from this type of signal at a reasonable roll-off. Class D amplifiers, which use Pulse Width Modulated data, do a similar thing, typically employing an inductor and capacitor. But because power handling is not an issue, Classé dispense with one type of reactive component, the inductor. They write:

The filter designed for the SACD 2 is a 2nd-order passive filter. At 12 dB per octave roll-off, it does not offer the out-of-band rejection of a more complex, steep slope filter. However, with a corner frequency at 35 kHz and 96 dB of attenuation at 2 MHz, this filter provides more than sufficient noise suppression to prevent both audible artifacts and intermodulation distortion. In addition, this filter has excellent time domain performance with virtually zero group delay from 100 Hz to 10 MHz – a level of performance that cannot be attained with steep slope filters. Because the filter is passive, composed of high quality capacitors and resistors, it is sonically pure, with no added distortion or noise.

Classé notes that THD analyzers cannot distinguish between harmonic and non-harmonic distortion. Therefore this figure is usually expressed as THD + Noise. In the case of sigma-delta converters noise increases with frequency and moving the noise out of band is a major function of a well-designed circuit. Because noise is a more serious issue with 1-bit (DSD) than multi-bit (PCM) data, the temptation is to design a steep filter to produce an impressively low THD + Noise measurement. But Classé balances one measured specification (THD + Noise) against another (Time Domain Response). They write that “[w]hile other designs offer lower THD + Noise specifications, they cannot match the sonic performance of the Omega SACD 2.” I will have more to say about sonic performance later.

The output circuit is a zero-feedback current-gain stage required to drive the interconnect cables. Zero-feedback has the advantage of preventing external noise, such as RFI picked up by audio cables, from entering the unit. Zero-feedback also makes the use of closely matched, highly accurate and stable components more critical.

A couple of years ago I made some tracking test discs using strips of opaque drafting tape with widths ranging from 0.75mm to 3.0mm. These pieces of tape run straight across the diameter of the disc. A transport’s ability to track these discs gives a crude indication of its ability to track a badly scratched CD. (Commercial test CDs use much more sophisticated tracking tests.) The SACD 2 was able to play the 0.75mm disc with no audible side effects. It failed to read the next size up, 1.5mm. Note that my Accuphase DP90 fails even on the 0.75mm disc; it reads the TOC but will not play any of the bands.

Caveats. Before I go on to talk about the music, in fairness to a prospective purchaser I must note some of the operational ideosyncrasies in the SACD 2. The Time function setting (“track count up,” “track count down,” “disc count down”) is lost when a new disc is inserted, defaulting to “track count up;” whereas Repeat is retained even when a new disc is inserted. This topsy-turvy arrangement is not what I would expect, that the Time function would be retained and Repeat lost when a new disc is inserted. The Pause button must be pressed and released quickly; if it is held down, the machine cycles back and forth between Pause and Play. The display readout shows Track and Index, rather than the more typical Track and Total Tracks. The majority of my CDs show only a single index point per track, though this feature is very handy when multiple index points are present.

The most serious problem is Fast Forward and Fast Reverse. They aren’t. These functions move painfully slowly, at only about two-times normal speed; with regular motion in SACD mode, and rather jumpy and irregular motion in CD mode. (Note that my Accuphase moves at more than 24-times normal speed in Fast Forward mode.) On the other hand, how many times have you used FF or FR in the past twelve months? Me too: maybe two or three times. The root of this problem, and the Time and Repeat anomalies, is the transport mechanism. The Sony 555ES is not simply a piece of expensive hardware; it has several logic boards of its own, with their own proprietary codes. Classé chose the 555ES over similarly priced Philips transports because they found it sonically superior. The trade-offs were the above functional anomalies. The codes generated by the SACD 2’s remote control must work when it is wired together with other Classé products (there is a special port on the rear for this purpose). As I understand it, the circuitry in the remote control could be modified so that FF/FR would progress at typically high speeds. I don’t know how practical this is, but it would be a boon if Classé were to offer, at a nominal charge, such a modification for those who do not intend to use the SACD 2 in conjunction with other Classé products, and who consider the absence of a practical Fast Forward and Reverse a serious design flaw. I do.

[It didn’t work. I ended up wanting to own the SACD 2 regardless.]

SACD anyone? I have reviewed a number of Telarc discs in these pages over the past several months. My evaluations were based solely on the CD layer of these hybrid discs, and Telarc’s recent Gramophone award for sound quality confirms my opinion. One aspect of the SACD vs. CD debate (if, indeed, there is a debate still going on) had to do with source media. How do we know that the engineering skill expended on the SACD version was also expended on the CD version? There is a potential conflict of interest for any company touting SACD as superior to CD. The Telarc hybrid discs, in my opinion, obviate this problem: we have here some of the very finest sound on record, impeccably represented in both formats.

I can categorically state that, played on the Classé Omega SACD 2, the SACD layer of these Telarc discs sound better than the vast majority of CDs I own. Among the few possible exceptions are the CD layers of these same Telarc discs. I don’t say the CD and SACD layers sound exactly alike, but the differences are so finely nuanced that the time required to change modes makes pinning them down very difficult. One engineer recommended months of extended listening, suggesting that over time CD sound would become readily distinguishable from SACD sound by a certain harshness, a certain discomfort felt in the inner ear. Would I choose an SACD over a CD? Absolutely. I know from forays into digital theory that DSD encoding is inherently more accurate that PCM. I like the simplicity of the data, that it’s based on pulse density, and on relative change, not on quantized values. Does DSD recording make skillful engineering and careful manufacture any less important? Silly question.

The ears have it. Fortunately, I am writing not about formats, but about a particular machine. A sound expert, an audio engineer, might be able to explain the factors that constitute the "sound” of the SACD2 in comparison with the “sound” of the Accuphase/Bel Canto. He would know that audio electronics and transducers are, essentially, voiced, just like pipe organs and violins. He would know that certain kinds and amounts of distortion may actually be pleasant to the ear. (To wit, the extremely favorable review of the $350,000 (yes, three hundred and fifty thousand dollar) Wavac SH-833 amplifier in Stereophile. This 550-pound absurdity had dreadful test results, full of distortion at reasonable power levels, ringing on the square waves, etc.) He would know that the notion of absolute accuracy is simplistic, and that an amplifier with magnificent specifications may sound cold, or harsh, or dead. He would know that a successful design is a matter of trade-offs, and that it depends on listening tests as much as on mathematics. Not being an engineer, and having a long and humble history of liking what I like and hearing what I hear (rather than what I’m supposed to like and hear), I can only say that the Classé SACD 2 makes the most enjoyable music I’ve ever heard. Imagine the sound of a fine piano. Then imagine the same piano with a soundboard of the choicest Sitka spruce, with straight, tight, regular grain, carefully dried and lovingly worked: that’s the SACD 2. It’s got a wonderful warmth and richness, a poise and effortless intimacy. I’ve never thought of the Accuphase/Bel Canto combination as being in any way uncomfortable to the ear, but in comparison with the SACD 2, it seems just a bit bright, a bit analytical. I dislike the term “analog-like” because it seems to me worse than meaningless, and yet the magic one hears with a good slab of vinyl played on a superb table and cartridge is undeniable. The same magic is applicable to the sound of the SACD 2.

The music. Clifford Jordan’s Live At Ethell’s (Mapleshade 56292) has a terrific and well-deserved reputation. Played through the SACD 2 all of its special qualities have just a little something extra. I’ve heard this CD on a number of systems besides mine, but I cannot recall such a relaxed yet precise sound. Clifford’s sax is harmonically rich without overly calling attention to itself, and the subtlest nuances of sound seem so natural and right. The quartet really does seem almost present in the room. Spatial relationships are clear and solid, like you could step right up on stage and shake hands with these fine musicians. Clifford’s gone now. The music is here to stay.

I next went to a little known set of Bach harpsichord music played by Ralph Kirkpatrick on a superb Chickering/Dolmetsch instrument from 1908 (Music & Arts CD-976). I know something of the struggles that went into making this set of four CDs, having corresponded briefly with the son of the sound engineer. This instrument used leather rather than quill plectra and employed pedal rather than draw knob stops, but in other respects was close to the harpsichord Bach would have played. The leather imparts a slightly softer voice to the plucked strings than does quill. Talk about an intimate experience! The recording was close-miked (most harpsichord recordings are) and you can see right into the instrument, feel the plectrum attack the string and hear the luscious overtones. One is apt to forget that this is a monaural recording made over fifty years ago, to close one’s eyes and readily imagine the room transformed into a richly fitted, aristocratic drawing room with a fireplace, walnut wainscoting, long, damask curtains, and a few intimate friends in high-backed easy chairs, being treated to a personal recital by Mr Kirkpatrick. (It is unfortunate that Kirkpatrick chose to ignore Bach’s repeats in the Goldberg Variations. I much prefer it with the repeats, both for aesthetic reasons and for its greater playing time.)

Telarc’s recent release of Berlioz’s Requiem (Telarc SACD-60627) represents, in my experience, a type of music that is particularly difficult to reproduce in any adequate fashion. Large scale orchestral and choir recordings tend to lose coherence and definition when things get loud, and the orchestra and choir Berlioz called for are huge. Up to the point where my living room becomes overwhelmed, the delineation of parts is remarkably clear and stable with the SACD 2. It is of course easier to imagine four jazz musicians in your living room than a vast orchestra and choir, but if you close you eyes and hunker down in the sweet spot, the illusion unfolds naturally. It’s quite wonderful. Somewhere I read a negative review of this disc, based on the reviewer’s requirement that a requiem be based on fear, wailing and the gnashing of teeth, whereas (according to him) this performance is too peaceful to be taken seriously. Well, I find this sense of peace (like the requiems of Fauré and Duruflé) totally in character with the music, and in no way detracting from its seriousness. Does he expect Frenchmen to write music like Germans?

I found it quite interesting that the repeatable sonic differences I heard between the SACD 2 and the Accuphase/Bel Canto were present even when I played the SACD layer in the former and (of course) the CD layer in the latter. Both the CD and SACD circuits in the SACD 2 have a similar, characteristic sound. This is no accident. These circuits were carefully voiced to Classé’s ideal, and any trade-offs were very consciously decided upon.

I turned to Diana Krall’s All For You (Impluse! IMPD182). This CD is a luscious sonic treat, close-miking Krall’s throaty, intelligent voice. (The rather uninspired accompaniment may be partly explained by the jacket notes, which tell us that the musicians were in sound booths, isolated from one another in the grand tradition of non-jazz.) Both the Accuphase/Bel Canto and the SACD 2 delivered the goods playing this CD, but their handling of ambience was noticeably different. The Accuphase/Bel Canto combination created a fine illusion of being transported to a club, whereas the SACD 2 rather transported the club to my livingroom. Does that make any sense?

I changed strategy at this point. I removed the digital cable from the Accuphase and plugged it into the SACD 2 digital output jack (RCA). I was now able to rapidly switch back and forth between the internal DAC and the Bel Canto DAC2. I found the sonic signature of the SACD 2 driving the Bel Canto closely matched that of the Accuphase driving the Bel Canto. I did not pursue any differences diligently since what I wanted to establish was simply whether the characteristic sound of the two front ends was attributable to the DAC circuits.

For a given disc, there is no way to say absolutely that one source component is doing a more accurate job than another. One wasn’t present when the recording was made, does not know the characteristics of the instruments or the venue, nor those of the equipment used to make the recording, nor is one privy to the decisions of the sound and recording engineers. That’s why we audition a component using many and varied discs. And auditioning many discs, classical as well as jazz (and acoustic rock), does reveal a subjective truth.

Me and Classé. This is first time I have chosen to review a piece of audio equipment, so I can’t generalize about the attitude of reviewers, but it was my intention from the get go not to covet the SACD 2. (Due, as I’m sure you’d never guess, to financial constraints.) I was going to evaluate a machine, that’s all. I was not going to ponder another upgrade to my system and start figuring the angles by which I might raise thousands of dollars. I tried to convince myself that my Accuphase/Bel Canto sounded just as good, but merely different. (In a sense it does. To its credit, it resolves the last nuance of reverberation in quiet passages slightly better than the SACD 2.) I told myself that I’d have trouble getting used to having to set Time to “track countdown” each time I put in a CD. Nor would I ever get used to finding that two and a half hours have gone by and I’ve been hearing the same CD over and over, because I forgot to turn off Repeat. (This actually happened when I was cooking and not paying close attention). And then there is the almost useless Fast Forward and Fast Reverse. What if I’m reviewing a CD and I need to replay a passage 7:31 into Band 1? How long before I’d lose patience, standing there for over three minutes, running the read head forward (and accelerating toward MTBF because these operations require a brighter laser beam)? I did my damnedest to discourage any desire for this machine.

It didn’t work. I ended up wanting to own the SACD 2 regardless. The sound, the wonderful sound. Despite its functional ideosyncrasies, it is a beautifully crafted, thoughtfully engineered machine with a voice of its own, a voice to which I could listen for years to come.
 

Russell Lichter

                           _________

SACD-2 Specifications
Analog output:
Balanced (XLR): 1.2V
Unbalanced (RCA): 2.2V
Power supply
Transformer size: 55VA custom toroidal x 2
Number of capacitors: 18
Total capacitance: 84,600µF
Digital section
D/A converter: Crystal CS4397 stereo delta-sigma x 3
Up-sampling: 24 bit/44.1/48/88.2/96/192, user selectable
AES/EBU output: 1x
S/PDIF output: 1x
Software support
CD, CDR, CDRW, HDCD, SACD
Weight & measurements
Net weight: 30lbs
Width: 17-3/8”
Height: 4-3/4”
Depth: 15”
Cost: $8,000


Info: Classé Audio, Inc.
5070 François Cusson
Lachine, Québec
H8T 1B3, Canada
Phone: 1-514-636-6384
FAX: 1-514- 636-1428
contact: sales@classeaudio.com
website: www.classeaudio.com