Meridian 203 D/A processor Robert Harley page 2
If Bitstream offers a sonic improvement in low-cost CD players, I'm all for it. However, my real interest lies in which technology will provide the ultimate digital playback. If Bitstream doesn't advance the state of the art in digital decoding, it would be a mistake for high-end manufacturers to embrace it out of marketing pressure or ease of implementation. Unless it can compete sonically with the very best conventional converters, Bitstream will be a step backward for the high end.
It's a common mistake, however, to focus on a particular technology rather than the implementation of that technology. A digital converter's musicality is affected by a host of factors, not just the D/A conversion method. The way in which the incoming digital data stream is handled, grounding, power supplies, analog output, analog filtering, and pcb layout are just a few of the factors that affect a processor's sonics. It is wise, therefore, to consider the entire design when evaluating a new technology like Bitstream. Both my previous experiences with so-called "1-Bit" converters have left me unimpressed. (See my reviews of the Harman/Kardon HK7500 CD player in Vol.13 No.4 and the D/A section of the Sansui AU-X911DG integrated amplifier in Vol.13 No.3.)
Getting back to the 203, one of its most interesting features is a jitter-reduction circuit that processes the incoming data stream before decoding. This proprietary circuitry is hidden beneath a metal shield toward the rear of the pcb. It reportedly reduces jitter by a factor of 100, which in turn improves the audio performance. Rather than use the ubiquitous Yamaha or Sony S/PDIF decoder chip, the 203's incoming-data processing circuit uses the Philips 7274, which I haven't seen before in an outboard digital processor. Meridian feels that the 7274 decoder provides a more stable phase-lock loop (PLL) than the common Yamaha and Sony S/PDIF decoders. This PLL must lock to the incoming data stream, introducing as little jitter as possible in the recovered clock. The loop filter and oscillator are shielded under a can, and each have their own isolated DC supply and ground planes in the pcb.
Bob Stuart has done some interesting research on the effects of jitter, specifically as related to the effects of PLL parameters. He has found both objective measurements and subjective correlations between sound quality—particularly low-level detail—and jitter introduced by loop errors. The 203's attention to this problem is extraordinary, especially in a $990 machine. I think more and more designers will address these issues as the general awareness of jitter increases (footnote 2). Incidentally, this circuit has the ability to lock to any sampling frequency between 32kHz and 48kHz, rather than being limited to the typical three discrete sampling frequencies (32kHz, 44.1kHz, 48kHz). This makes the 203 ideal for use with CD players with variable speed operation.
Overall, the 203 is a simple, cost-effective, no-frills machine that incorporates some interesting design innovation. I am particularly intrigued by the incoming-data processing circuit that reduces jitter in the recovered clock. In addition, the dual-differential Bitstream approach appears to have a sound theoretical basis.
I auditioned the Meridian 203 over a three-week period in my reference system. I listened to it on its own and compared it with the similarly priced Proceed PDP ($1295) reviewed in Vol.13 No.6, the exceptional Theta DSPro Basic ($2000) reviewed in Vol.13 No.8, and the Melior Digital Control Center ($2250) reviewed elsewhere in this issue.
The first time I heard the 203, I was immediately impressed by its performance. I've auditioned and reviewed a wide assortment of digital processors lately and have developed a pretty good feel for what level of performance one gets at the various price levels. The 203 far exceeded what I've come to expect from a $1k converter.
The 203's presentation was smooth and fluid, with natural rendering of instrumental timbres. There was a certain liquidity and roundness to instrumental textures that I found particularly musical. Often, processors in this price range tend to be a bit sterile or cardboardy, failing to convey an instrument's texture. Not so with the 203. It had a remarkable absence of glare and hardness through the midrange that provided a sense of smoothness, inviting the listener into the performance. Voice and other instruments with mostly midrange content were round and silky. The flugelhorn and sax on a CD I had engineered were palpable and liquid. The overall presentation tended to be forward and immediate rather than laid-back.
I found, however, that the treble had a measure of hardness and a slight tendency toward aggressiveness. Cymbals lost a little of their delicacy and shimmer. Instruments rich in high-frequency harmonics, like the violin, took on a slight edge. Compared to comparably priced digital processors, this character was quite minor and didn't significantly affect the musical experience. I wouldn't call the treble bright. Rather, I would characterize the 203 as having a lively, open, and airy quality through the treble.
The soundstage was particularly impressive for an affordable decoder—transparent and uncongested, and accompanied by a distinct sense of three-dimensional layering. Instruments behind one another were easy to distinguish from their neighbors. Each instrument retained its individuality in the presentation during complex passages, without blurring or homogeneity. In addition, image focus and specificity were similarly good. Try the remarkable sense of space and depth on Three-Way Mirror (Reference Recordings RR-24CD), which I find quite revealing of a component's ability resolve spatial nuance. The 203 presented tight, precise instrumental outlines, perhaps contributing to the ability to hear individual parts in a performance. The soundstage was very wide, giving the presentation an open, airy feeling. The impression of space around instruments was particularly apparent on Handel's Water Music (Harmonia Mundi HMU 907010). In addition, instruments were well-placed laterally, with a solid center image. Overall, the 203's soundstage competes with processors costing twice as much.
Bass performance was excellent by any measure. The 203 had a good sense of weight and heft, without being bloated or tubby. The extreme bottom end was tight and punchy, while the midbass was full, round, and rich. The kick-drum on James Newton Howard and Friends (Sheffield CD-23) was dynamic and deep, with a feeling of effortlessness. Pitch definition was superb, with the ability to follow rapidly flowing bass-lines (Stanley Clarke on Light as a Feather, Polydor 827 148-2). Throughout the auditioning, the 203 infused music with rhythmic urgency and drive which, coupled with the lively treble, conveyed the life and energy of the music.
My primary criticism of the 203 is a tendency to obscure inner detail. Finely woven textures just didn't come through, instead becoming slightly blurred. Percussion instruments toward the rear of the soundstage lacked the bite and fine detail heard through other processors. Similarly, subtle instrumental textures seemed lost in the instrument's body. Finger noises on guitar strings, for example, tended to be muted and less clear, reducing the sense of realism from the reproduced instrument.
I next compared the 203 with the Theta DSPro Basic, Melior Digital Control Center, and the Proceed PDP. First, the 203 was no match for the Theta. The 203 had many of the Basic's qualities, but not to the same degree. The Basic's soundstage was deeper and more transparent, its presentation highly detailed by comparison. They did share, however, a sense of liquidity, bass definition, and openness. Nevertheless, the 203 was not embarrassed next to the Class A Theta.
I found the 203 closer in overall performance to the Melior Digital Control Center, but different sonically. The MDCC was "undigital"-sounding by comparison, with smoother treble and softer textures. In addition, the MDCC was more detailed, with better resolution of subtle instrumental nuance. The MDCC had a sense of ease and relaxation rather than of immediacy. Both had similar soundstage presentations, but with the 203 sounding more up-front and vivid. Consequently, the 203 tended to have a greater sense of openness and soundstage width, but slightly less depth. The 203 had a better presentation of dynamics, especially in the low end. There was more substance and weight in the lower registers through the 203, but not at the expense of sounding fat.
Next to the Proceed PDP, I preferred the 203's soundstage presentation, smoother treble, and bass articulation. However, the PDP was better at revealing detail in an instrument and also in portraying the transient attack of percussion instruments. The PDP's bass presentation tended to be warmer, while the 203's was tighter and faster. What really tipped the scales in favor of the 203 was the smoother treble and more liquid midrange textures, making it sound less digital.
The Meridian 203 sets a new standard of performance in under-$1k digital converters, offering a level of musicality previously unavailable at anywhere near the price. Specifically, the 203 had many attributes of much more expensive digital processors. These include a wide and deep soundstage that created a convincing illusion of size, and instrumental textures that were liquid and smooth, without glare. In addition, the 203's senses of weight and effortlessness in the lower registers were excellent by any measure. This all adds up to a lot of processor for the money. My two criticisms are minor and should be taken in the context of the 203's reasonable price. These are a tendency to obscure fine detail and a slightly forward, hard treble. However, in relation to the 203's many strengths and the performance offered by similarly priced digital converters, such faults are acceptable.
I am encouraged by the fact that this level of musicality was achieved with Bitstream conversion. Like it or not, designers are embracing Bitstream and we are likely to find these chips in many more digital products in the coming months and years. (See my CES report on new digital products in Vol.13 No.9.) This doesn't mean all Bitstream converters will sound good; but with the right implementation, Bitstream has the potential of musical performance.
Another aspect of the 203 I found interesting and impressive was the economy of design. Careful attention was paid to aspects of the circuit that affect the processor's sonics (like the incoming-data processing circuit), while money wasn't wasted on frills or tweaky details whose sonic contributions to the unit's overall performance are questionable or marginal.
The Meridian 203 is, in my opinion, the best value available today in $1000 digital processors. All things considered, the 203 ranks in Class B of Stereophile's "Recommended Components," outdistancing similarly priced units in Class C, while falling short of Class A performance in the areas noted. If you can live without multiple digital inputs and the last measure of detail, the Meridian 203 will provide eminently musical digital playback without requiring a major investment.—Robert Harley
Footnote 2: See my review of the Stax DAC-X1t (Vol.13 No.8) for a brief discussion of how R-2R ladder converters work. For an excellent explanation of Bitstream decoding, see Peter Mitchell's "Industry Update" in Vol.13 No.1.—Robert Harley