Bel Canto DAC 2
Sugg. Retail: $1300.00 US
Musical Fidelity A3 24
Sugg. Retail: $1199 US
(Reprinted from the Summer/Fall 2002 Audio Ideas Guide )
Since audio reviews generally rely heavily on comparisons, and, more specifically from the cataloguing of differences gleaned from such comparisons, this review might be a little shorter than usual. You see, the products in question here are similar in more ways than any two I’ve reviewed before.
We have here two digital to analog converters aspiring to high-end status, priced almost identically, and at roughly the US $1200 mark, nestled squarely in the moderately affordable zone between multi-thousand dollar “reference” processors and multi-hundred dollar “entry level” boxes.
Closely in step with current digital audio fashions both are up-sampling converters, bumping up incoming data signals to 192 kHz (user selectable 192 or 96 kHz in the case of the Musical Fidelity) and increasing the word length size to 24 bit before turning the samples into analog waveforms. Both accept digital data at 96 kHz or less from either coaxial or toslink optical inputs and output analog audio through gold plated RCA outputs only. The Musical Fidelity, however, also provides a digital output allowing you to feed another device or system if desired, the signal from which is fully re-clocked in the process.
Both DACs employ multi-bit, Delta-Sigma converters (The Bel Canto’s is from Burr-Brown) and, since the higher sampling frequencies push the nasty digital harmonics created by the conversion process much further out of the audible frequency spectrum, both machines take advantage of much more gradual roll-off filters rather than the “brick wall” filters in 44.1 kHz DACs (44.1 kHz DACs clamp down hard on any signal above 20 kHz whereas 96 and 192 kHz devices can roll of the top end much more gently at closer to 96 kHz, helping to preserve analog like transient response). This, many believe, is the main reason behind the perceived increase in resolution gained by up-sampling, a process which increases the clock speed but cannot, without complicated interpolation technology, actually increase the temporal resolution of a signal that was originally recorded at a lower sample rate. For a more thorough discussion of up-sampling and interpolation see my review of the Perpetual Technologies P-1A and P-3A digital converters.
The Same, but Different
For all their commercial and technological similarities no one is likely to confuse these two boxes based on their physical appearance. Housed in the same chassis as the rest of the A3 line the Musical Fidelity is big for a DAC, and, like my A3 CR preamp and power amp surprisingly heavy. It’s also very pretty, assuming you like shiny things with nifty looking blue LEDs. The full-size aluminum chassis and beefy, silver faceplate are certainly contributors to the A3 24’s 10 Kg mass, but the bulk of the heft comes from Musical Fidelity’s signature choke regulated power supply. These chokes, basically brick wall filters for incoming AC power, can be found in most MF components and are often credited for the lively and transparent sound of Musical Fidelity’s recent gear. With lots of empty space inside the relatively small DAC circuit board can be kept well away from the power supply.
As with most DACs operation is dead simple. Plug it in via detachable power cord, connect a digital source (or sources) to one of two input jacks, push the power button, choose 96 kHz or 192 kHz upsampling via the rear mounted button and watch the beguiling blue indicator lights telling you you’re powered up and locked on to the source. With the unit drawing only ten watts and running cool, I had no compunction about leaving it on all the time, which is exactly what I did.
Like Perpetual Technologies, and Audio Alchemy before them, Bel Canto has opted for the smaller is better approach. The brick-shaped DAC 2 is functionally minimalist, designed to keep signal path lengths to a minimum. Input and output connectors are mounted directly to the narrow internal circuit board, coax and optical inputs at one end and RCA outputs at the other. With cables sprouting from both ends, and its non-descript black chassis, the DAC 2 is less elegant than the A3 24. Points lost for style, however, are picked up for efficiency. Being a DAC it doesn’t demand much user interaction and with its compact size, can be tucked away behind or beside other components, ideally positioned very close to your CD transport to minimize cable lengths. Similarly simple in operation the DAC 2 has only one control, a switch selecting either the optical or coax input. A light on the output side of the brick changes from red to green when the DAC 2 locks onto an incoming signal. Like the Musical Fidelity it draws only 10 watts and, with no power switch, is on whenever it’s plugged in. Neither machine ever acted up, failed to lock up to an incoming signal or otherwise behaved aberrantly in any way.
When the Musical Fidelity processor arrived I still had the Perpetual Technologies digital to digital processor (P1-A) and DAC (P3-A) on hand for a brief period overlap so sonic comparisons were certainly in order. With both processors fed by the P1-A (which up-samples to 96 kHz, mounts multi-pronged attacks on jitter, and actually interpolates data that would have been recorded if the source were 24/96) the two DACs sounded truly excellent, and definitely revealed subtle, but distinctive characteristics. Slightly darker tonally, the Musical Fidelity also sounded a touch more refined than the less expensive P3-A. At the same time the P3 had a very slight edge in overall detail, owing in part to a slightly brighter, drier sound. Both DACs unequivocally blew away the internal DAC in my Rotel RCD-951 CD player but I came away with the feeling that the MF was the smoother operator, it’s suave, more analog-like sound making it more natural sounding and less fatiguing over the long haul. This impression only increased when the P3-A was stripped of the advantage of the US $349 Monolithic Sound power supply designed to drive both it and the P1-A. Running on its internal power supply the P3-A sounded decidedly coarser and a touch aggressive by comparison with bass now a little underdamped.
When the A3 24, this time without the aid of the P1-A, was pit against the Perpetual full court press (P1-A, P3-A and the Monolithic power supply, a combination costing over US $2000), the advantage shifted to Perpetual. With the P3-A synergistically matched to the P1-A, the MF lost some smoothness, air, and musicality by comparison, the P3 benefiting considerably more from the P1 than the A3 24. As you can already infer, however, the best sounding combination was the P1-A (powered by the Monolithic power supply of course) feeding the A3 24. This chain was the smoothest, most musical, most “un-digital” sounding digital I’ve ever enjoyed in my system. It also helped reveal something which set the Musical Fidelity processor apart from so many others: long term freedom from listener fatigue. I’m not sure it’s as quite at the level of SACD or DVD-A, as MF’s ads claim, but digital gear at this level can certainly inspire joyful excavations of long-unplayed CDs.
Since I was getting such sweet sound from the Perpetual/MF combo I was sorry to see the P1-A go back to the manufacturer. With the A3 24 now left to its own devices a little of the magic was missing, but not enough to keep me from enjoying it thoroughly. No flash in the pan, the A3 24 seemed to get better the more I listened to it; a combination of vanishingly low listener fatigue and a slow break in period perhaps. This is not something than can be said of most other digital gear I’ve auditioned.
A late joiner to the party the Bel Canto Dac 2 showed up a few weeks after the Perpetual pieces had left. After a week or more of break in I substituted it for the A3 24 and, over another couple of weeks formed some general impressions without doing any direct comparisons. Luckily recent trips to Hollywood’s Amoeba Music (easily the best record store I’ve ever patronized, and this in a town with more great record stores than any city I can think of) had furnished me with ample fodder for the formation of said impressions, namely in the form of some excellent sounding CDs. Cornershop’s Handcream for Generation, Wilco’s Yankee Hotel Foxtrot, and Neil Young’s Are You Passionate? for example are not only great discs which I urge you to add to your collection, they also sound fantastic and particularly un-digital, even on CD. And, just to give you an idea of how deep a deep catalog store Amoeba is, I found two excellent CDs by Toronto acoustic guitar phenom Don Ross which I had never even heard of before (Passion Session and Huron Street are the titles). These, from a label called Narada, are minimally miked, audiophile grade ammunition and sound correspondingly excellent. Since they were used I got em’ for 7 bucks a pop to boot.
Listening to these discs, and many others through the DAC 2, it was pretty clear that Bel Canto had no intention of being embarrassed by the likes of Perpetual Technology or Musical Fidelity. My initial impression was of velvety, analog-style sound with excellent soundstaging capable of putting instruments well outside the speakers. I was particularly enthralled with the Wilco and Neil Young records, the latter having a “live off the floor” quality that puts Neil and his band (in this case Booker T’s MG’s have been substituted for Crazy Horse) in the room to an almost eerie degree. Transparency? Most definitely. When that kind of thing starts happening, and you just keep pulling out disc after disc, it can usually be chalked up to the musicality of the system (not to suggest that the quality of the music doesn’t play a huge part). When you can close your eyes and have the room melt away it’s about as good at “virtual reality” gets.
So, yes, as these listening sessions proved, the DAC 2 was impressive right off the bat. For all its strengths, however, it didn’t leave me feeling that it had necessarily outclassed the Musical Fidelity. And here, my friends, is where we return to the theme of similitude. Once I started making direct comparisons (via remote control on the MF A3 CR preamp) it became clear just how similar these two DACs sounded. When comparing DACs this way it generally doesn’t take terribly long to establish a feel for the character of each machine, especially when long passages are played without switching. In this case it was frustratingly difficult to even hear the switch! With my Rotel as transport, or with my Panasonic DVD A-310 DVD player, via optical or toslink, when compared back to back from the same transport the two DACs sounded almost identical (I used two identical runs of Ultralink Ultima interconnect cable to the preamp and the stock power cords, by the way). Even when I thought I had some consistent sonic difference to latch onto, repeated back and forth comparisons would rarely reveal something repeatable. Never before have I encountered two components so sonically deadlocked.
Even more concentrated, intent listening, with the most challenging source material in my collection did, over time, start to reveal some very subtle, but identifiable differences between the two DACs. On Air’s 10,000 Hz Legend, for instance, the DAC 2 seemed just a hair more open with a slightly wider soundstage and somewhat crisper transient edges for a vaguely more three dimensional effect. Tonally, in terms of overall resolution, imaging and bass control, however, I couldn’t hear any consistent differences. They both sounded great. More or less the same, but great.
On the Kundun Soundtrack and Patricia Barber’s Cafe Blue, as good a test of digital resolving power as I’ve come across, the sound was sumptuous on both processors, with maybe an extra whiff of air and a centimeter of stage width from the Bel Canto. At the same time I felt that the A3 24 delivered a slightly smoother sound, the top end and mids ever so slightly lusher. Both featured fast, engagingly tactile microdynamics and extraordinary resolution with extended, firm, and well damped bottom end. They may lack the sophistication of digital from the likes of dCS and Levinson, but there were few faults to find in this pair of processors.