In fact in another context, phase is used to modulate a carrier frequency in order to convey information, just as are amplitude and frequency in AM and FM radio. Only, in the case of audio signals, the “information” encoded in the phase variations of the carrier is undesirable, it is noise, whereas the “carrier” itself is the desired audio signal. And if I understand the diagrams correctly, it is a type of distortion that applies in the analog as well as the digital realm, altering the frequencies of audio signals. Very slightly to be sure, just as the products of IMD are very small, but I presume audibly.

I admit to a certain curiosity to know which area – digital or analog – is most effected by noise modulation, is most improved by the VBS1, but I'm not sure anyone can say. However I would like to know what the delivery of battery-like power to the high current end of a stereo, the power amplifiers, would do to the sound. Such an experiment would at least be suggestive as to what areas of a DAC are most effected by noise modulation.

A few paragraphs above I proposed the question, Does the power cord supplying the VBS1 make a difference? The answer to this question is an indicator of how well the Virtual Battery Supply works, how well it isolates the DAC3VB (in this case) from line-borne noise. I was excited but not altogether surprised to discover this very question addressed in the VBS1 white paper:

“We have confirmed this isolation by performing listening tests with the VBS1 where you literally UNPLUG the VBS1 from the wall and listen for any change in the audio performance – there is no change – nor is there much, if any, sensitivity to the specific type of power cord used to drive the VBS1...”

As Mr Stronczer wrote in an email, “I am just using the stock cord — tried my uber-expensive after-market cables and could not reliably hear a difference.” I too tried swapping in the OEM cord and had essentially the same experience. Make no mistake about it, if a power supply works well enough to render proprietary power cords virtually indistinguishable from mass produced OEM cords, it's not just another cool design, it's revolutionary. At least from an audiophile's point of view.

The VBS design has been through various configurations. At one time it was built around a linear power supply preceded by extensive line filtration and followed by low-frequency filters to eliminate 50/60Hz artifacts. But it turns out that higher frequency noise is easier to filter effectively than lower frequencies, so Bel Canto experimented with a switch-mode power supply (inherently highly regulated, unlike typical linear designs), the main disadvantage of which is the production of a lot of high frequency noise. The final VBS1 has common mode LC filtration at the AC input, followed by low-noise, high-speed bridge rectification. Solid state rectifiers, which are used in virtually every piece of AC-powered electronic equipment on the planet, in addition to conducting in a single direction, store and release energy that dissipates as a burst of high-frequency noise cycling at twice line frequency (100/120Hz). The soft-switching diodes used in the VBS1 produce relatively low noise and produce it at higher frequencies, making it easier to filter the output. The bridge rectifier is followed by electrolytic capacitors (reminiscent of the MKII upgrade to the REF amplifiers), then an over-rated buck regulated SMPS. This in turn is followed by a “massive bank of LC filtration” – compound inductors, paralleled electrolytic capacitors and polypropylene capacitors. The electrolytics used at this stage of the VBS1 have a good safety margin for a 12V circuit. They are wired in parallel for a filtering/storage capacitance of nearly 500,000uF, a rather staggering value that can deliver large bursts of current just like a storage battery. (A typical low voltage linear power supply has an output filter capacitance in the range of several thousand micro-farads.) The final polypropylene capacitors filter any residual high frequency noise and increase instantaneous current delivery. Noise filtration in the VBS1 begins at a few hertz and is 100db down at 100Hz. With a 20 watt load (the DAC3VB idles at 7 watts) the noise level measures around 6uV over a 30kHz bandwidth, a level of quietness achieved by only a few, very expensive analog preamplifiers.

“Spectrum analysis of this design from the low audio frequencies to well beyond the audio frequency band revealed a noise floor within a few decibels of the spectrum analyzer noise floor. This represented a 10X to 100X reduction of the noise floor of the traditional linear approach. This reduction in noise was particularly evident in the lower region of the audio frequency band, the critical midrange where the ear is most sensitive. Indeed in the low frequencies the VBS1 provides an even lower noise floor than a lead-acid battery!”

The 12 volt output of the VBS1 can directly power the Bel Canto CD2 player, but the DAC3 requires 16 volts. This is accomplished with a proprietary power board that replaces the original board in the DAC3. This circuit is a sophisticated boost converter SMPS designed by John Stronczer that employs a custom-made transformer supplying the plus and minus rails of the output as well as providing a high level of isolation. This transformer, plus a low-noise SMPS designed to control slew rate in order to limit HF noise, followed by further HF filtering, result in a noise level of better than 10uV over a wide bandwidth.

The Bel Canto DAC3 has received a lot of praise in various quarters, including Stereotimes, making it to several “best of the year” lists. And now the introduction of the VBS1 power supply elevates it's performance to a remarkable degree. For a reviewer such as myself having what can be described as a reference-level digital to analog converter is a privilege and a delight.

THE SOUND.

Initially at least, in auditioning the VBS1, I found myself unwilling to pay very much attention to the usual objets de culte like imaging, transient response, sound stage, detail, dynamics (which in any case were good prior to the DAC upgrade). Rather, I was continually captivated by the sheer beauty and realistic presence of the music. Which is precisely what happens, certainly what ought to happen, when listening to live music: it should captivate. I've been to enough performances to understand that experiencing live music is not about the ruminations of audiophiles or whether I am able to hear the oboist turning a page of music. No, it's about the ecstatic experience of music itself. And that is precisely what is enhanced by the VBS1, a music so true, so neutral, so unencumbered, that other considerations are not merely elusive, they seem unimportant. There's an easiness, a relaxed presence, a breath-taking purity about the sound (all highly subjective terms, all open to numerous interpretations). A sort of “airiness,” an open, lucid space between me and the music. It goes far beyond the limited dimensionality of the “image behind the loudspeakers” to something deeper in space, something more palpable and more clearly defined, the venue as well as the timbre. Speaking as a poet, the experience feels like emerging from an enclosed space into a lush meadow; from experiencing sunlight and springtime through a window, to actually being there. Can it get better than this? No doubt, no doubt. But, given a well-recorded and engineered CD, I don't believe I've ever heard any stereo, anywhere, that's impressed me more.

In a sense, improving the quality of a stereo system in an acoustically compromised room is a matter of compensation. I remember a visiting neighbor, a high-end audio dealer, saying (some years ago now) that my system did a particularly good job on solo piano music. This may have been a sort of back-handed compliment, for certainly I was aware that one of the great weaknesses of my stereo was the reproduction of large scale music, especially when a large choir was involved. There is simply no bypassing room acoustics. Getting that right is primary, yet I suspect it is the single most neglected aspect of setting up a successful stereo. Unfortunately, sometimes it cannot be gotten right. If, for example, in an acoustically critical corner of the room there is a partial brick wall surrounding a wood burning stove so that the use of sound absorbent material is not possible; and there is no practical loudspeaker position that gets around the problem. In such a case compensation is the only route, and it is the route that I have been following for years, as my home-made, burlap-wrapped fiberglass baffles and heavy window curtains testify.

It dawned on me, rather surprisingly, that in the weeks since the VBS1 arrived I have been listening mainly to several versions of Faure's Requiem, precisely the kind of music that has been problematical for this stereo. And at least three of those versions are quite large scale, the dubious – but I think magnificent – 1901 orchestration with choir, soloists, organ, strings and horns. While neither power cords nor electronics can fully compensate for an imperfect acoustic space, the VBS1 and DAC3VB have moved this stereo a giant step in the right direction. There is so little noise, so much pure detail, I am able to see so much farther into the texture of the music, that even listening to the Requiem is a relaxed experience, free of audiophilic compulsions. When Fischer-Dieskau sings the Libera me, when De Los Angeles sings the Pie Jesu (EMI Classics 7243 5 66946 2 0), their voices are breathtakingly real. De Los Angeles's singing here is simply one of the most beautiful sounds I have heard in my life. Visitors who have been fortunate enough to walk in on one of the several daily performances have all commented on the extraordinary realism and beauty of the music. One guy, who has listened to this stereo every couple of weeks over the past few years commented that every time I make a change he thinks to himself, “This is it, it doesn't get any better,” only to find that with the next change it does. But I don't think I'd seen him so astonished, so lost for words before.

JS Bach, Goldberg Variations, Martin Schmeding, pipe organ (Cybele SACD 030.802). The Goldberg Variations is one of the few compositions published in Bach's lifetime, and one of the even fewer for which he specified the instrument, a harpsichord with two manuals. Ever since Wanda Landowska restored Bach's keyboard music to our attention with her trusty Pleyel, a harpsichord-versus-pianoforte debate has raged. Actually, I'm not sure it's raging any longer. Despite indignant purists (“How dare they ignore Bach's clear instructions?”), the piano has won out. There are hundreds of piano versions, compared to a relative handful on harpsichord. But pipe organ? How utterly unsuitable for this intimate music! A catastrophe, or worse, a boring performance. However, take an exquisite instrument like the Dresden cathedral Silbermann organ (completed 1750, an instrument Bach actually played), add an organist of extraordinary taste and sensitivity, and you discover yet another way this resilient music can be beautifully performed. (In 1944 the pipes, the heart of the instrument, were removed to a monastery for safety, shortly before the fire bombing of Dresden.) So impeccable are Martin Schmeding's tempi and choice of registration that you'd think the Goldbergs were written for that instrument and written in a spirit of delight and playfulness. I am particularly pleased that he plays all Bach's repeats, a “purity” of performance so often ignored by purists on piano and harpsichord alike. The level of tonal and ambient detail on this hybrid disc is as good as I've heard: and with the VBS1 the realism is quite wonderful. I find there is notably more ambient detail following the conversion of the DAC3 for the VBS1.

Beethoven, Waldstein sonata, David Allen Wehr, piano (Connoisseur CD 4263). I am particularly fond of this performance. For one thing it's well done, and I find Wehr's approach refreshing, he's shaken off some of the inertia of “tradition” and found his own voice. For another I have long-standing respect for the Connoisseur label and for owner Alan Silver, who's published some marvelous performances that otherwise might not have seen the light of day. And finally because this recording was made in Utica, a town in upstate New York where I lived for years, ate numerous antipasti and pizzas, picked wild strawberries, and saw my first Jackson Pollock. The venue, First Presbyterian Church on Genesee Street, is very live, lots of reverberation. In the past I found this objectionable, even questioned the wisdom of the sound engineer, but with the VBS1 and by honing-in the volume level, the realism is impressive. Much like a live performance, yet capturing the nuanced voice of the piano, a Yamaha CF111S, beautifully and apparently without close miking.

Beethoven, String Quartet in C-sharp minor, Petersen SQ (Capriccio 10 510). Older quartet musicians are prone to comments like, “When it comes to performing Beethoven's late quartets, life experience is all important.” I agree; but like many truisms, this one is simply not always (or consistently) true. Take the Petersen String Quartet, relative youngsters who formed the quartet in 1979 while still students, posing for the album cover (in 2000) in casual clothing and longish hair. Add to this that the C-sharp minor quartet is generally regarded as the most perfect, most transcendental composition Beethoven wrote (and the always-astonishing fact that he was nearly stone deaf when he wrote it). Yet I find this a wholly wonderful performance, and its emotional and spiritual range doesn't seem limited by the relative youth of the musicians. One returns to quotidian life as from a vast spiritual voyage.

Inevitably I've begun to get used to this level of sound quality, though every so often I am still astonished. I suspect that, if there were a way to conveniently step back to the pre-VBS1 DAC3, I'd be pretty shocked at the change. It seems there's really a lot more music on a CD than we're generally able or used to retrieving.

12/04/2009

Addendum: I remember an audio newsgroup thread discussing the possibility of editing out Glenn Gould's infamous mumblings on his studio recordings. (I don't know if anybody ever spoke to Gould about this, asked if he could please not vocalize while he was playing, particularly as his mumbling seemed to have no melodic relationship to the music. But mumble to himself he did, right up until the end.) I was on the other side of the fence: GG's extra-musical accompaniment is an integral and endearing part of his performance and I argued that it should by no means be expurgated. Listening just now to his second studio recording of the Goldberg Variations (CBS-Masterworks MK37779), the only piece of music he recorded more than once, his mumbling was very audible, even more audible than before the VBS1. Which lead me to this addendum, because Gould's sub rosa vocals are indicative of a stereo's ability to delineate and clarify, and there are a few more things I'd like to add, or reiterate, in that regard.

Even if I were clever enough and articulate enough to enumerate all the improvements since installing the VBS1, the whole remains more than the sum of its parts. And it's the whole, the aesthetic experience itself, that's so tough to describe. I finished this review and sent it to Clement a few days ago, now I could relax and just listen without having to write anything. But as I've played additional CDs from my collection, I've been astonished (yes, that's the word) at how real the music seems sometimes. Recordings I'm totally familiar with stop me in my tracks. You might call the Bel Canto VBS1/DAC3VB a sort of truth machine. You get more of what's really on the CD. In most instances this is an improvement; not only nuances, but distinct passages in the music that were obscured before, buried in the subtle distortion, spring to life and are obvious. So that I've said a number of times, “It's as if I've never really heard this before.” Not just improvements to the sound quality, but an overall improvement to the aesthetic experience itself. Performances that were very good become radiant, suddenly the language of the music is revealed with greater detail and aesthetic logic. But there's a catch. Some performances that I thought perfectly good have become less coherent, they make less sense from an aesthetic point of view. As well, some CDs sound worse, because there are fundamental problems with the way it was recorded and/or produced and there's no longer a subtle haze of noise obscuring its contours, sort of like eating a Macdonald's Big'n'Tasty without the catchup, mayo, pickles and artificial flavoring (as manufactured in a chemical plant in the central Jersey swampland). All of this is precisely what I look for in audio equipment. With the Bel Canto Virtual Battery Supply powering their DAC, the idea of accuracy has taken on a new meaning. A truth machine indeed.

Price:

DAC3VB $2695
Virtual Battery Supply $1495
Upgrade board for DAC3 $595

Bel Canto Design, Ltd.
221 North 1st Street
Minneapolis MN 55401
USA

Tel: 612-317-4550 (9AM to 5 PM CST M-F)
Toll-free (866) 200-7342
Fax: 612-359-9358
www.belcantodesign.com