I differ from some of my colleagues, though, in believing that straight-up A/B comparisons can be useful. All but the most subtle changes in the character of sound, I've found, can be discerned in level-matched A/B switching—and this approach can expose common reviewing errors, like mistaking a small difference in volume for a difference in quality. Playing the same passage over and over, switching between (or among) different source components, is an effective way to hear even subtle differences in spatial performance, reverberation tails, and other manifestations of low-level resolution.
I started my listening to the Bartók by running the output of my analog front-end through a Mytek Brooklyn ADC I have on hand. I then ran the 24/192 output from the ADC into two different DACs: the Bartók and the Mytek Brooklyn DAC+ (footnote 4). I connected all three analog outputs—direct from the phono preamp and from each of the two DACs—to inputs on my preamp. Both DACs have volume controls, so I used them to match the DACs' volume to the volume of the vinyl source, monitoring levels with my iPhone's DecibelXPro app.
Why such a complicated test? I wanted to start with a true analog source—I only played records that had been recorded and mastered in the analog domain—and compare it directly to music that had been digitally processed. This, then, was a three-way comparison: I could compare the two DACs to each other and each of them with the original analog. Would I be able to detect a unique digital signature that had imposed itself on the music? Would it be worse with one DAC than with the other? Pressing buttons at random on the preamp's remote control, I soon lost track of which source I was listening to at a given moment.
Fun idea, boring outcome. I listened to tracks from several albums—but could hear no differences among the three inputs. It was easy to tell why: There was too much noise in the source. Whether it's because of the quality of my vinyl front-end, the condition of records I played, or the intrinsic limitations of vinyl, anything added or subtracted from the sound by the ADC and DACs was swamped by noise and distortion from the analog front-end. I would not call this test successful, but I did learn from it.
It convinced me that 24/192, with a high-quality ADC and DAC, is sufficient to capture the sound of analog—anyway, my analog. PCM proved a transparent container, preserving the sonic character of vinyl in my system while adding or subtracting nothing audible.
I moved on to long-term listening, now with digital sources, using my Roon Optimized Core Kit (ROCK), which I built on an Intel NUC, directing data to the Bartók via the network connection, from Qobuz and Tidal and files collected on my Synology NAS.
The Bartók, in the context of my current system, consistently and unambiguously revealed the character of the recordings it played, with clarity, pinpoint imaging (even at low volumes), excellent image depth, fully saturated tonal colors, and no noticeable emphasis on any part of the frequency spectrum. This is the highest praise. Consistent with my expectations, there was no discernible "digital" sound.
On "Top Hat," from Andy McKee's album Sound Roots (Mapleshade 4432), Alex Foster's tenor and Ryan Kisor's trumpet were uncannily tangible and textured, surrounded by a convincing sonic space (not my space but the space in the studio, or a good approximation). The ride cymbal struck a great (I think natural) balance between sweetness and that little bit of harshness I typically hear in real cymbals, live.
I listened to a lot of recordings, but my listening notes are boring—essentially repetitions of what I wrote above: Recordings sounded like themselves. Mostly these were good recordings, and they sounded very good. With recordings that have that AM radio character, my system sounded like an AM radio. I was able to listen with pleasure to the particular sound of each recording.
The pursuit of transparency
There's something to be said for those DACs that have been influenced, directly or indirectly, by the needs of pro audio. Pro audio needs DACs that approach transparency, to the point at which several passes through the DAC (round-tripped via a companion ADC) change the sound negligibly. That puts consumers at a huge advantage, since the music only passes through our DACs once—any alteration in the sound will, therefore, be a fraction of what it would be in a studio.
DACs that achieve a high degree of transparency must sound very much alike: If A and B both equal C, then A must equal B. Therefore, DACs that do have their own sound—those that stand out from the crowd—must deviate from transparency. This isn't from listening tests or technical theory; it's the most basic logic.
Toward the end of my time auditioning the Bartók, I went back to the Mytek Brooklyn (in fact I now used the Brooklyn Bridge, to take advantage of its Ethernet input). I set up two zones in Roon, then grouped them. I matched volumes. Then I listened to a few tracks, switching between them rapidly—now with digital sources, delivered via Ethernet from my collection of files plus tracks from Tidal and Qobuz—and also listening to the same track for several minutes, via one DAC and then the other.
It was not hard to hear a difference between the Mytek and dCS DACs, although that difference was subtle enough that I doubt I would have noticed it if I had not had the ability to switch rapidly back and forth, although the effects of the difference could still have made themselves known over time.
The main difference was this: In the middle of the soundstage, where it's usually deepest, there was a slight smoky haze with the Mytek. This had two subjective consequences. The illusion of space around the instruments (usually piano) was better with the dCS. Especially when a double bass was centered and not a piano, the bass's higher notes were woodier; this same difference somehow uncovered more treble energy. One track I heard this on was "Lonnie's Lament," from John Coltrane's album Crescent (heard via Qobuz in 24/192, originally on Impulse!). Images were dryer, less spot-lit, a little more sharply defined via the Bartók, with clearer edges. Curiously, I found this difference easier to hear on older recordings than on newer, cleaner ones.
After I'd finished with my listening, dCS alerted me to a possible error in my methodology: Grouped zones in Roon may not be bit-perfect. If the two DACs aren't both receiving the same, unaltered data, the test is invalid. But in my case, the data apparently were bit-perfect: When I played an MQA file through each of two grouped zones to both MQA-enabled DACs, both indicated that they were decoding MQA, which, according to both Roon and MQA experts I talked to, is a clear indication of bit-perfect playback. Indeed, dCS writes in the Bartók manual, "MQA decoding is not possible if the original MQA data has been changed." Enno Vandermeer, Roon's CEO, told me via Facebook Messenger: "It's certainly possible" to get bit-perfect playback in grouped zones, "especially on a stable network, We just can't guarantee that the slave zone will be bit-perfect."
Just to be sure, I decoupled the zones, did some additional listening, and heard some additional differences, always subtle, always in the Bartók's favor.
Wrapping up
The Ring DAC is based on technically solid ideas, with clear, definite advantages over other DAC technologies. There is value here—absolutely—but, even though this is dCS's least expensive component that can play music, it's still not inexpensive. A well-designed chip-based DAC can achieve superb measured performance, both measured—consider the Benchmark DAC3 HGC I reviewed for the November 2017 issue of Stereophile—and subjective.
I asked dCS's John Quick whether the best chip-based DACs have closed the gap with dCS's Ring DAC: After all, one main problem with chip DACs has been processing power, and there's far more of that now than there was a couple of decades ago. He acknowledged that chip DACs Big IC companies, he pointed out, tend to go where the money is. There's a risk that in their efforts to make DAC chips more mobile-friendly—more compact or energy efficient—the unique needs of audiophiles will be sacrificed, or at least neglected. After decades of improvement, DAC chips could start to get worse. He's making a serious point. At the moment, though, chip DACs are very good, approaching the performance of dCS's Ring-DAC technology.
After all these words, I still haven't written about filter choice (I listened mostly with Filter 4), DXD vs DSD upsampling (I tried both but mostly listened with the default DXD upsampling), or the Bartók's headphone amplifier, which the company seems to have invested much thought and resources in.
Before I began to listen seriously to the Bartók, and later, when I was done with serious listening, I spent a lot of time with it playing music in the background at low to moderate volumes. I write "in the background," but I often found my attention drawn to the music, and it was always a pleasure. Sure, we audio geeks put music at the center of our lives, but there is much to be said for components, and systems, that can be enjoyed unobtrusively.
Someday, I hope to hear the company's grand flagship system, the full Vivaldi stack, in a familiar system. For now, the Bartók seems ideal: How much better could the Vivaldi be? (I could ask the same about expensive, top-end DACs from other companies.) Until then, I'll consider the Bartók the state of the art, while noting that there are other, cheaper products nipping aggressively at its heels.
Footnote 4: I chose the DAC+ because, having reviewed the very similar Brooklyn Bridge just last month, I feel familiar with its sound. I chose the DAC+ instead of the Brooklyn Bridge because it has an AES/EBU input.
Fun idea, boring outcome. I listened to tracks from several albums—but could hear no differences among the three inputs. It was easy to tell why: There was too much noise in the source. Whether it's because of the quality of my vinyl front-end, the condition of records I played, or the intrinsic limitations of vinyl, anything added or subtracted from the sound by the ADC and DACs was swamped by noise and distortion from the analog front-end. I would not call this test successful, but I did learn from it.
It convinced me that 24/192, with a high-quality ADC and DAC, is sufficient to capture the sound of analog—anyway, my analog. PCM proved a transparent container, preserving the sonic character of vinyl in my system while adding or subtracting nothing audible.
I moved on to long-term listening, now with digital sources, using my Roon Optimized Core Kit (ROCK), which I built on an Intel NUC, directing data to the Bartók via the network connection, from Qobuz and Tidal and files collected on my Synology NAS.
There's something to be said for those DACs that have been influenced, directly or indirectly, by the needs of pro audio. Pro audio needs DACs that approach transparency, to the point at which several passes through the DAC (round-tripped via a companion ADC) change the sound negligibly. That puts consumers at a huge advantage, since the music only passes through our DACs once—any alteration in the sound will, therefore, be a fraction of what it would be in a studio.
DACs that achieve a high degree of transparency must sound very much alike: If A and B both equal C, then A must equal B. Therefore, DACs that do have their own sound—those that stand out from the crowd—must deviate from transparency. This isn't from listening tests or technical theory; it's the most basic logic.
Toward the end of my time auditioning the Bartók, I went back to the Mytek Brooklyn (in fact I now used the Brooklyn Bridge, to take advantage of its Ethernet input). I set up two zones in Roon, then grouped them. I matched volumes. Then I listened to a few tracks, switching between them rapidly—now with digital sources, delivered via Ethernet from my collection of files plus tracks from Tidal and Qobuz—and also listening to the same track for several minutes, via one DAC and then the other.
The Ring DAC is based on technically solid ideas, with clear, definite advantages over other DAC technologies. There is value here—absolutely—but, even though this is dCS's least expensive component that can play music, it's still not inexpensive. A well-designed chip-based DAC can achieve superb measured performance, both measured—consider the Benchmark DAC3 HGC I reviewed for the November 2017 issue of Stereophile—and subjective.
After all these words, I still haven't written about filter choice (I listened mostly with Filter 4), DXD vs DSD upsampling (I tried both but mostly listened with the default DXD upsampling), or the Bartók's headphone amplifier, which the company seems to have invested much thought and resources in.
Before I began to listen seriously to the Bartók, and later, when I was done with serious listening, I spent a lot of time with it playing music in the background at low to moderate volumes. I write "in the background," but I often found my attention drawn to the music, and it was always a pleasure. Sure, we audio geeks put music at the center of our lives, but there is much to be said for components, and systems, that can be enjoyed unobtrusively.
Footnote 4: I chose the DAC+ because, having reviewed the very similar Brooklyn Bridge just last month, I feel familiar with its sound. I chose the DAC+ instead of the Brooklyn Bridge because it has an AES/EBU input.
