Because the 2.0 designation refers to a firmware update, all exterior and interior parts of the Bartók 2.0 remain identical to those of the original Bartók. The inputs on the back panel include an Ethernet connection (required for network access), two AES/EBU inputs, three S/PDIF inputs (one RCA, one BNC and one TosLink), and a USB/PC input for playing music files stored on hard drives. Three BNC connectors enable users to add a dCS clock for better jitter rejection. Owners of a dCS Rossini or Vivaldi SACD transport can utilize both AES/EBU connectors simultaneously to send encrypted DSD data to the Bartók’s Ring DAC. This is a very complete set of inputs, lacking only HDMI, but considering the stereo-only target market for the Bartók, I don’t see this as a problem.

Once you’ve chosen a music source, either from a network or a hard-wired source, you can choose how the Bartók will decode the digital files into audio. The Bartók's Ring DAC runs at the DXD data rate of 24 bits/352.8kHz. All incoming signals are upsampled to DXD speeds. If you are playing 16-bit/44.1kHz data from CDs or a stream, you can choose from four different filters. Filter 1 has the sharpest roll-off characteristic, resulting in the best rejection of out-of-band noise and distortion. The tradeoff is that it has the least sonically efficacious transient response of the upsampling filters. Filters 2 through 4 are progressively less effective at rejecting noise and distortion and have slower roll-offs above the sampling frequency, but they exhibit better transient response.

Another option is to convert the DXD data to DSD or DSDx2 right before the analog output. You can choose from five DSD filter choices. The general rule from the Bartók owner’s manual is “. . . if Filter 1 sounds too harsh, then try Filters 2-5 as they progressively reduce out-of-bandwidth noise, sounding smoother.” One caveat: Filter 4 is included for ease of testing at dCS and puts a brickwall filter above 20kHz, so it is not a good choice for listening.

Now let’s discuss what's changed about the original Bartók: the 2.0 firmware upgrade, which owners can install themselves, as TAB writer Vance Hiner did. dCS Vivaldi and Rossini owners have enjoyed this for a few years now, and Bartók owners clamored for a version to apply to their units.

If a music source is at 176kHz or a higher sampling frequency, two additional filters become available. Filters 5 and 6 use special Gaussian filters that have almost no overshoot on transients. The tradeoff is a more aggressive roll-off characteristic, requiring high sampling rates to maintain good frequency response in the audible range. Filter 5 is the newest choice available and comes recommended by John Giolas (Vice President of Sales and Marketing for dCS) as a particularly good-sounding filter for DSD listening. I found it to be my favorite.

The upgrade adds two new mappers to the Bartók. Mappers are high-speed circuits used to control the Ring DAC and are a fascinating part of the inner workings of the Bartók 2.0. Without going down a very deep technical rabbit hole, dCS explains that “The Mapper decodes the data to a minimum of 32 digital signals, one corresponding to each binary state, then scrambles the order of the signals. The Mapper randomizes small matching errors in the Ring DAC core, converting what would be harmonic distortion to a small amount of extra noise.” Why would this be considered a good thing? Well, dCS believes that by randomly selecting the bits to be fed to the Ring DAC, any possible errors are minimized, as well as compensating for performance variations as parts age.

Users are able to listen to the “Classic Mapper” by choosing Mapper 2. This mapper was used with the Bartók I reviewed last year. The newly added Mappers 1 and 3 run at over twice the speed of Mapper 2. dCS maintains that this increase in operating speed contributes to the improved sonics of the Bartók 2.0. The improvement is described as “more detailed, expressive, tonally-complete and full-bodied.” While I would agree with this assessment, I would add that I hear the improvements more as a reduction in distortion or noise surrounding individual instruments or voices. They are more easily located in complex mixes and the images emerge from inky-black backgrounds. The off-mike whisperings on Holly Cole’s Temptation [Metro Blue CDP 7243 8 31653 2 2] fooled me into thinking someone was talking to me from the other room. The vocal realism was startling.

Of the new mapper choices, Mapper 1 exhibits the lowest possible distortion in the output, or, if you prefer, the greatest linearity. In contrast, Mapper 3 minimizes odd-order harmonics in favor of even-order harmonics, somewhat similar to the behavior of a single-ended tube amp. Because of how they handle distortion, some have considered Mapper 1 as being more “solid-state” in tone, while Mapper 3 is more "tubelike." While I can appreciate that Mapper 1 has speed and bass power that I equate to solid-state gear, I don’t think anyone with a tubed-output DAC will be fooled into thinking Mapper 3 mimics the sound of tubes. Instead, Mapper 3 has a softer initial attack and a richer tone, so the tube descriptor is helpful in characterizing the differences. I ended up choosing either Mapper 1 or Mapper 3, depending on the type of music being played. To my ears, Mapper 2 doesn’t hold up very well in comparison to the new mappers, but do try it for yourself.

After attaching the Bartók 2.0 to your network via an Ethernet cable and downloading the Mosaic 2.0 app, you can select from a variety of music sources: UPnP for networked data or devices, and USB for memory sticks or external hard drives. Music streaming choices are Deezer, Qobuz, Tidal and Spotify, and Mosaic 2.0 includes Tidal Connect plus Spotify Connect. Apple AirPlay is supported, too. Finally, Radio lets you access Internet radio stations and your favorite podcasts.

I did spend some time streaming AirPlay during the Bartók 2.0's break-in period and was pleasantly surprised by the audio quality. AirPlay previously sounded tonally thin, but with the Bartók 2.0 it is thoroughly listenable, even if it gives up some resolution to Qobuz’s high-resolution files. While I wouldn’t choose to select AirPlay if the same title is on Qobuz, it sure is nice that the Bartók 2.0 can elevate AirPlay’s sonics when I discover something interesting on Apple Music.

If you subscribe to Tidal, its available MQA files will signal the Bartók 2.0 to choose the M1 filter and implement full unfolding of those files. You could also override the M1 filter and use one of the first six filters, but that defeats any MQA advantages, so it’s hard to imagine selecting this option. My experiences playing Peter McGrath’s MQA recordings on the Bartók 2.0 were extremely positive.

One bit of exciting news for many streaming listeners will be the addition of Spotify Connect and Tidal Connect. I devoted some listening sessions to Tidal Connect and, unfortunately, became somewhat ambivalent about its sound quality. It did sound good at times, but the character of the recordings changed, with the bass becoming fuller and lacking control. And image placement was not as precise as when the Bartók 2.0 was being fed the same material using Tidal through the Mosaic 2.0 app. This was a bit surprising, as I’d hoped Tidal Connect would be equal in sound to Tidal through Mosaic 2.0 and offer a different user interface for those individuals used to Tidal’s layout. In my system, I chose to stay with the Mosaic 2.0 app for the best sound quality.

I still like the Bartók’s understated, minimalist design, but after seeing dCS gear at shows, I feel the optional silver finish adds a bit more visual interest. The review unit was finished in matte black and equipped with the headphone option, like last year’s review unit. One thing that did not change from last year is that the Bartók 2.0 sounds much, much better when left on all the time. I do turn off the screen and logo lighting, but I never switched the unit off unless there were thunderstorms in the forecast. Granted, almost all digital gear sounds better fully warmed up, and I now know why this is so beneficial for the Bartók 2.0. The quartz crystal oscillators used inside are extremely sensitive to thermal changes and their temperature needs to be precisely maintained for optimum performance. dCS actually heats the crystals to that precise temperature, which is the reason that I felt the performance was significantly better after at least sixty minutes of warming up. The Bartók 2.0 runs barely warm to the touch, so my advice is to leave it on at all times for the best sound.

continue to be impressed with the Bartók 2.0 as a source and as a preamp using the XLR inputs of my BAT VK-60 amp. This combination was used for part of my review of the Wilson SabrinaX and it worked equally well with my usual Altec Valencia and Opera Consonance M12 speakers. I concede that I may be losing some low-level resolution as a result of the high sensitivity of these speakers. I generally use the volume knob setting around -20dB for comfortable levels. I did try the Bartók 2.0’s variable output setting at .6 volts, but it was just a little too soft for satisfying listening, so the 2V setting was my best compromise. At that, the low-level reverb tails were well reproduced, so I may not have as much of a problem as it might seem. And I should mention that when using the Bartók 2.0 as a preamp, I appreciate the inclusion of an actual volume knob on the faceplate, when it would have been far easier for dCS to force me to use a virtual volume slider of the Mosaic 2.0 app.

In my review of the original Bartók, I stated my preference for the DSD setting over DXD by a large margin, and I eventually settled on DSD upsampling with Filter 3 for all my listening. The 2.0 update has changed that. The new mappers allowed me to enjoy DXD for most recordings, and I only used DSD for specific types of recordings. Quite a personal change of opinion, but one of the beauties of the Bartók 2.0 and Mosaic 2.0 was being able to sit on the couch and make these switches rapidly on my iPad. Here are three examples of how I used DXD and DSD.

I listened to quite a bit of piano music this past summer and even did a couple of recording sessions for solo trumpet and piano, so my ears were attuned to piano tone. A favorite performer right now is Vikingur Olafsson, and I particularly enjoy his album Debussy-Rameau on Deutsche Grammophone [24-bit/192kHz stream on Qobuz]. I enjoy his technical facility and musicality, his knowing when to add just the right amount of sentiment without being overly romantic. The problem for me is that DG has chosen too distant a perspective for the piano, blunting the percussive attack of the hammers. This is not a deal-breaker, but even turning up the volume didn’t help add realism to the performance. With the Bartók 2.0, I had a solution to this problem. I select DXD, Mapper 1 and Filter 1 as a way to emphasize the leading edge of each piano note. I heard Olafsson’s recordings with better keystroke attacks without a trace of hardness. I tried some of the other filter choices, but Filter 1 worked best. Attending a recent vocal recital by Christine Brewer, accompanied on piano by Jennifer Judd, reaffirmed my choice for DXD and Mapper 1 as reproducing the best approximation of live piano percussiveness with my system.

However, my preferences were reversed when I listened to Steely Dan’s Two Against Nature CD [Giant 9 24719-2]. Ricky Lawson’s hi-hat and snare slaps on “Gaslighting Abbie” were a touch too harsh for my speakers, so I experimented and settled on DSD, Filter 5 and Mapper 3. This took the edge off cymbal harmonics and sounded much more natural to me. The effect of switching mappers or output choice was not like using an equalizer in any way. The mappers only affect the nature of the overtones, not the strength of high frequencies, leaving the treble fully extended. What a luxury to be able to sit on the couch and modify the digital decoding schemes to better serve recorded music. Mapper 3 was also a good choice when massed-string recordings veered toward strident or edgy. I was not tempted to try to catalog a filter/mapper/upsampling setting for every album, although one could certainly do that.

Occasionally I chose the DSD and Mapper 3 settings when just listening for pure pleasure. On Hakan Hardenberger’s newest release, French Trumpet Concertos on BIS [Qobuz 24-bit/96kHz stream], the trumpet tone was so beautifully captured that I chose DSD and Mapper 3 for their greater warmth and vividness. Hakan’s performance on the first track, “Allegro et Cadence” by Henri Tomasi, is so effortless and musical that I found the DSD/Mapper 3 enhanced my enjoyment. This setting shifted my attention from the initial attack of the notes (the trait I relished when using DXD and Mapper 1) and redirected me to the tone of the trumpet. This effect was subtle, and if Mapper 3 was emphasizing the even harmonics, it was still musically truthful and well worth trying with favorite orchestral recordings. I also enjoyed how beautifully the Bartók 2.0 rendered the string tone of the Royal Stockholm Philharmonic on this release. This is not your typical audiophile release, but it is a recording I’d certainly recommend when auditioning the Bartók 2.0.

I should also mention that my experience with the DSDx2 setting sometimes resulted in subtly better sound, but it was hard to pin down the differences. It was most apparent when using headphones with Alon Sariel's Plucked Bach [PentaTone PTC 5186985]. Switching back and forth between the two DSD settings, I felt DSDx2 offered greater resolution and better texture in the hall's reverb. When sitting across the room listening to speakers, it was much harder to discern any change. If I were a full-time headphone listener, however, I think I’d choose DSDx2 over DSD, if for no other reason than it produced a more clearly defined sense of acoustic space.

It might be appropriate to ask why dCS has given users so many filtering, upsampling and mapping choices. Surely one combination of these choices is the absolute best. In his article on the Bartók 2.0, Vance Hiner wrote that dCS developed the filter choices as a result of research into psychoacoustics. Each filter choice has to do with one's listening preferences and biases. I’ve spent quite a bit of time thinking about that statement, and here’s my view on the broad number of choices. I own speaker designs ranging from full-range, single-driver models to vintage horns to electrostatic/dynamic hybrids to LS3/5a monitors. No two of these speakers sound remotely similar and yet each has a number of compelling sonic virtues. Visitors who regularly attend listening sessions at my apartment have their speaker favorites and dislikes. Each guest has different listening tastes, and some even have their favorite choice of tube amps to better mate with each pair of speakers. These listener preferences even extend to certain combinations of amplifiers and speaker cables.

Now, keeping in mind the idea of individuals expressing hearing differences or listening preferences, what if a company could devise a way for a DAC or streamer to offer multiple settings aimed at accommodating those differences in hearing or musical taste, particularly if those settings were available without degradation of the signal or loss of fidelity? With the Bartók 2.0, owners have the flexibility to tailor the sound according to their own preferences. Or if there’s a change elsewhere in the system, whether that’s new cabling, new amps, or new speakers. Owners can also take some solace in the upgradability of the decoding system whenever dCS discovers a higher-performance option. One click downloads the upgrade to your device and a new DAC is ready to play. While it’s expensive for companies to develop new programming, I feel that some level of future-proofing is advantageous for both sellers and consumers of high-end digital audio equipment.

I spent less time with the Bartók 2.0's headphone section than I did during my original review, but it sounds as good or better than last year’s sample. I still like the Bartók's Crossfeed and Expanse settings for music that has a lot of ambient cues. The circuit is designed to minimize the feeling of sound being trapped between your ears when listening via headphones. The Expanse settings came about, through additional research by dCS, as an improvement over simple crossfeed circuits that have been used previously. A release that responded well to either setting was Plucked Bach. With the arrangement of Bach’s Cello Suite no.1, BWV 1007 on archlute, the recorded sound was warm and reverberant, and Crossfeed or Expanse added lushness without swamping the sound with too much echo or reverberation.

I did use the unmodified, or standard, headphone feed when I was concentrating on the words and musical delivery in simple tunes. Becky and Rachel Unthanks' album Diversions, Vol.4: The Songs and Poems of Molly Drake [Qobuz 16-bit/44.1kHz stream] features plaintive tunes written by Nick Drake’s mother. Particularly when listening to the sister’s rendition of “I Remember,” adding any effects to the music diluted the song’s sorrow about the failing of long-term love affair or marriage. I much preferred confining the sound squarely between my ears without Expanse or Crossfeed.

hat a reunion this turned out to be. Instead of adding a thicker faceplate or glitzy touchscreen, dCS has focused on features and performance with the Bartók 2.0. Excluding dCS’s own Rossini and Vivaldi DACs, I can’t think of another competing product with a greater level of user-adjustable sonic parameters than the Bartók 2.0. Once you begin to explore the new Mappers 1 and 3, both with DXD and DSD, I think you’ll discover that the 2.0 version easily exceeds the performance of the original. The new Bartók 2.0 with its accompanying Mosaic 2.0 app is simply brilliant.