The dCS Puccini Plays DSD
Input: 2 RCA Coax S/PDIF, 1 Wordclock
Output: 2 RCA Coax S/PDIF, 1 Wordclock, RCA (single-ended) or XLR Balanced analog output
Dimensions (H x W x D): (W) 460mm x (D) 410mm x (H) 101mm / (W) 18.0” x (D) 16.1” x (H) 4”
Weight: 12.1kg / 26.6lbs
Availability: through Authorized Dealers
US Distributor's Website: dCS North America, LLC
UK-based digital music pioneer dCS (Data Conversion Systems, ltd.) has long been known as a company whose products reflect cutting edge digital technology. Yet one of the big paradoxes in the 25-year history of an organization that was part of the original working groups that developed the SACD format, and that originated DoP (DSD over PCM), the protocol that enables transfer and playback of DSD music files packed into a PCM frame, is that it has been slow to incorporate DSD file playback into its products.
dCS products have long been capable of upsampling PCM data to DSD, and many have also been capable of playing the stereo DSD layer of SACDs. The first dCS product capable of DSD file playback per se, however, was the four-stack, state-of-the-art Vivaldi system, which was released in September 2012. All other products in the line required software upgrades.
dCS’s stand-alone Debussy DAC received the DSD upgrade, which should technically be called “The DoP upgrade on the DAC’s digital inputs,” at the end of 2012. Then followed software upgrades to the Scarlatti DAC in July 2013, the Puccini CD/SACD player in September 2013, and the Paganini DAC in November 2013. It should be noted that while a host of DACs can now decode double-rate DSD – 5.6448 mHz sampling rate, aka DSD 128 – and dCS’s Ring DAC is certainly capable of doing as much, the company’s current upgrade only enables playback of DSD 64 (SACD resolution). The rationale for this – that very little DSD 128 material is currently available – suggests that another software upgrade will arrive in due course. With the release of Sony’s new flagship music server/DAC, the HAP-Z1ES ($1999), which not only plays double DSD but also includes a DSD “remastering engine” that upsamples most audio formats to double DSD, the fire has already been lit under dCS’s tail.
This review concentrates on DSD file playback on the Puccini CD/SACD player, which I use in conjunction with dCS’s Scarlatti Clock Plus. Much too soon after I successfully installed the Puccini’s 1.41 software upgrade pre-release in mid-August 2013, my entire reference system was dismantled so that we could stage our house prior to listing it for sale. Anyone who has undertaken the challenge of selling and buying a home while attempting to meet all work and family commitments can understand the lag time between the software release and the appearance of this review.
Although installing the DoP upgrade on the Puccini’s digital inputs was a piece of cake – I inserted the upgrade disc into the player, followed the instructions, and took the dogs for a walk while the installation proceeded – the upgrade was only successfully achieved after I discovered a bug in the initial 1.40 prototype. After its installation, when I began to play back DSD files on my Macbook Pro using either Audirvana or Pure Music, I heard what I thought sounded similar, but not quite identical to monaural playback.
Conference calls with Rob Robinson of Pure Music and John R. Quick of Tempo Sales (the U.S. distributor of dCS), as well as emails to Damien Plisson at Audirvana, confirmed that their products, as well as my Macbook Pro, were AOK. Eventually, dCS discovered that a glitch had left the Puccini playing left channel DSD information through both left and right channel outputs.
With that bugaboo fixed, I am happy to report that DSD file playback proceeded flawlessly. Whenever I cued up a DSD file in Audirvana, the Puccini recognized it and indicated DSD on its screen. Although the Scarlatti Clock Plus that I use with it could not recognize that it was passing DoP per se, its screen indicated the 176.4 sample frequency with which DoP is packaged.
Before I discuss the results my listening tests, I want to share information about dCS’s development of the DoP protocol, which I’ve edited down from a Skype chat with company head David J. Steven. Steven, 33, grew up learning the ins and outs of the audio industry from his father, company founder David Steven. The two worked side-by-side for a year before the elder Steven passed away in 2009, and his son took over leadership.
dCS started in radar and military work, and continued with it until the Cold War cooled. Then we moved into professional recording, and made some of the first hi-res DACs.
We’ve gone through peaks and troughs. At one time, there were 25 developers. Our company now consists of 17 individuals, with 6 full-time employees on the technical team. Everyone in our current core team has been with the company at least 10 years – some long enough to make every product we’ve ever made break in one way or another.
Andy McHarg, Technical Director of dCS, came to me August before last  with the idea that we could basically pack DSD into a PCM frame. At that time, we’d seen the move to DSD audio and network streaming. But we also knew that everyone worked with the silver disc. It had been an even playing field, where everyone knew what they were getting and how to work with it until network streaming, firewire, USB, and DSD came along.
Andy came up with the idea of how to transmit DSD over PCM. Up to then, everyone could do 44.1 up to 192. Andy had figured out how to arrange DSD within a 176.4 frame so that basically, as long as the DAC knew what to do with 176.4 data, none of the hardware upstream had to change.
DSD is 1 bit data that has only two states, a 1 and a 0. But you have this huge sample rate, potentially 64 times CD, so the math fits quite nicely into a 24/176.4 stream. The first 8 bits are the marker that tells the DAC that it is DSD, and the next 16 bits are the audio. It’s not converted to PCM; it stays DSD all the way. The DoP protocol was almost like just passing the data through without having to start from scratch.
After Andy explained to me how the DoP protocol would work, I contacted Rob Robinson from Pure Music and asked him if he could enable Pure Music software to support DSD. Within a week, he had developed a way, using one of Pure Music’s pull down menus, to select the dCS method for transferring a DFF and DSF file. The software would pack the data in such a way that we could then play it back.
From there we started to contact some other companies, including Charlie Hanson and the guys at Ayre, and Andreas Koch at Playback Designs. Claude from Merging Technologies, who make the Pyramix work station, and the ESS guys were also involved – every part of the recording and playback chain was covered – so once everyone agreed on the protocol, it really started to gather momentum.
Andreas had been working on a similar type of implementation, but I don’t think it was as simple as Andy had proposed. The more people we got involved, the more it snowballed. Most of the group deferred to Andreas and Andy, who moved it forward and drove it through to being adopted.
We did a couple of demonstrations in the UK, and built up a groundswell for how DoP could be accomplished. The legwork had be done to by the DAC manufacturers, who had to adopt the software in the DAC so that when it saw the stream of data, it immediately identified it as a DSD file and knew how to process it. But that was better than having all the playback guys go off and come up with different ways to play back DSD.
I pitched DoP to a bunch of people, including Damien of Audirvana. There was no USB class that would support DSD, and getting a change to the USB classes would have taken forever. There was no operating system support for DSD playback in either Windows or Mac, and we didn’t want to develop special drivers which would have been a pain in the ass, so the easiest thing to do was to let the DAC do the difficult legwork and piggyback on an existing format (24/176.4). We convinced people there was no point in trying to approach Microsoft or Apple. That helped keep us on track, and it came to fruition pretty quick.
The rapidity with which DoP caught on took us by surprise. It took us awhile to catch up with all our products. I was expecting it to take awhile, and that there’d be some objections. But Andy and Andreas got hold of it, and published a draft of a white paper. People kicked it around, we made a few tweaks, and within six months, it went from being a great idea to becoming the standard.
As DoP started to snowball, more and more people got involved. It was hard enough keeping up with the traffic, never mind rolling out our own upgrades. And because we have 15 products, and all of the hardware within our units is developed by us and maintained by us, we had to change, verify, and test 15 proprietary code streams.
At the same time, some of the chip manufacturers, specifically ESS Sabre, got involved as well, so everyone who used their DAC chip immediately had DoP support before we did. But again, that’s just because all the lines of code inside dCS gear are written by us.
Every dCS product has what’s essentially a motherboard inside it that can be programmed to be a master clock, DAC, or transport. It’s the same hardware; it’s the software that changes. That was why it was very easy for us to adopt DoP.
The chip that’s in the Puccini is a Field Programmable Gate Array (FPGA): it’s basically a computer chip with millions of configuration options. There’s no DAC chip or microprocessor from someone else; this is what makes our software so flexible.
True DSD Playback?
As I was putting the finishing touches on this review, I came across Michael Lavorgna’s Nov. 6 Q&A with Andreas Koch (see Q&A) of Playback Designs. It includes the following exchange:
Michael: Some manufacturers, including Gordon Rankin of Wavelength Audio, have pointed out that there are no current production DAC chips that handle DSD natively. If this is the case, are all DSD DACs that use current production chips converting DSD to PCM internally?This led me to wonder, given that dCS’s method of D-to-A conversion uses proprietary technology rather than an off-the-shelf DAC chip, if dCS products actually handle DSD natively. After I posed the question to David J. Steven, he sent the following reply from Andy McHarg. I don’t pretend to be a technical whiz, but I am capable of cutting, pasting, and editing information from others far more qualified to address such matters.
Andreas: Most DSD DAC chips, if not all, lowpass filter the DSD signal to get rid of the high frequency noise (see Fig.2) before the signal gets converted to analog. The resulting signal behind this lowpass filter (and before the actual analog conversion) may still have the same sample rate as the original DSD signal, but it is no longer 1 bit. So can this still be considered DSD?
It is all a matter of definition: DSD, or Delta-Sigma Modulation, can be encoded with more than just 1 bit, and PCM can have a very high sample rate. When looking at the criteria of word length and sample rates only, the boundary between DSD and PCM can become fuzzy. Sometimes it is more useful to distinguish DSD from PCM in the frequency domain and look for the characteristic behavior in the higher frequency bands, as pointed out in Figure 2 above.
Since I believe that the source of the sonic difference between DSD and PCM lies in the difference between how these signals compare in their behavior for higher frequencies, I also believe that filtering a DSD signal with an aggressive filter to flatten the upper frequencies will make it behave and sound more again like a PCM signal.
The reason why chip manufacturers like to add an aggressive lowpass filter at the input to their DACs is simple: the analog output measures better. Whether it sounds better with real music signals instead of measurement tones is an entirely different question. Similarly, most audio manufacturers and even end users who do not understand DSD are mostly concerned with Signal-to-Noise performance even at high frequencies and would not choose a chip with a frequency response that is not completely flat and optimally low all the way up to Nyquist.
With that, the answer is yes, most if not all DSD DAC chips convert to PCM before converting to analog.
That opens the door for discretely built DACs that don’t have to follow the criteria of measurements with sine waves, but rather the listening experience with real music signals.
The short answer is: Andreas makes some good points in his interview, namely that, especially within the innards of the DAC, the distinction between PCM & DSD is somewhat blurred.
The dCS Ring DAC is a multi-bit, noise shaped architecture that you can view as lying somewhere in between DSD & PCM. Inherent to DSD is an awful lot of ultrasonic noise, due to the aggressive noise shaping required to get in-band performance with only one bit. This ultrasonic noise, much like the "image" effects you get with a NOS DAC, can cause problems with equipment downstream. Amplifiers in particular are designed with a certain input bandwidth in mind, so having lots of energy at places where the amp isn't expecting it can cause misbehavior, ranging from IM products in the audio band to the amplifier shutting down. To alleviate this problem, the SACD specification requires, in addition to a wideband low pass filter, the ability to switch in an optional 50kHz filter to limit misbehavior of downstream equipment.
The Ring DAC format itself is 5-bit / 2.8224MHz. This means that dCS has a DSD processing block implemented in the FPGA that can convert between DSD and the 5-bit format. Because we have effectively 32x more resolution, we can use this to implement a selection of different filters without degrading the original signal.
In all dCS products, each filter removes more ultrasonic noise than the last. Filter 1 is minimal filtering, but may not be compatible with all amplifiers. By distinction, Filter 4 filters out nearly all of the ultrasonic noise, effectively limiting the bandwidth to be much more like PCM. Because all of this processing is done inside an FPGA, we are able to modify the processing block should we come up with an improved scheme in the future.
So, as always, the answer to the question is, (a) “It's a bit more complicated than you might think," and (b) "It depends on the choice of filter.”
To discover how the sound of the Puccini’s DSD file playback compared to that of the DSD layer of hybrid SACDs, I picked three native DSD recordings – Mahler Symphony No. 2 from Michael Tilson Thomas and the San Francisco Symphony (SFS Media hybrid SACD, Blue Coast Studios DSD download), Vivaldi’s La Cetra by violinist Rachel Podger and the Holland Baroque Society (Channel Classics hybrid SACD and DSD download), and a Wagner compilation by Ivan Fischer and the Budapest Festival Orchestra with soprano Petra Lang (Channel Classics hybrid SACD and DSD download)—that I had in both DSD file and hybrid SACD format. As it turned out, in the limited time I had for listening comparisons before my system was put into storage, I found Lang’s sound so unattractive that I basically skipped that recording.
For reasons I cannot explain, my copy of Pure Music, which had been set up by Rob Robinson, refused to cooperate; with limited time before I had to put my system into storage, I bailed on trouble-shooting, and played all DSD files via Audirvana. My Macbook Pro, dCS Puccini, and dCS Scarlatti Clock Plus were all supported by Synergistic Research Tranquility Bases, in turn powered by a Transporter Ultra SE, Power Cell 4 SE and Synergistic power cables. Pass Labs XA 200.5 Class A mono block amplifiers and Wilson Audio Sasha loudspeakers completed the chain, held together and powered by Nordost Odin cabling save for a Synergistic Research Active USB cable on the computer. Although supports were not identical under all equipment—I used a mixture of Nordost Brass and Titanium Sort Kones and Magico Q-Pods—I did my best to create a level playing field when I compared DSD files to SACD playback.
Switching from SACD to DSD file playback required a simple remote control shift of inputs on the Puccini. Once set to USB input, both the Puccini and Scarlatti flawlessly recognized files as DSD, and treated them accordingly. That’s really all there is to say about that.
As someone who has spent many years carrying Fischer’s Budapest Festival Orchestra recording of Mahler’s Symphony No. 2 room-to-room at audio shows, it was a shock to revisit MTT’s version and discover its greater wealth of orchestral detail. I was also struck by the SACD’s “amazing width and breadth.” But listening to the DSD files of the same recording via Audirvana, I was struck by its greater richness of tone. At 6:25 into the first movement, for example, I noted how well the file conveyed the emotional transition from light to dark, as well as the sweetness of Concertmaster Alexander Barantschik’s Guarneri violin (formerly played by Jascha Heifetz). As I continued my comparison, I thought the SACD conveyed a bit more light, and the DSD file more midrange richness. I also noted that file playback set the orchestra a little farther back than did the SACD. But did one sound “better” than the other? Not to my way of hearing and feeling music. Each provided a different window on what to me sounded very much like musical truth.
How much perceived differences in sound had to do with the Puccini itself, rather than the choice of file playback software (Audirvana), the influence of different sounding equipment supports under different sources, and/or the fact that, by adding a Synergistic Research Active USB cable to my otherwise all-Nordost cable chain, I had contradicted Nordost’s statement that their cabling performs optimally when it is not mixed with other brands, I could not determine. I did try the three other USB cables I had on hand – Light Harmonic’s Lightspeed, which didn’t stand a fighting chance because no one had told me that it had not been broken in prior to shipment, and needed many hundreds of hours to come up to snuff; Nordost’s Blue Heaven; and AudioQuest’s Diamond – and determined that the Synergistic cable yielded the best sound in my system. So that’s what I stuck with throughout my listening tests. When Nordost eventually releases a higher level USB cable, it will be time to listen once again.
Next, I played DSD files of the first two movements of La Cetra’s first Vivaldi violin concerto on computer. “Tremendous sense of acoustic space– beautiful depth to bass viols and the lowest notes of the organ,” I wrote in my notes. Switching to the SACD, my listening companion, Bay Area Audiophile Society member Jeremiah Horn, found the sound “less thick, lighter, and less weighted toward the lower octaves.” He also felt the violin sounded a bit more “complex” which I expect means that he heard more harmonics and detail in the sound. Whether they were in fact present, or simply recognized after hearing an unfamiliar recording for a second time, I am not certain.
Even before the upgrade, the dCS Puccini/Scarlatti had impressed me as the finest digital playback system I’ve ever heard in my reference system. Nor was I alone. At the last meeting of the Bay Area Audiophile Society that we held at Casa Bellecci-Serinus, BAAS members had the opportunity to hear, for the first time, the latest additions to my reference system: the dCS Puccini/Scarlatti Clock Plus combo, supported by Tranquility Bases, along with the Wilson Audio Sashas and Synergistic Research’s active USB cable. (They were already familiar with the sound Nordost Odin cabling and Pass amplification, as well as the acoustics in my room.) The sound was so convincing, and so musical, that several BAAS members left muttering that they had no choice but to go home and immediately figure out what component to upgrade next.
The installation of dCS’ DoP upgrade on the Puccini’s Ring DAC’s digital inputs was a piece of cake. In listening tests, differences in tonal balance and soundstage positioning were evident when comparing the sound of stereo DSD file playback with that of the stereo DSD layer on a hybrid SACD of the same music. Whether these differences were a result of the choice of music file playback software, a difference in the sound of equipment supports, the particular USB cable, or inherent in dCS’ electronics and software was impossible to determine. What was beyond question was the excellence of sound, regardless of medium.