Northern California south of San Francisco is called Silicon Valley. This area is known for high tech innovation and education. The area in and around Cambridge, England is called Silicon Fen. Much like Silicon Valley, Silicon Fen is known for high tech innovation, businesses, and the tech-centered University of Cambridge. It's fitting that audio innovator dCS is located in this cradle of digital technology.
Driving through the Cambridge countryside it would be simple to pass right by the dCS factory. dCS' very nondescript building is where its less than twenty employees produce some of the industry's best products. dCS prides itself on building its products by hand, in low volume, in the United Kingdom. As I walked around the building I stopped to talk to engineers in a few different departments. Each engineer was happy to discuss how dCS products are designed. I was captivated listening to the fanaticism that goes into building each component. Ray who designs the chassis and metal work for all dCS products spent an extended period of time explaining the difficulties of producing such high quality metal finishes. Producing products year after year that look identical is critical. Many customers purchase one dCS component at a time and will not accept two components from the same series without an indistinguishable finish. If the metal's color is off just a smidgen the customer will reject the product as it's placed in a system with existing dCS components. In addition to the external build quality Ray showed me some dampening materials that were in for R&D and explained how dCS chassises are designed using layers of metal. It's a fascinating process to say the least.
Each dCS product is subjected to hundreds of tests. Part of this is done with custom dCS computers that test every interface at all supported sample rates and produce a small library's worth of results. dCS stores about 300 MB worth of documentation containing every minute detail for all components before they are shipped out the door. In the rare circumstance product is returned due to a problem engineers can quickly compare its current test results with those of the product before it left the dCS factory.
Adjacent to the auto-testing computers at dCS is the repair department. My tour guide Martin proudly noted the lack of current generation products in for repair. I believe I only saw three or four products on the repair rack and each product appeared fairly dated. Based on my experience with dCS components this was not surprising.
dCS, DSD, USB, Etc…
The day after visiting dCS in Cambridge I traveled to Amsterdam for the press event in which dCS debuted its DSD over USB implementation. The event was held at Harry van Dalen's Rhapsody Sound & Vision. Readers fortunate enough to live near Amsterdam should definitely visit Harry, Michael Poland, and the rest of the Rhapsody staff. These guys know good sound and made my visit very enjoyable. The hospitality of Benelux importer Martin Odijk from More Music was also extremely gracious. There's nothing better than spending time with great people who also enjoy great music. A fun time was had by all.
I hope everyone is hungry for alphabet soup. Let's start by covering some acronyms.
- dCS -> Data Conversion Systems
- DSD -> Direct Stream Digital
- USB -> Universal Serial Bus
David Steven Jr. and Andy McHarg were on hand to present dCS' DSD over USB solution. Above all both gentlemen continually said the goal of dCS is to develop an open standard that any playback software developer or DAC manufacturer can implement. dCS is interested in furthering DSD playback in the industry rather than simply its own components as this is the only way for a format to survive. According to dCS there were a number of technical challenges to implementing DSD over USB in the most open way possible.
- There are no open interfaces for allowing the streaming of DSD source material.
- No USB class supporting DSD currently exists.
- No operating system support for playing back DSD exists.
- As DSD is fundamentally different from PCM, utilizing an existing interface has dangers if the wrong mode can be selected by the user (for example this could result in high?level wideband noise ).
- Any solution needs to ensure easy implementation for playback software developers.
- Any solution should ensure there is no need for special drivers (other than USB audio class 2).
- Most importantly the solution must ensure that playback quality is of the highest standard.
The dCS solution, and proposed open standard, to these challenges involves an ingenious method of packing pure DSD data so that any DAC capable of accepting 24/176.4 data can be modified to accept this pure DSD data. So everyone is clear I will reiterate this method does not involve converting DSD to PCM data. Rather this proposed open standard is a method of streaming pure DSD data directly from the computer to the DAC. One more time, there is no DSD to PCM conversion. Period.
In the words of dCS:
"Working together with Amarra (Sonic Studio) and Pure Music (Channel D), the dCS team specified a method of packaging the DSD data so that any DAC with 24/176.4 capability could in principle be modified to accept it.
The solution is based around the fact that DSD has a sample rate 64 times that of CD, but each sample is 16 times smaller, giving a data rate of 4 times CD, which may be packed into 16/176.4. This means that we can utilize a "standard" PCM sample rate of 176.4k, and if each sample is 24 bit wide, then we can utilize 16 bits for the raw DSD, and another 8 bits to allow the receiving DAC to establish that it is receiving DSD rather than PCM, as follows: (Graphic omitted but available in this PDF document)
As you can see, the top 8 bits of each sample are 0xAA. This represents a negative DC offset, which won't be seen in real?life audio, so we can use it as an indicator for the DAC that the data it is receiving is DSD rather than 24/176.4 PCM. Additionally, it means that if we try and play the DSD stream via a DAC that doesn't support it, the output is noise, but it is 48dB down ? the user can tell they are playing back an unsupported format, but not in a speaker?destroying way. For PC?audio based solutions, if the playback software is DSD?aware, it can output 0xAAAAAA samples by default to represent a safe DSD mute.
The first dCS product to be modified to accept DSD over USB is the Debussy DAC. To modify the DAC, we leveraged the inherent flexibility in the dCS Signal Processing architecture, by modifying the FPGA that deals with signal processing to look out for the special indicator samples. When these are detected, the samples are unpacked into native DSD, and fed into the DSD processing block that Debussy shares with the other DSD?capable dCS DACs."
After the technical presentation at Rhapsody were listened to pure DSD playback from a MacBook Pro running Channel D's Pure Music. The laptop was connected to the dCS Debussy DAC via AudioQuest Diamond USB cable. The DSD music used for demonstrations was provided by Turtle Records' Bert van der Wolf who was on hand to provide background information about each recording. The rest of the audio system comprised a Spectral DMC-30SS Series 2 pre amp, DMA-260 power amp, dCS U-Clock, MIT cabling, and Avalon Acoustics loudspeakers. It was very neat to see iTunes displaying DSD data and listing the sample rate as 2.8224 MHz. Pure Music effortlessly sent this pure DSD data to the Debussy producing wonderful sound quality. One track in particular contained a solo drummer and produced astounding dynamic range. The drums sounded like they were literally being played in the room. The overall sound quality I heard from the remaining DSD tracks was on par with Bill Schnee's 24/192 live to two-track recordings for Bravura Records.
dCS plans to make the DSD upgrade available to all components containing USB inputs. As dCS uses an advanced logic board inside all its products this DSD capability can be enabled through a simple software upgrade in the field.
At the conclusion of the press event at Rhapsody HiFi+ Editor Alan Sircom, The Audio Beat's Chris Thomas, and I headed to the mastering lab of Turtle Records' Bert van der Wolf. Bert has a full dCS, Spectral Audio, and Avalon Acoustics multi-channel system for working on DSD recordings. The first two or three pure DSD tracks I heard in 5.1 surround really threw me for a loop. I didn't like the sound at first as it wasn't what I'm familiar with in my own room or any room for that matter. Once I acclimated to high quality multi-channel sound I was addicted. When done right multi-channel audio can deliver an immersive experience like no two channel recording I've ever heard. Whether I focussed on the front three channels or the front right and rear right channels I could sense where I was in the acoustic space of the recording. Instead of hearing the typical reflections of a listening room I heard the reflections of a concert hall and a small club where some of the demo tracks were recorded. If high level multi-channel playback was more practical (think space and money) I'd have it in my listening room in a heartbeat.
I'd like to thank Martin and the More Music team, Harry, Michael and the Rhapsody team, Turtle Records' Bert van der Wolf, and the dedicated Computer Audiophile readers who came to the computer audio seminars for making my stay in Holland so pleasant. I can't wait to return and see my new friends once again.