This review page is supported in part by the sponsor whose ad is displayed above
Source: Zanden Audio Model 2000P/5000S
Preamp: Wavac PR-T1
Amp: AudioValve Baldur 300 monoblocks
Speakers: Kharma Midi Exquisite, Kharma Ceramique Subwoofer
Cables: Argento Serenity Master Reference Extreme Edition digital XLR, Acrolink DA6100 digital RCA, Argento Serenity Master Reference (SMR) Special Edition between DAC and the Pre, Argento, Serenity Master Reference Extreme edition between Pre and the Amp, SMR Extreme Edition speaker cable and Serenity SE between the output of the Wavac PR-T1 and the subwoofer; SMR-H power cord x 2 on the amps and all other power cords by ORB of Japan (model HC-150ACW)
Stand: Finite Elemente Master Reference Rack (stock platforms upgraded to custom-made wooden platforms from Acoustics International)
Power line conditioning: HT-4G AC passive strip for the source components via SMR power cord as the AC wall connection; power cords on preamp and amp connect directly to wall
Sundry accessories: 2 x super-cyro technology wall sockets from Japan
Room size: 11' w x 15' d x 10' h
Review Component Retail: CHF 42,500 [Hong Kong price]
Modesty is a form of virtue. This is my true feeling towards Thierry Heeb, Chief Executive Officer of Anagram Technologies [below], after I interviewed him on the high-profile project of the Heritage DAC under the Orpheus Laboratories banner, a former subsidiary of Anagram now run as an independent company. An expert in the field of applied mathematics (algorithms) who previously worked as a crypto-analyst in the Swiss Department of Defense before joining Goldmund, he subsequently cofounded Orpheus Laboratories with Florian Cossy (another expert in analog electronics who designed the complete hardware package for the Heritage converter). Thierry is an inventor with 4 patents pending approval, one on sample rate conversion, two on digital amplifiers and one on system synchronization. To my surprise, he told me that a successful DAC does not solely rely upon on the digital section, no matter how fancy the digital theory behind it is. The other 3 pillars -- the clocking scheme, the analog section and the power supplies -- are equally as important. The challenge of the Heritage Project was to apply equality to the execution of these four pillars. Even when the four integral parts are considered technically perfect from the engineering/measurement perspective, their final interface in the final development stage still has to undergo stringent tests of human ears in numerous sessions with real musicians.
This preview focuses primarily on the background development of this DAC, the design philosophy and the technologies embodied. The actual review will most likely follow about one month from now. Along the line, I will gather listening notes from four other Heritage owners (two of them who have been using the Orpheus One SE DAC for 6 months already.) The Heritage is an asynchronously upsampling DAC of a type invented and implemented by Anagram Technology. Before going into the details of the asynchronous approach, let's have some basic ideas on traditional methodology first.
The standard approach
In the standard approach, the transport reads data from the CD and then modulates clock and data into a single data stream (bi-phase modulated) before sending the S/PDIF signal to the converter which then locks to the incoming data stream by receiver. A PLL inside the receiver (phase locked loop, a very delicate electronic circuit) recovers the clock signal and separates it from the data to regenerate the clock signal before sending it to the D/A conversion chips. The amount of jitter will depend heavily on the accuracy of the recovered clock signal because it is used to synchronize (clock) the D/A chip(s).
The asynchronous approach
In Anagram's asynchronous clocking approach, the recovered clock signal and data after arriving at the DAC's receiver end are directly transferred to three 800MMAC DSPs which upsample the incoming data stream to 24bit/768kHz (the standard offering by Anagram Technologies to OEM parties is 24/384). This asynchronous upsampling technology Anagram has exclusively licensed to Orpheus Laboratories, encompassing a "5th generation sample rate converter algorithm" and "data-to-system synchronization (DSS) architecture". These are advanced features that go beyond my ability to comprehend thoroughly but I will try to explain the uniqueness of the 5th generation algorithm inside the Anagram/Orpheus digital signal processor.
Most other upsampling algorithms yield an output which is "n" times faster than the input, but n is always an integer (typically 4, 8, 16, 32 or 64) of that input. In the Heritage, "n" is not necessarily an integer but can be any rate depending on the input signal. It ensures that for every input sampling rate, the output after the upsampling process is completely independent of the input sampling rate.
With respect to the DSS architecture, I was told that the merit of such technology is to ensure that all processors inside the Heritage operate at a synchronous rate with the audio data stream. All switching noise induced by the processors is therefore synchronous with the new sampling rate after the upsampling conversion. As a result, "asynchronous switching noise", which usually contaminates analog signals after the D/A conversion, won't be present.
b. Clocking stage
After the asynchronous upsampling process, the original clock recovered from the transport is completely decoupled from the interpolated data stream. A quartz master clock with sub pico-second accuracy is physically located next to the DA conversion chips to resynchronize the new data stream in accord with its own clocking rate (this clock is referred to as the time-lock module). The DSP that performs the upsampling thus acts as a clock isolator between the clock recovered from the incoming S/PDIF stream and the local master clock. The clock used for the D /A conversion process no longer is the clock recovered by the PLL inside the receiver. The local master clock assumes full control and is used to synchronize with the DAC chips. Data going into the four Delta-Sigma type DACs in true parallel balanced configuration are then internally upsampled to 6.144MHz before the actual D/A conversion process takes place after the upsampling and reclocking processes.
c. The D/A conversion
Inside the D/A chips, the 768kHz data stream is split into two 384kHz halves. A random bias signal for each chip is added. The random signal for each D/A chip is different but their sum is zero. At the output of the converters (4 per channel, true dual differential), the positive phase of the twin converters is added to the negative phase of the opposing pair. The first group of converters receives the audio signal plus random noise and the second group of converters receives the inverted audio signal plus random noise. Once added, the random noise cancels out. The same is done with the negative current output of the first group and the positive output of the second group. This results in a total of eight current outputs per channel to further reduce quantification noises.
The importance of clocking
Thierry regards other school of thoughts on clocking schemes as interesting too but does not view them as the most effective solutions for minimized jitter. To him, the presence of an external clock source may introduce more problems since clock signals have to travel through cables which are vulnerable to RF and EMI contamination. Moreover, adjustable features such as a VXCO (voltage controlled oscillator) or OCXO (oven controlled oscillator) are usually required with external clocks which are inherently vulnerable to jitter.
A better alternative is to slave the DAC to the clock provided by the transport. However, it is then usually the 'word' clock transmitted since the master clock signal is usually in excess of 10MHz. Such ultra-high frequency signal is even more vulnerable and sensitive to external noise interferences. Transmitting word clock information also has to undergo a PLL stage to recreate a master clock in the DAC by multiplying the incoming word clock signals. Even very good PLL circuits have difficulty in eliminating low-frequency jitter.
The analog section
The analog stage of the Heritage was designed primarily around the principles of high bandwidth and simplicity. This stage is fully balanced. It utilizes two I/V and two gain stages per channel (one for each of the balanced halves).
a. Current to voltage stage
The I/V stage is build differentially using AD797 ultra-low noise op-amps and the highest grade Vishay resistors as well as ultra-low ESR Leclanché capacitors. The I/V stage is usually where most of the audio signal degradation occurs. It is thus critical to design ultra-wide bandwidth I/V converters that do not introduce time domain distortions from insufficient speed. The Heritage I/V stage hence always operates under very high current conditions and is biased in pure class A. Operating at high current makes this stage less sensitive to low-level noise. S/N ratio is elevated in tandem with high currents (tens of milliamperes since we're dealing with low-level signals).
b. Gain stage
The I/V conversion stage employs individual discrete regulators for each power supply. These regulators are extremely fast, purportedly much faster than the best off-the-shelf integrated regulators on the market. After the I/V stage, the audio modulation is directed to the gain stage. Thierry prefers to call that an impedance adaptation stage because the audio modulation reaching it post I/V is already at full swing. These impedance adaptation stages offer high input impedance and low output impedance to minimize sensitivity to the final load (the preamplifier input).
The final output is summed at the XLR output to provide a differential signal of the highest quality. The Heritage DAC is optimized for balanced operation and the XLR socket becomes the preferred connection.
The power supply section
Even the power supply has been optimized, using a novel type of discrete regulators which remain very quiet and fast during balanced operation. Each output stage runs off its own dedicated regulators to provide ultra-low noise DC to the analog circuitry. The Heritage uses a dual-mono power supply with twin 180VA toroidal transformers. The reservoir of energy allows the power supply to provide enough current on demand. The digital section has its own regulated supply as does the clock. PCB quality is superlative using gold-plated traces and dedicated rails for various power supplies and ground paths.
The final dual chassis was developed by acclaimed industrial designers and is built by specialized subcontractors in the Swiss Watch Valley to the highest mechanical quality standards. Even the footers underneath the casings are a completely new design and custom fabricated.
The role of the human ear
Originally, the intention was to not include an analog filter for the Heritage but numerous listening tests with golden-eared Mr. Philippe Montavon and orchestral musicians convinced the design team to include an analog low-pass filter in the analog stage. The absence of an analog low-pass filter could introduce high-frequency noise from the delta-sigma converter architecture. The listening panel's final verdict was to include an analog filter with a very gentle slope.
Initial listening impressions
The technological achievements of the Heritage DAC require substantially burn-in time alongside fine-tuning before meaningful observations and opinions can be drawn in an objective manner. One of my most valuable and knowledgeable high-end audio comrades is J. Lam, who initially tested the Heritage with me for three days when partnered with my Zanden 2000p transport. I think it is fruitful to share his words since he had chosen the live recording Get a Life by renowned Hong Kong male vocalist Eason Chan. This is a recording he knows inside out because (1), he had recently discussed its recording quality with Eason Chan himself after an interview on Metro Radio; (2) he attended the concert that was recorded; (3) he also talked to the recording engineers of Cinepoly in charge of this album
In the words of Jlam: "Eason Chan is my favorite male Hong Kong singer not only because of his personal charm but for his consummate vocal skills. He studied at the prominent Vocal Faculty of the Royal Academy of Music in the UK. He applied many techniques learnt there to his songs where most pop singers have not undergone equivalent formal training. It probably explains why the quality of his voice during the final two concerts maintained the very high standards of his very first sessions. His appropriate use of vibrato enhances certain lyrical lines from time to time with varying degree of intensity through the most delicate variations in his singing tone; transient emphasis on vowels and consonants in some lyrical lines; as well as masterful breath control to recharge energy in the midst of high passages. All of these subtle variations convey a very emotional experience. The Heritage reveals all such subtleties even to the extent of seeing the microphones temporarily shaded by his moving hands, incurring subtle alterations of depth perspective. Yet the vocals never decouple from the big picture, the surrounding, emotionally highly charged audience who collectively acts as multiple layers of background chorales.
"The dimensionality of this recording is phenomenal. Recording engineers at Cinepoly used as many tracks as possible to track different sounds at different locations inside the coliseum. Numerous microphones were used in audience section seats. Other microphones were located at the very front near the stage, others again placed to the far outer edges of the auditorium. At the mixing stage, the production team spent thousands of hours to mix these pre-recorded tracks and recreate as complete an immersive experience from two channels as possible.
|"The Heritage DAC has immense spatial resolving power. The distance perspective is conveyed not only between Eason Chan and the applauding audience sections sitting in quadrants surrounding the stage, but also the virtual space of the listeners themselves with varying depth perspectives. Note that the experience is not that of the fixation of images in certain virtual spaces but moving images interacting with one and another. It must be the case because so many microphones were located everywhere inside the coliseum. Through the Heritage, I can feel the collective sentiment of the audience when they sang along with Eason during the last few tracks of disc 3 in Get a Life. The achievement of such an enveloping sound field is not achieved through inflation of air but rather by the accuracy of retrieving numerous very small building blocks of the entire soundstage in the x, y and z axes. In certain songs where the background chorus requires a certain degree of energy to reach specific key notes, I can immediately lock onto the corresponding charge of energy by the listeners, trying to somewhat breathlessly catch up with the singer."
|Concluding introductory remarks
The very last question I asked Thierry during the interview was about what makes him proudest of the Heritage DAC. In his exact words, "... the whole project went very quietly. We neither wanted to indulge any marketing hype nor make any superlative statements about our proprietary technology. We just wanted to see how far D/A technology can go with our vision of digital theory applied from the perspective of live music. The Heritage project allowed us to make a DAC purely for musical enjoyment. It is the transformation of digital audio theory into emotional experience. We think that the asynchronous approach and our clocking scheme are especially unique. The project enhances our understanding of the importance of clocking schemes for D/A conversion. It was extremely rewarding to bring Florian's and my concepts to a level where development cost was not a constraint. Also, Florian's analog expertise was paramount to this project. Even if you can deliver the purest digital signal, it's not worth much without a corresponding ultra-high quality analog interface and power supply. I would also like to thank Mr. Philippe Montavon who chaired the listening panel with other fine musicians during the entire course of the development."
This is the first time that this writer has devoted so much introductory space to the technical aspects of a review component. If I have misunderstood certain technicalities, I hope fellow readers will forgive my innocence on the subject matter. I have learnt a lot from Thierry and his Heritage team not only on the technical aspects at the very forefront of digital technology but also his spirit, vision and determination to fulfill his dream despite the rocky road of research and development. The actual achievement, however, does not invite modesty though it remains, in my opinion, a virtue that is very much reflected in the particular outlook of the Heritage. It deserves, at the very least, a silent ovation for a very competent team that's deliberately shunned the lime light for this product launch.