Peter Qvortrup On Digital
Jack: Peter, please share with us your opinion on the basic differences between CD and vinyl sound.
Peter: My view is that digital cannot resemble the original because of the inaccuracies introduced at the point of entry in the digital domain. The errors and omissions introduced by all current and past methods of conversion are so great that, at best, all we get is a card board copy of the original and it is clearly audible, as a quick comparison between a decent turntable and even the most expensive CD replay set up will quickly reveal.
This is because much like the flawed assumptions used by mathematicians to create the financial “innovations” which lie at the heart of the current financial crisis, the fundamental assumptions that created the way we digitize the analogue signal also fail to describe the content of dynamic wide band signals and the way they flow. The end result is a set of mathematical formulas that are sadly lacking in their ability to model the full range of variables in a music signal, and as a result when we come to make the analogue to digital conversion process, the bar is set at a level which is lower than it needs to be, resulting in conversion technology which does not do justice to the analogue signal it is being presented with, resulting in an anemic digital version of the analogue original.
It is therefore no major surprise that the digital medium, as we know it, lacks authority, authenticity, immediacy, instrumental medium and density, dimensionality, and overall presence when compared to its analogue version, to the point where even some digital recordings sound better on LP than the CD (the reverse is of course also occasionally true, but for sake of a proper comparison, we should always compare early all-analogue recordings on LP with their CD counterparts from the early to mid 1980s, my experience has always been that AAD recordings generally sound better than ADD, and ADD generally sounds better than DDD), which leads me to believe that the signal damage goes beyond merely the digital conversion process itself, as it would appear that the longer the signal stays in the digital domain the more damage it suffers, which may also explain some aspect of what I hear in music servers.
The high “resolution” formats like SACD and DVD-A are no better, just different versions of the same problems that beset Redbook, but unlike standard Redbook, the SACD/DVD-A converters are virtually impossible to improve upon as the 1Bit system they use do not allow removing the oversampling and digital filtering, for example, so we are reduced to playing with component choices and power supply configurations, which is like putting lipstick on a bulldog, to use a currently popular phrase!
Jack: Then Peter, how did you come to build some of the most musical sounding CD players, DACs, and Transports?
Peter: Well, from the onset I disliked CD with a vengeance, but over some years I was increasingly faced with the opposing choices or dilemma if you like, of preferring the quality of the purely analogue source but also a great need to be able to hear a lot of the music I liked that only came out on CD. So, I had no real choice but to try to work out how one makes a digital-analogue converter that does not completely sucks the life out of any musician’s best efforts, including when the recording engineers have also not done their greatest job, which of course many LPs also suffer under, but which to me at least manifests itself as far worse when it is on CD than LP.
I always felt that the main improvements were available in the DAC, so that is where our main efforts have been concentrated. However, recent development work has shown that the CD transport has equally great potential for improvement, so considerable efforts are now given to find out how far the CD transport can be taken: We now employ a valve power supply in our best CD transport for part of the circuit!
Perhaps a little history?
In 1991, I set about with my engineer at the time, Guy Adams (of Voyd Turntables fame!) to develop a digital-to-analogue converter which has a less digital signature. This work led to the use of a transformer as the I/V interface, a practice we patented across the world, and the DAC3 was born in 1993. I have always been doubtful about techniques that purport to improve the signal, so sometime in 1994 I asked Guy whether it was possible to remove the oversampling and digital filters from the converter. Guy felt it was not feasible after speaking to various people at Burr Brown, so in late 1995 I mentioned it to Andy Grove, and Andy said he would look at it, a prototype was made a week later and the sound was a revelation, although it needed some fine tuning to get rid of the out of band interference.
We released the first non-oversampling (we call it 1xoversampling) design as the DAC5 Special in July 1997, and the rest is really history. We have spent the years since then refining the circuit, I/V interfaces and power supplies.
We added dedicated 1xoversampled CD players in 1998 and started work on CD transports in 1997, the first one being the CDT Two, unfortunately the Korean manufacturers who were building all the CD products for us went under in 2000, and it took us another 4 – 5 years to develop a replacement for the CDT Two. The CDT Two/II was released in late 2004.
Jack: One last question Peter, what are Audio Note’s future plans for digital playback?
Peter: We continue to refine the circuitry and power supplies and we keep discovering incremental improvements to add to the circuits, components and materials. As an ultimate statement, Andy has been working on two major projects, the first is an all discrete 20 – 24 bit converter, and the second, which is more interesting, is a completely new conversion system which we originally designed as an out-of-house project for a company that does investigations into molecular resonance in materials; initial experiments look promising, so we shall see.
On the CD transport side, we are finishing development of the CDT Five, where a couple of valve in part of the power supply has shown that even here, valves are superior to semiconductors.
I think that covers it!