I will share some stories with you of our history over the last 50 years – half a century of making audio equipment. There are a lot of stories, some of which I cover in my memoir 99% True, while others have gone untold. Still others are lost in the mists of time.
One such story involves the creation of the PS Audio Digital Lens.
A decade or so after the introduction of the CD player, we audio manufacturers were totally immersed in the challenge of extracting all we could from this new medium. At that time, vinyl still trounced digital (and not by just a little). CD reproduction through an external DAC was a fledgling category just out of the nest and fending for itself. Even the finest of DACs in that day were hard, bright, and somewhat two-dimensional sounding.
There was a lot to learn back then.
At that particular junction in time – the mid 1990s – I had sold PS Audio and had moved my family from California to Colorado to join Infinity founder Arnie Nudell in designing and building loudspeakers and the electronics needed to make them sing. The name of our company was Genesis Technologies.
This deviation of mine, from a pure electronics company like PS Audio, to a loudspeaker/electronics company like Genesis, left me twiddling my proverbial thumbs (not something I can tolerate for long). I had already designed all our servo electronics and power amplifiers and was working on an entirely new idea in amplification that would one day become the Stealth Amplifier (featuring user-adjustable Class A bias from 1 percent to 100 percent and a novel MOSFET pure-current amplification output stage that has still not been duplicated to this day). I was itching to get back into the digital fray.
During that period, my fellow designers and manufacturers in the high-end were expending all their efforts at raising digital performance levels by focusing on the all-important analog output stage. During that era, we saw everything from big vacuum-tube outputs to high-speed ultralinear analog whiz bangs. A few companies were experimenting with different DAC architectures, like the dCS Ring DAC (around 1989), and the PS Audio Ultra Analog entry (around 1993), but for the most part, the bulk of companies were basing their designs on off-the-shelf DAC chips from Texas Instruments and others.
(All these DACs were still what we refer to today as R2R DACs that were limited to 20 bits or so – like the Ultra Analog version from PS Audio – the exception being the 24-bit "sort of" performance that dCS achieved.)
I, on the other hand, was headed off in a very different direction.
At that early stage in the technology's development, no one was jumping up and down about superior sound quality to analog. In fact, any digital product even approaching a pale semblance to analog was hailed as a miracle wrought by wizards.
There weren't many miracles or wizards.
While others were focused on perfecting the output stage of the D-to-A converter, I became enamored with the idea of starting at the very beginning. How much did the bits themselves matter? The popular notion of the time was that one of the beauties of digital was that bits are bits. As long as you got the 1s and the 0s in the right place, not much else would matter.
Oh, the hubris.
One of the clues that got me headed down the "quality of bits might matter" rabbit hole was a curious CD transport we owned at Genesis. Our good friend and fellow audiophile crazy, a New York-based importer of audio equipment by the name of Victor Goldstein (now sadly deceased), had been haranguing Arnie and me about the wonders of a very pricey French transport from a company we knew primarily for its vacuum tube products. Jadis.
The Jadis JD-1 transport was a beautiful beast – a sculpted metal temple of gold and black that housed a decent top-loading transport from Philips. Victor was so excited about this new transport that he just shipped us one on loan, certain we would buy it as our new reference in the Genesis listening room. He was not wrong.
Incredibly (remember, this was back in the early ’90s when only DACs mattered to sound quality), CDs played on this transport were remarkably better than those same CDs played on any other transport we had on hand. And not by just a little bit.
The Jadis JD-1 transport. Courtesy of SkyFi Audio.
WTF?
Same bits. Completely different performance. How was that possible? What sorcery was at play here?
I lay awake at night going over the possibilities and always, I came back to the same conclusion. Something in those bits had to be different (a better conclusion than that the world was turning upside down).
One Saturday morning I left the house early and headed into work determined to discover what could be different. After hours on the bench armed with nothing more than my trusty capture scope to grab hold of those "identical" bits I began to notice something unusual.
I noted previously that I had been driven crazy by the remarkable performance improvements of a product that made no sense: a CD transport producing the bits feeding a DAC.
What in the world? Bits are bits. Every engineer and scientist willing to offer their opinion agreed. The beauty of digital is its ability to be endlessly copied without degradation – transported around the world over every means possible, from telephone modems, satellite relays, and underwater cables to a newfangled optical interface. Bits are bits. Get them right and without drop outs or confusion, and the only differences possible would be in the conversion process of producing analog.
Only, a new transport we had acquired was turning our world upside down. Those identical bits sounded remarkably different when played through the Jadis JD-1 transport than through any other medium we had.
And to make things even weirder, it didn't matter where we had obtained those bits. Commercial CDs as well as CD-R copies of those same CDs Arnie and I had ripped and then burned all sounded remarkably better.
In my investigation, armed with little more than a fancy capture scope, I narrowed it down to the only difference I could see. A difference that made zero sense whatsoever – in fact, so little sense that I continually rejected the observation as mere noise and clutter. Couldn't be.
That feeling you get when something you see on a 'scope that just doesn't add up. Courtesy of Pexels.com/cottonbro.
When you're faced with the impossible, and it is the only option available to you, it's a good idea to let down your guard and accept it as the best possibility. Run with it to see where it takes you.
The output digital signal from the Jadis had two major differences from any other digital output I could see: amplitude and shape. Instead of the industry-standard output level of 1.2 volts (or so), the Jadis output was a whopping 4 volts. On top of that, the output square wave shape was different: smooth and without any visible over or undershoot. It looked to me like a beautiful square wave that any analog or tube designer would be happy with if their analog circuit had output it.
But this was digital bits – bits that would never see an analog circuit – bits that were used only to trigger an input flip-flop on the DAC. This could not matter!
WTF? I had to know.
It didn't take more than a day or two on the breadboard to build a little op-amp circuit that would take in a standard output from a transport and amplify it up to the 4 volts to match the Jadis. This circuit got slapped onto the output of our previous CD reference transport and I pressed play. Wow.
Did it have the magic of the Jadis? Yes and no. I would give it an 80 percent. (Over the next few months, I experimented with the wave shape, using a softer JFET input stage op-amp to round off the corners of my larger output square wave, and this filled in the remaining 20 percent of the mystery for that transport.)
The output size and shape mattered. What came next was even more surprising.
I started to form an idea. What about making a separate piece of equipment that would function as a digital in and digital out? It would take the wildly variable (and cheesy) master clock signal from the CD, put it into an intelligent buffer (memory), and output it through a fixed high precision low-jitter clock into a completely separate output stage with its own power supply and wave-shaping circuit, to provide a 4-volt beautiful square wave, without any jitter.
Think of this like a large water tank feeding a city. The tank gets filled over days and, because it is a big buffer, it can output an absolutely perfect water stream of constant pressure to as many or as few people as need its water.
I presented my brainstorm product to Arnie Nudell, who got all excited and said, "In essence, we'll clean up the signal, get rid of any timing and jitter errors, and refocus that data into a perfect output."
"Exactly.” I said. Any idea what we could call it?"
Arnie, the nuclear physicist turned speaker designer said, "obviously, it'll have to be the Digital Lens. As in optics, it focuses a blurry image into a sharp image."
Bingo. We now had a concept, a path and a reason for the path, expected results, and a cool new name.
Heck, all we needed now was…something that worked.
The original Genesis Technologies Digital Lens.
To be continued…
Header image: Paul McGowan with the Infinity IRS V loudspeakers at PS Audio.
0 comments