Naim CD 3.5 CD player It's About Time!
I asked Naim's Julian Vereker to describe his digital philosophy. The following was completely off the cuff, but on a subject he has obviously thought deeply about.—Wes Phillips
"In the analog domain, information is stored in terms of amplitude—either voltage or level of magnetization, or the mechanical movement of a stylus in vinyl. In the digital domain, it's stored in terms of time—that is, when something becomes a one or a zero. Therefore the time element is absolutely crucial when you come to convert the digital stream back into analog. Any noise in the digital datastream can—and does—cause non-signal-related distortion. Even at very low levels, this is extremely disturbing.
"We were amazed at how quiet the power supply has to be in order to get around those sorts of problems. And, of course, the board layout becomes extremely important. In terms of noise within the system, digital noise is a big factor—digital outputs on the modern chipsets can be turned on or off, and having it off reduces jitter noise by a large amount. People say, "Oh, it's only a little wire out," but the effects of putting that tiny wire onto the digital out and turning it on destroys the jitter performance by at least an order of magnitude.
"When we started thinking about making CD players, before we made our first one back in 1989, no one had mentioned jitter yet, and it was only afterwards that people started calling it that and trying to measure it. We knew you couldn't measure it directly because the minute you put a probe on, it starts broadcasting the noise around the circuit and other parts pick it up. This makes it a bit like the Heisenberg Uncertainty Principle—the minute you start looking at it, it's not the same as it was before you began looking at it.
"Eventually, when Paul Miller came along with his test set, he was flabbergasted to find that our players measured about four times better than anything he'd measured before. "How did you measure this?" he asked, and we told him we hadn't actually measured it, we just looked at it sideways. By the way, I think that Miller's program is absolutely amazing—he wrote all that code himself. It's really quite good."—Julian Vereker