Esoteric P-03 Universal Disc Transport and D-03 Digital-to-Analog Converter

A Transparent Window On The Source

One of the great mysteries of digital audio (to me, at least) is how CD transport-mechanism quality affects the sound. I’m not talking about the differences between transports as a whole, but of the mechanism that spins the disc and reads the data.

The sonic differences between transports are the result of jitter, or timing variations, in their digital outputs; this is now a welldocumented phenomenon.

But here’s the conundrum: Every CD transport mechanism recovers the same ones and zeros from the disc, whether that mechanism is a flimsy plastic job found in a $39 player or the massive VRDS-Neo mechanism in the $17,200 Esoteric P-03 transport reviewed here. You can prove this by conducting a bitfor- bit comparison between the audio data recovered from either player. So what has changed?

True, those audio data are only part of a very complex datastream that comes off the disc. They undergo significant processing to extract the audio information.

Nevertheless, one would think that processing and clocking that data through integrated circuits (and sometimes buffers) would remove any timing errors (jitter). And if that were the case, then why not recover the data with a cheap mechanism and employ a high-precision clock to correct for the mechanism’s timing inaccuracy?

There are two more pieces to this mystery. Nakamichi’s 1000 CD transport, which has an acoustic seal on the door to its slot-loading mechanism, sounds better with the door closed. Apparently, acoustic energy from the loudspeakers impinging on the disc and transport results in a slight, but audible, degradation of the sound. Once again, the bits are the same, door open or closed.

The third piece of this mystery has baffled me for nearly 20 years. I was working in a CD mastering lab (where we transferred CD mastertapes to disc on a million-dollar laser mastering machine in a clean room), and part of my job involved trouble-shooting odd technical problems relating to mastertapes and replicated discs. A client for whom we had made discs was unhappy with the results, reporting that the discs didn’t sound as good as the mastertape. I compared the data on the mastertape with the data recovered from the replicated disc (using a CD-ROM pre-mastering system) and found, not surprisingly, that the disc and tape were bit-for-bit identical. The extremely talented electrical and optical engineers I worked with (who had designed and built the mastering machine) dismissed the artist’s claim, saying: “Bits is bits.” Unfortunately, I was unable to compare the sound of the disc with the mastertape through the same digital-to-analog converters (this was back in the day when mastertapes were on ¾" U-Matic tape decoded by a Sony PCM 1610 or 1630).

Partly out of my own curiosity and partly out of the desire to please the client, we cut another master disc and replicated new discs—this time on a different mastering machine. The client reported that the new disc sounded as he intended, and went away happy. But I was left with the question of how two CDs, each containing identical data, could sound different. But now I was armed with two discs that could be played back on a high-resolution system, and my own listening confirmed that the second disc did sound better than the first. (A similar paradox, which arose many years later, is that a CD-R made from a CD often sounds better than the original CD.)

The next step was to look at the physical differences between the two discs—discs with identical ones and zeros but with different sound. An analysis of the pit and land lengths on the two discs showed that the inferior-sounding disc had greater variations in those pit and land lengths—in other words, jitter was encoded in the disc’s physical structures. (Specifically, a histogram of the frequency variance in the discs’ nine discrete pit and land lengths showed the inferior-sounding disc had a wider bell curve than the better-sounding disc.)

The question remains: How do timing variations in the raw bitstream recovered from a CD make their way into the analog output signal? If the sound is different, then the signals must be different. Doesn’t precise clocking eliminate transport-induced jitter? Would better clocks have removed the sonic differences I heard between the two replicated CDs? And how did sound impinging on a disc played in the Nakamichi transport affect a change in the analog output signal?

These puzzles lead to the question that opened this review: Does a massive and elaborate machined-metal transport mechanism in the Esoteric P-03 Universal Disc Transport sound better than a cheap plastic job? The P-03 is the ultimate expression of transport quality. This 71-pound, $17,200 device takes the task of recovering data from a CD quite seriously, employing what is unquestionably the best built and most elaborate mechanism yet devised for CD playback (save for the mechanism in Esoteric’s $25k P-01 transport).