If you discount the fact that digital transports can sound different—it matters not whether they stream digital files off the cloud, retrieve them magnetically from a hard disk or optically off a CD—you've not paid sufficient attention. As the first reviewer of today's Lektor wrote, "from the first seconds of the piano intro to Autumn In Seattle, a Tsuyoshi Yamamoto song written for Winston Ma for the eponymous album, it was immediately clear that the Ancient Audio transport presented the music differently than my Ayon transport." As for files vs physical media, "I compared the sound of CD over the Lektor and the same recordings on file both as CD quality and higher resolution. Extending word depth to 24 bits as well as sample rates beyond 44.1/48kHz improved the files over their Redbook versions. But it was also true that a CD played back over this CD transport sounded better than any file except maybe for DSD." [At right, a Lektor from Jarek's Granite Age.]

But there's more. Even control software makes a difference. Compare streaming through iTunes or Windows Media Player to Audirvana, HQplayer, PureMusic, Roon & Co. Though everything remains in the digital domain, software controlling signal routing has an effect. Ditto up/resampling algorithms and digital volume algorithms like Leedh. As Jarek explained, this extends to firmware controlling a disc-spinning transport servo. Somewhere during the digital Stone Age, CD was perfect sound forever. Jitter, stamper quality, error correction, brickwall filters, laser tracking servos and the lot didn't matter. 1s were 1s, 0s were 0s. Life was simple.

As any digital audio engineer tells us today, the reality is very different. Everything matters. How much is simply for our ears and wallet to decide. Like John Darko puts it, there a day'n'night differences like the sound of his new Lisbon room pre/post professional Vicoustics room treatment. And then there are 'audiophile' differences which to his YouTube viewers may or may not matter much if at all. As overall system resolution goes up and the owner's hearing keens, previously obscured or super-marginal differences can become more significant. What value we attach to them differs from person to person. Even in that context speaker differences are greater than amplifier differences which are greater than DAC differences which are greater than… and so forth. On that totem pole digital transport differences sit rather low. But once everything else is shined up just so, they can still be decisive enough to trigger a clear preference. I think that painted a fair and even-handed intro for today's subject. Filling in comparative details are a discontinued Denafrips Avatar top-loading transport with Philips CDM4/19 drive…

… and this superceded Simon Audio Lab AIO used as top-loading Sony/Sanyo transport with coaxial output. Of the three, only the Denafrips can be set to 352.8kHz pre-DAC upsampling. None of them use a slot or drawer type to require that we first place a stabilizer puck atop the CD spindle before hitting 'play'. Only the Lektor goes nude. That makes leaving a CD on the only dust cover. And following the everything-matters mantra, manufacturers of top-loading transports have different views on the ideal size, material and weight of a disc stabilizer. Some cover the entire CD and are quite light, others far smaller yet decidedly heavier. That didn't yet mention treatments with squalene oil, edge paints, a Furutech destat/mag spin or 'remastering' commercial discs to recordable CD to improve the sound. Such tweaks made the rounds during CD's heyday. Some makers even exposed their spinning discs to blue light. None of it would go down well with today's YouTubers so consider it mentioned in passing only.

Likewise for digital signal transmission really being an analog process of two rapidly alternating voltage states which represent the ones and zeros. Accurate timing of these switching voltages demands steep rise times. Too much rounding of a square-wave's edges equals jitter. That's time-shifted registration of when exactly a high-voltage one becomes a low-voltage zero. Other potential jitter creators are insufficient cable bandwidth and mismatched cable termination impedances causing signal reflections. So sonic pilgrims with max ambitions pay attention to digital cables of proper bandwidth and impedance. Simply substituting a spare analog interconnect isn't a good idea. In this context advanced users tend to prefer BNC sockets. They could fault Jarek's choice of RCA socket when his Lektor is an unapologetically hi-end not mid-level consumer proposition.

Denafrips socketry.

Full-on audiophile reptile Jarek went with this earlier 3-box Lektor Grand player. It implemented I²S on four discrete specialty cables per channel. That's before our sector defaulted to first S-Video then HDMI and RJ45 plugs to execute external I²S transmission with a single multi-core cable albeit zero consensus on pin assignment. It's why external I²S is prone to disappoint DACs that won't shake hands at all despite sporting the matching socket; or still invert phase or channels. That's unavoidable unless a transport uses dip-switch pin changes or executes them with multi-function switches à la Denafrips. Wherever a maker can't guarantee a perfect match no matter the transport/DAC combo, one is nearly best off without external I²S. As to upsampling in a transport, I've only come across it with Denafrips and Ensemble. Far more arguments exist which claim lower jitter for leaving externally routed digital signal at its native sample rate thus upsampling to a DAC.

Now our review table is set.

Where's the meal?