Downstairs. Here I removed our Elrog ER50 fitted direct-heated/coupled triode preamp from Vinnie Rossi for a COS Engineering D1. That marries a Burr Brown DAC with custom filter to analog volume and inputs. The Nagra fed its analog unbalanced input, our music iMac the USB input of either machine through Audirvana 3 for very easy comparisons. Amplifiers off the COS were LinnenberG's 200-watt Liszt monos with lateral Exicon Mosfets. Speakers were our usual Audio Physic Codex 4-way. All cabling was by Allnic except for a KingRex red double-header USB cable.

Cueing up the same Dvorak wind serenade for its familiar timbres—I loved to play it during my clarinet days—the most obvious difference between Taiwanese and Swiss converters were the upper bands. The treble of the COS was more informative and developed. That prioritized illumination, outlines, focus and created an overall leaner more quicksilvery temperament of high relief and reflexes. With the Nagra, timbres subtly shifted. The clarinet's German bore had nearly become French. Given the upstairs preview, that wasn't unexpected. The obviousness of it still surprised. However, its means or the how weren't as great as the effect or the what.

For confirmation, I spun up Schubert's Octet with the Quatuour Modigliani, Sabine Meyer & friends. Whilst a very different recording and composition, the same held true. The COS was more top-down lit up, spatially teased out and more emphatic in the upper harmonics.

The Nagra was earthier, rounder, sweeter and weightier. If across its color spectrum the COS was a bright summer's day, the Nagra was more autumnal. If the COS was a platinum diamond ring, the Nagra set a ruby in rose gold.

The difference was about tone, textures, how defined the recorded venue manifested as reflective space and how the sense of color gravity communicated. To take the easy but still true way out, the Nagra really did sound like DSD, the COS like pristine PCM. Or, the Nagra sounded be-tubed whilst the COS was typical solid state.