For the goods under the hood—their current CD player at right acts as our very solidly built digital precursor still with a 24/192 engine—I tickled chief John Laurel's feathers. Relative to marketing whereby to mint global brand awareness, he still didn't strike me as a prized song bird these many years later. Hence my ineffective radar on their status. For today's deck, were the 6922 for voltage gain? Pure buffers? Cap or trafo coupled? Max output voltage/Ω on RCA/XLR? Balanced or convenience XLR? If the former, how were the RCA symmetrized? If the latter, how the XLR desymmetrized? What type of filtering in the power supply? Isolation transformers on the digital inputs? What USB transceiver - Amanero, XMOS, Cypress Semi, C-Media, M2Tech, other?

And there you thought civilized reviewers weren't nosey. This one has quite the schnotz. How in-depth makers elect to reply, if at all, depends. Are they protective of their IP? Some also hate getting specific. One breed of audiophile prejudges all gear based on familiar bits and bobs. Sabre chips? Naw. Xpresso clocks? Oy veh. UpTone capacitors? Ayee. I jest. Just so, armchair engineers do believe that sonics are predicted by a laundry list of ingredients, safely from the distance without any listening at all. Stay mum or feed their illusion?

We already know about the dual 9018. Here their on-chip volume control and selectable filters are part of the game (fast/slow for PCM, 50/60/70kHz cut-off for DSD); plus there's a DPLL digital phase-lock loop. An unexpected novelty? The digital balance control as enabled by the use of dual mono DACs.

Each channel gets its own toroidal power transformer plus the tube buffer its own EI core. Specs promise 118dB SN/R and 130dB dynamic range. Input options equal one of everything - coax, BNC, I²S, optical, AES/EBU and USB. Stock glass is ElectroHarmonix. Finally, Apple's stylish remote wand is compatible for volume and input switching but not included. All that was the formally published stuff. Now we get to the extras. The first is from Terry London. "I just became aware that you are reviewing Canary Audio’s KD-2000 DAC and believe you are in for a real treat. I reviewed it for and found it to be a great music maker. My reference DAC is the Concert Fidelity battery-driven 040 DAC. It costs almost double. And yes, the KD-2000 was not quite on its level but a close second. I did find that the stock tubes shipped with my review sample did not let the DAC 'sing' as beautifully as it did with NOS Mullard tubes." That type conclusion is quite typical for inveterate tube rollers. TAS reviewer Dick Olsher belongs to them, too. For my tastes however, gold-pin Mullard 6201 in my Aqua Hifi LaScala MkII had instantly fattened up the midband, reduced the outer bandwidth and slowed down transients to get muddier and skew the tonal balance. I decisively favour the stock commoner's glass which that deck was voiced around. Cages for canaries. Just so, it's a fair reminder that tube-based gear makes for plug'n'play. Roll to your heart's content. Only believe your own experience. Have fun!

On that note, UPS did by managing to lose Lynn's shipment for more than a week. When they eventually retrieved it in Geneva—they finally sent her this photo to confirm that this was it, with the shipping box hopefully part of the deal—their lost'n'found section shipped it to Cologne before rerouting it to CH. No fun. Here now are John Laurel's answers to go with final deliverance. "The tube buffer section and power supply are my design, the digital circuit is from my partner Gery who also designed the Magnus product line. In the digital field he is more knowledgeable. So I let him handle the rest of the design. The KD-2000 is FPGA based with auto detection between PCM and DSD signals. All three 3 transformers are epoxy-potted inside the metal can. Each channel has its own dedicated toroidal transformer for the power supply and the EI type supplies the tube buffer which is capacitor coupled. The whole circuit is fully balanced. The RCA outputs merge the balanced signal. An input transformer is used for the AES/EBU input. The two coaxial and the optical inputs are driven by LMV7219 comparators, the I²S by a 74LVC245 comparator. Max output voltage on RCA is 3V, 6V on XLR. The output impedance on RCA is 5.2KΩ and on XLR it is 2.6KΩ. The USB transceiver is the Italian Amanero Combo384 module." Now I asked John whether the FPGA replaced Sabre's filters and upsamplers and/or implemented the digital PLL. With the high-ish output voltages, I also asked again whether the 6922 contributed voltage gain. Finally, relative to factory photos I'd requested, I was told that none were available.

"The 6922 are solely used as output buffer for the unbalanced RCA outputs. There is no voltage gain. The RCA outputs are summed with a transistor amplifier built with discrete components, not an opamp. The FPGA functions as the DAC's control center. It controls the input signal, the display and implements the digital volume control. It also interfaces the ES9018 to operate in mono mode." This prompted more questions. Did their FPGA control Sabre's on-chip volume for dual-mono mode plus balance? Or was Sabre's on-chip attenuator bypassed and executed instead by Canary's own code? And just to be sure, I had to reconfirm that the XLR outputs bypass the tube buffer since neither their website nor Terry's review make mention thereof. "Yes, the 6922 only operate on the RCA outputs. You are also correct that the digital volume and balance are from the on-chip function of the ES9018 controlled by the FPGA." In conclusion, the KD-2000 is a dual-path machine. The shortest path is via the XLR outputs which tap the balanced DAC signal directly. The longer path is via the RCA outputs which use a discrete transistor solution to sum the balanced DAC signal to single-ended, then follow it with a no-gain triode buffer. One deck, two flavours.