Topping the desk. With the iFi and Wattson decks getting simultaneous digits off Singxer's SU-2 bridge with Qobuz Sublime streaming the cloud or Audirvana Origin serving up local files—I split my monitor screen to show both windows—the Final D8000 planars were the resident plug'n'play load. On commercially mastered fare like Anoushka Shankar's new live album Between Us, 9:00 got loud enough. On 'purist' recordings like Al Gromer Khan's new Ambient Religion whose median level sits a good 20dB lower, I nearly maxed out Madison whilst the iDSD Pro Signature only hit 10:30. Hover mouse for loupe enlarger to see actual settings. Where the iFi bench-presses our Susvara, Madison would choke and lose its lunch. It's a minimalist network player with added headfi. On appropriate loads that's sonically far from a convenience feature. On raw drive it relies on you sucking tunes from reasonable cans. For final housekeeping declarations, Audirvana's 64-bit r8brain upsampler renders my local files as 176.4/192kHz. iFi's FPGA upsampler then adds its own x 4 multiplier for 705.6/764kHz processing. For its digital filter I prefer Thorsen Loesch's Gibbs Transient Optimized aka GTO. Madison received the same x 4 upsampled data from Audirvana then applied its own Anagram-evolved math. Qobuz ran native so often 24/96 due to my subscription tier and selections. [As someone who buys the music he fancies most, their Sublime tier pricing is significantly lower for hi-rez files than the standard tier's Redbook versions. That's another reason to get that Sublime shine.]

The iFi staged wider and spatially showed more connective tissue even without engaging DSD resampling. Madison played it a bit more compact, its more obvious blackness swallowing up some image halos. The net effect was that the iFi sounded a bit bigger and fluffier, the Wattson tighter and tauter. Operators distrustful of operational amplifiers would be bothered by their ears not being bothered. As far as I heard, the Swiss choice for IC-based output stages only reflects in this minor dryness and spatial compression/focus playing counterpoint to iFi's greater fruitiness and expansiveness.

Whilst the Brit will handle heavier loads to be more universal or 'serious', on the D8000 I mostly heard sonic leveling. That's because Madison's drier textures weren't chalkier or paler. Its color intensity closed that back door through which dry specimens of class D often intrude. Here I'll mention chip amps à la Clones and Moonriver. Their IC outputs likewise combine dynamic and color crunch with grip that's a net effect of dry + taut. In paint terms, think matte not gloss.

Time to scuttle headfi and listen over speakers in the main system where a Sonnet Pasithea's variable reference voltage on her R2R ladders drives an Enleum AMP-23R into Aurai Lieutenant speakers. Madison did her own variable thing via Leedh, just without Sonnet's metal IR remote. Imagine this Wattson app on a smartphone instead. It adds a bass EQ feature not otherwise accessible; and Tidal Connect is coming. Back to my system, incoming digits of local files start life on my iMac's FusionDrive buffered in 32GB RAM by Audirvana. That also minimizes background threads and enables exclusive access.

Those digits then output via a double-header USB cable to a Singxer SU-6 bridge which applies buffering, reclocking on Crystek 957-25 OXCO and advanced FPGA shenanigans powered by super capacitors for virtual batteries. The Dutch then received those already dry-cleaned digits via AES/EBU, the Swiss over coax. To compare just required changing over one set of analog Allnic interconnects to the amplifier. Not having a 5V/2A linear power supply on hand, I again used the stock medical-grade Meanwell switcher. Inveterate tweakers discounting all such solutions as inferior to linear power by design might look at a $125 LHY LPS25VA from Singapore. Playing to similar headspace are my desktop speakers. As actives, they're DSP corrected for impulse response and linearity. Yet their external class D amp with embedded signal processor only takes a balanced analog signal. To many armchair pundits its subsequent A/D conversion renders quality assessments of preceding DACs moot just as they claim for comparing cartridges via digitized versions of needle drops. Needless to say, I did try Madison as DMAX driver and—would you believe it—heard the same difference as over headphones. Also mirrored was seemingly maxing out available SPL on Al Gromer Khan type productions. I wrote 'seemingly' because I discovered unexpected life inside the two uppermost LED. But if the possible 4.2Vrms on XLR aren't tapped because the music you play hovers at ±0.5V instead and your amp/speaker combo needs more, you will need extra gain from an active preamp. Here iFi's otherwise excessive 11.7V option makes a big difference.

Meanwhile on YouTube music videos, Madison's half mast was once again the sweet spot. How the engineer set Leedh's attenuation gradient across his scale felt expertly weighted for normal scenarios. If it were a full circle, 12:00 o'clock would be the golden mean. With a half circle, it's 9:00. Voilà. That's Madison's essence: being for normal scenarios and users. All tweaky nerdy decisions are made on our behalf then hidden from view or access. If we can't mess with the recipe, we can't screw it up. As the first Polish review put it, "I'm in fact somewhat surprised by the choices made by the engineers related to sound…measurements and theory are one thing while practice is something completely different… the Wattson Audio player sounds significantly different from for example the ferrum or Mytek players I reviewed some time ago…"

Having myself reviewed the new Mytek Liberty DAC II a few months earlier as well as both ferrum decks, I completely concur. Madison plays chocolate ice cream to their lemon sorbet. I'd group Chord's aesthetic into the same sorbet class, Audio Note into Madison's. That such a small component could set such a decisive course will surprise only those who equate substance with size; or think that all D/A converters sound the same being just 1s and 0s. Elephants amongst us—for our excellent memory not wrinkles—recall how Audio Aero loved coupling sub-miniature tube output stages to their Anagram digital engines. Today Vermeer do it again. Wattson now walk a not dissimilar path with high-feedback opamps. It's not raw ingredients which fix the result. It's how the engineer implemented and seasoned them.

This is a 48-min. concert. If you don't have time for the full hit, track 3 is my favorite. You can jump to it by selecting 'watch on YouTube', then clicking on 'more…' in the description box below. With YouTube's compression algorithm throwing away data it thinks are 'masked' by louder sounds, Madison's tuning plays a particularly winning hand.