This review page is supported in part by the sponsors whose ad banners are displayed below

Round 1. Wave 40. SPEC vs. FirstWatt. This system came conceptually closest to the designer's own vintage ideal of SET + horns even though my lower speaker sensitivity and lack of front-horn loading still differed significantly. On the one hand this made for a very typical low-power tube-friendly proposition vintage by design. On the other hand it didn't make for a scenario where at least traditionally we'd have expected a class D amp to do well. But times change. Take the forthcoming Greek Kalypso speaker from Rois Acoustics [right]. It will combine a whizzer-fitted Tangband 8-inch widebander with a folded rear horn and built-in PCM-to-PWM true digital amplification by Audio4Soul driven from either a USB or coax input. My recent DIA-250 review had shown how modern high-power class D amps can conjure up valve-reminiscent warmth and fullness into counter-intuitive loads well within the first-watt window. As such my Swiss soundkaos eggs were ideal to test Shirokazu Yazaki's claim that their approach to class D combines the best of traditional transistor and valve topologies. And so too were my simpler-than-triodes SIT1s ideal to juxtapose ultra-refined modern SET sound with specs better than equivalent valve amps. How would the RSA-M3EX strut its stuff in that context?


To simplify A/B/A swaps I wanted to run the RSA-M3EX in amp-direct mode preceded by the same Nagra Jazz preamp which drove the Nelson Pass monos and only subsequently compare the Nagra/SPEC combo to pure spec. Whilst the Jazz into the 35dB-gain Job 225 is dead quiet—the measured S/N ratio of my actual Nagra unit is 114dB—the SPEC in 20dB fixed mode was very noisy in the seat even when the Jazz was set to 0dB gain. In variable mode and with the Nagra set to unity gain, that noise disappeared. I thus would eliminate the Nagra from the go and only assess the RSA-M3EX as an integrated amplifier. In that mode and at the actual volume setting used, the SPEC was dead quiet whenever I paused my source.

The upshot of this juxtaposition against a €12.000 valve preamp + $10.000/pr single-ended transistor monos was a far from wildly dissimilar sound. My twice-priced separates were simply the more informative in the upper half of the sonic spectrum. This translated into more specific depth layering and related separation. I also had the more illuminated upper harmonics and subjectively greater speed with predictable outcomes on transient perception. The Japanese integrated played it darker, heavier and lusher. According to modern parlance and how we use the word 'resolution'—pixel count—it stepped that down. The emphasis instead shifted on something more sumptuous, languorous and rich. In tube terms it was in fact more vintage 300B (Western Electric or Sophia rather than Emission Labs) than the leaner quicker flavor of a 50 or 10Y.

If we drew a horizontal line with end points of fast/lean and relaxed/warm at the left and right respectively, the Bakoon AMP-12R would sit at the far left followed by the Crayon CFA-1.2. The Goldmund/Job 225 would sit between those two but a tier lower to indicate less refinement. The Jazz/SIT1 combo would take up the approximate midpoint. The SPEC then settled down rather farther to the right. How much farther I'd determine vis-à-vis the Gato DIA-225.

Quite a bit. I no longer had the DIA-400 on hand to determine whether that might overlap. That's because the big Dane had struck me as rather more fulsome and lush than its smaller sibling to walk our line way over to the right. With all that implies the RSA-M3EX sounded plainly louder than the DIA-250. It was chunkier. Meatier. At the same time it also projected more forward. This reduced depth and the Gato's more clearly walk-about stage. If we visualize a sea of violins—I was on a kick through various Louis Spohr clarinet concertos on the Hyperion label, with Michael Collins on blackwood—the Gato showed more white caps and ripples. More wind on the surface. The SPEC played it as a more homogenous mass. Less wind. For the clarinet's timbre the difference was akin to that between B-flat and A lengths. The longer instrument sounds darker and warmer.


This gets us directly at the live/playback dichotomy. As reported elsewhere, I'd recently attended our classical music fall festival in Montreux's Stravinski Hall. We'd heard the London Philharmonic under Charles Dutoit with Mussorgsky, Rachmaninov and, fittingly, Le Sacre du Printemps. Sitting in row 7 of the floor's left third the sound there was far closer to what the SPEC now proposed—massive, dark, with more wall-of-sound than vivisectionist staging—than what generally goes for high-resolution hifi sound. With the latter the obvious lack of actual visual data is replaced by quasi-visual playback cues like imaging, performer outlines, halos, pinpoint focus and heightened separation. None of that belongs with unamplified live music. 1:0 for SPEC. Hurray!
But now we turn tables to inspect the average studio production. Microphones stare down throats and F-holes to capture impressions at pornographic proximities and massively paralleled. Unless we did a Van Gogh on a number of folks simultaneously, no human ears would ever duplicate that. Still with clarinet but now Eddie Daniels and Nepenthe on the GRP label, this altered gestalt equalized the game 1:1 for everything left of the SPEC. Obviously close-mic'd multi-tracked recordings make up 95%+ of what we listen to. Now SPEC's readings diverge. They recalibrate such recordings to sound more like an acoustic live concert from a good distance. That's the very essence of this proposition from Japan.


The question any buyer has to ask is whether this undeniably attractive makeover has their vote. It's the old beauty versus truth versus beauty is truth conundrum. Here writers must bail. How one decides is entirely personal. It relies on what one uses as reference, what one expects from a hifi and how one weighs tone mass vs. transient speed amongst various trigger points. Based on my exposure to class D sonics and using our earlier graph but now with Hypex Ncore at the center, I'd peg early NuForce well to the left and SPEC far to the right. Most current ICEpower falls between Ncore and SPEC but closer to the latter. Gato's DIA-250 slots between ICEpower and Ncore. AURALiC's Merak monos are the 'poor man's Ncore' to sit right next to NC1200 as I heard it in Acoustic Imagery's Atsah monos. Where does that put SPEC? As Anssi Hyvönen of Amphion put it who provided his speakers to SPEC's Munich HighEnd 2013 exhibit, "it sounded very liquid and for lack of a better word analogue". Spot on!

That was my fix on the lay of this land. Up next would be the amp's behavior into my already SPEC-voiced Rhapsody 200 towers.