SINAD is just distortion and noise. It's usually easy enough to see from the SINAD vs. Power curve which is dominating. Usually, noise dominates at lower powers, because it is closer to being a constant than the power, and thus, the ratio of noise to power goes down as power goes up. This does not mean that noise goes down, it just means that noise does not rise as fast as power does. Most amps are noise-limited at low power, and the better the amp, the truer that becomes. Noise as a ratio usually presents as a line sloping downward.
Some amps have enough distortion to become dominant as the power goes up, at which point the steady downward slope dominated by noise turns up again, showing that noise is no longer dominant. We admire amps that don't have rising distortion with high power, but really I'm not sure how audible it is. Once the SPL gets high, lots of masking is occurring and our ability to discern distortion, which for most of us is not nearly what we think it is, become further limited. This amp showed high distortion at high frequencies, but I note that the AP had to be evaluating 45 KHz of bandwidth to see it--the harmonics of even 8 KHz are above what many of us can hear, particularly clearly. An 8K signal with a second harmonic that is 40 dB down will sound clean to us--I know I can't hear harmonics 40 dB down at frequencies above the bass register, let alone when those distortion products are above my range of hearing. If our hearing at 16K is already 40-dB attenuated, a further 40 dB down will be unhearable. (I'm not talking about intermodulation distortion, which can have products much lower down.) At power output levels and frequencies where most listening occurs, the SINAD is better than 60 dB. Where it really counts, this amp just isn't that bad.
With this amp, the SINAD curve was complicated by the fact that at the higher power-supply voltage as switched in based on power demand, the noise level jumped up somewhat.
But it's power specification was at 0.5% THD, which is just -46 dB. This example was doing 10 dB better than that at the onset of clipping and 20 dB better at usual usage levels, feeding 8 ohms, notwithstanding that the maximum power was a couple of dB (and only that) lower than spec. Given how little we know of the history of this amp, it still largely does what it claims to do, and I think a listener (who is just listening) would have to approach it pretty critically.
This is especially so given my own belief that clipping is the usual thing people hear, and this amp is going to be making some serious noise before it clips.
Would I choose it versus the similarly powered NC502MP Buckeye, which is roughly the same price? No. But I would choose it over a Benchmark if loud playing was a use case and cost was a factor.
Rick "pickier about SINAD upstream from speaker amps, because that noise gets amplified" Denney