The point "..there are things that cannot be measure..", is of course a valid point by itself, but it is also often irrelevant when apply to comparing audio gear. For example, typical bench test measurements of amps and DACs would include THD, Noise, Channel balance, Crosstalk, IMD, linearity, Frequency response among others. All of those are important to measure and each are well defined and understood by most. The PSA Stellar gain cell measured flat in FR so there should be no artificial boost and attenuation of any particular frequency within the audible range, so that one is out of the way, what's left then..
Where it did poorly was in THD+N (at high output level) and linearity (at low level) so if one listens at not too loud (such as ref) level, it should still be consider accurate enough for it it not to be picked out easily in a properly conducted AB comparison listening session.
Distortion is just that, it has a literal meaning. So if the gear does not distort the input signal, it literally means what goes in is what comes out, that is, the output may take a different form (e.g. digital t analog, and/or at different level, but the bottom line is, the original music waveform is preserved. It may be debatable regarding what is the threshold of audibility, is it 0.1%, o.05%, or 0.01%? 0.1% seems widely used, Gene would probably use that number, while others may say 0.05 or 0.01%. I would therefore argue assume it is not debatable that 0.005% should be below the threshold regardless of the harmonic contents. It doesn't matter if the harmonics are even or odd, if the total is at 0.005%! If it is 5% or even 2% then for sure it matters but while at 1% it may still be audible it may not be "easy" to tell, or at least a debatable point.
I don't doubt what Leemix heard/experienced, but it was so easy, then I have to wonder whether it was a truly apple to apple AB comparison. One is a DAC and the other is a prepro, so it takes some effort to ensure the comparison of the two as a DAC is valid one.
