Thanks for the mention, MrD. I agree with the mastering engineer, at least as a general rule of thumb. And especially if we define "distortion" as "deviation from accuracy," rather than more narrowly as harmonic distortion. Here are reasons that occur to me off the top of my head; I’m sure there are other reasons as well:
1) As alluded to earlier, it is widely recognized that certain even order harmonic distortion components, especially the second harmonic, tend to be subjectively perceived as contributing to richness and warmth.
2) As Ralph (Atmasphere) has pointed out in a number of past threads, if the distortion characteristics of an amplifier are such that harmonic distortion becomes vanishingly small at low power levels while increasing significantly at high power levels (as is the case with SET and some other amplifiers having single-ended output stages), since our hearing mechanisms use certain odd order distortion components as loudness cues the subjective result will often be a perceived increase in dynamics.
3) My understanding is that very small amounts of high frequency noise or hiss can be subjectively perceived as added "air" and ambience. That would seem to be a reason, btw, that some audiophiles report finding that shielded cables tend to produce a more closed in sound than unshielded cables. Shielding can potentially reduce high frequency noise in at least three ways: (a) As a result of increased capacitance; (b) by reducing RFI pickup; (c) by reducing ground loop-related noise, as a result of lowering the impedance between the circuit grounds and/or chassis of the connected components.
4) Here is an interesting quote from the manual for the DEQX HDP-5 I have in my system:
Room measurements typically exhibit a downward “tilt” from low bass to high treble of 6 up to 15 dB. This is caused by a number of factors including reduced dispersion and greater absorption in the room at high frequencies. Do not attempt to EQ your room measurement completely flat – that will most likely sound overly bright.
That rings true to me. And it indicates that optimizing in-room frequency response is a delicate balancing act, with the result unlikely to be anything that can be considered to be precisely accurate in an objective sense.
5) It was established in the late 1970s, as I recall, that the commonly used technique of reducing total harmonic distortion (THD) by application of feedback can often result in significant amounts of transient intermodulation distortion (TIM), which can be much more objectionable than THD. So in addition to adding richness, warmth, and in some cases enhancing dynamics, highish levels of THD may be appealing in some cases due to an associated minimization of TIM.
Regards,
-- Al