Klippel suggested a scheme which increased the values of distortion as frequency rises- i found it on a google search
(dont forget to use the advanced search function to search sites individually)
Also of note is not just a % figure overall-but how it varies with frequency on transducers,and how class A/b amps although measuring well can have HF artifacts,and class A reduce distortion with lower level more like Valve amps(but are only similar in that respect)Im planning on a 3.1 way active system with Class A eventually
Cheers!