Of course the very word "distortion" can mean all sorts of different things.
As soon as you DO ANYTHING (Eq-crossover-limiter et) you are "distorting" the incoming waveform-meaning that it is not the same as what is coming in.
Agreed that different types of "distortion" sound differently.
But when most people talk about "distortion" they are talking about clipped waveforms or harmonic distortion.
That is the problem when you "generalize" about a very complex subject.
Or try to apply the same "criteria" to everything-even when it does not apply.
yes I think we agree..mostly. Amplitude or phase response shifts vs frequency are generally not considered distortion but a purist could call any error distortion (I don't). Some deviations from ideal, like phase response can be hard to even hear on simple signals.
I consider distortion, artifacts or new signals that gets added to some original input signal and that weren't present in the original input signal. Distortion is usually measured using pure tone inputs to make the added artifacts easier to separate. (Note: standard distortion analyzers ignore phase shift).
When playing back complex waveforms like music, naturally occurring overtones may coincide with low order harmonic distortion components making them hard to detect and clearly identify as distortion. Distortion artifacts that don't land anywhere near musical overtones OTOH will be much easier to hear.
My point was that you can not compare amplifier distortion to loudspeaker distortion. Just like sound masking becomes less effective the further the distance apart the two pitches. As you've probably experienced we can often hear a fraction of a percent high order amplifier distortion despite playing though a loud speaker making 10x that and more. This is a classic misunderstanding about interpreting distortion specs in music playback systems.
JR
