Huh? Loudspeaker distortions can be orders of magnitude greater than good amplifiers, and will often dominate. Cross-over distortion at the levels seen in early transistor/transformer push-pull amps is very much a thing of the past, good solid state amps can be < 0.01% distortion, which is basically well below audibility, be it all 2nd order or 10th order.
The nice thing about THD is its simplicity - if the value is low eough you don't have to worry about the kind of distortion, and if you do have to worry about the kind of distortion then the residual or distortion spectrum are both much more imformative than any single metric. However just testing at 1kHz is insufficient, and the graphs of THD v. frequency and v. level are much more useful to have.
This is true of Gedlee metric too - testing only at 1kHz is not enough, many distortions rise rapidly with frequency.
The elephant in the room however is less nicely behaved distortions, like thermal and other memory effects, they are the trickiest to characterize and may be slipping through testing in some cases, being history-sensitive.
The nice thing about THD is its simplicity - if the value is low eough you don't have to worry about the kind of distortion, and if you do have to worry about the kind of distortion then the residual or distortion spectrum are both much more imformative than any single metric. However just testing at 1kHz is insufficient, and the graphs of THD v. frequency and v. level are much more useful to have.
This is true of Gedlee metric too - testing only at 1kHz is not enough, many distortions rise rapidly with frequency.
The elephant in the room however is less nicely behaved distortions, like thermal and other memory effects, they are the trickiest to characterize and may be slipping through testing in some cases, being history-sensitive.
Which metric or concept did you apply here? I am guessing you're wired to THD.
Why only 1kHz? You can apply the metric at any frequency just like THD.
There are no thermal nonlinearities in loudspeakers, I know, I checked.
"Memory effects" - what are those? In nonlinear analysis this usually means that the nonlinearities are frequency dependent, and "memory" refers to this effect. It is accounted for in the metric if you do it correctly.
