Subjective evaluations of audible distortion and compression in larger well-designed loudspeakers usually requires playing them at loud levels over an extended period of time (20 -30 minutes). With that, there are risks associated with damaging the listeners' hearing that most researchers are not willing to take. There are ways of mitigating this:
1) use a very large room and increase the listening distance or 2) make a binaural recording them at loud SPL levels and play the recordings back over headphones at reduced levels.
For these reasons, most companies rely on objective measurements to determine the max SPL limits of the loudspeaker where the measured distortion is above audible threshold and likely longer acceptable. If you have a Klippel analyzer you can measure the nonlinearities in the speaker, model/simulate and auralize them over headphones in a very controlled way.
www.klippel.de