Dynamic Linearity/Stability in Loudspeakers

[sardonx]

After watching some video interviews of Paul Barton on you tube, I found one of his comments interesting. He was talking about what criteria are important for designing loudspeakers and at the end said
"The weakest link in any sound reproduction system compared with real life nature is not the frequency response, is not the distortion. It's the ability to play the wide dynamic range that exists in real life and the wide dynamic range that our human ear/brain combination is capable of handling."

He said that a lot of speakers are not dynamically stable in that when the source calls for a 5db jump (for example) in output over an already loud passage many will only want to play 4db. I'm wondering if he was simply referring to compression due to temperature or something else. And if it's something else, what makes one speaker more linear dynamically than another?

[bbggg]

Apart from that, he probably also meant that most speakers' distortion increases rapidly the louder they play, which makes us want to turn the volume down lest we get a headache. Have you ever thought "hey, this live unamplified music is way too loud"?? Me neither. That's why most speaker makers specify distortion at 1 Watt or 2.83 Volts, if at all. 3rd harmonic distortion can actually be soft clipping, ie dynamic compression, and when joined by its loftier cousins-once-removed (5th, 7th etc) it becomes hard clipping.

I should expect independent testers to plot distortion at, say, 20 Volts in the very least, but no one does. Wonder why! Surely just about any non-miniature speaker on the market can take 20 bleeding Volts and survive. Indeed I would expect testers and makers to measure/specify maximum loudness levels for given distortion levels (say 3%) at all working frequencies, but I suspect this will happen right around when the Horned One takes up ice hockey.