Devialet's newest technology...?

[f1eng]

As far as I can understand it DSP can correct for some types of distortion but not all.
Let's assume that the signal is 100Hz and its overtone of 200 Hz. Any error in timing and level between these frequencies can be corrected through DSP like SAM if one knows the system. It also could correct for different behaviors at different absolut levels.
If that signal causes also some enclosure vibrations at 300Hz (enclosure and membrane vibratios are chaotic in nature and not necessarily idential to the stimulating frequency)any correction at 300Hz would not be effective as the signal does not contain 300Hz.

Do you know what the transfer function of an error is?

Cabinet excitation would be included along with everything else.

It is possible that the error transfer function is so complex that the inverted version of it can not be multiplied by the input signal in real time but since much of the impetus in the production of the second generation models was this capability I would imagine they fitted plenty of DSP power.

The excitation of cabinet by drive units is complex but entirely predictable. 200Hz won't excite 300Hz at a high level anywhere BTW, but if it did that output would be included in the transfer function and when inverted and multiplied by the input would cancel it.
Even if the sound is coming from the cabinet the correct amount of out-of-phase sound from the drive unit would cancel it at the listening position.

This is a clever application of maths.

Noise cancellation uses similar technology.