Hi,

The principle is electrically quite sound.

IF a driver produces large amounts of 2nd H at the normal listening level, BUT low levels of 3rd H, then adding about the same amount of 2nd H but in opposite phase cancels the distortion in the driver.

You could call it feed forward error compensation. It was also used on many records during the 60's, 70's and 80's in the form of a tracing simulator (which cancels certain types of distortion resulting from using spherical tip stylii).

If applied well you can reduce the HD of a given driver by 10db and more over wide ranges of SPL and frequency, it is better applied to active systems than passive ones.

You may find these articles illuminating:

http://usuarios.uninet.com.br/~edelima/

Ciao T