It was a video in which he explained the development of his patented tweeter array. He said he heard a 440Hz note played by a violin, and then measured the weight of the string the note was played on, which was 1/3 gram. He figured if he used a driver (or drivers) whose moving mass was no more than 1/3 gram to reproduce the sound of the violin, the reproduction would ipso facto surpass that possible from a driver(s) whose moving mass outweighed the violin string.
A smaller driver/smaller mass--> improved impulse response (IR) --> improved resolution/detail/clarity compared to a larger driver/larger mass playing in the same band.
Some designs will let a capable tweeter (small driver) crossed over/play as low as possible. His approach for the midrange handled by small tweeter sized drivers have this benefit (IN THEORY). Theory and execution may not always line up depending on how the co was designed, drivers themselves, etc.