Thats pretty cool, and i like that he is trying to fix the typical bad ESL dispersion using electrical segmentation. i wonder how it ended up measuring off axis?
Not so sure if i like the V shaped bass bin behind the panel though - why mess with the figure 8 radiation pattern when you dont have to?
just place the bass next to the speaker and it will be in the dipole null, and not make any problems.
Having rebuilt my old esl63 a couple of times, i know from experience how much of a pain it is to tension the mylar juuuust right, only to have it tear last second, so props to this guy for seeing it through, and making everything from scratch.