After picking on audiophiles for poor testing I made an equally silly mistake.

Grayson told me an interesting experiment yesterday: Play pink noise through both speakers, with the same signal.

./pink-phase 1 1 | aplay -f cd

You'll get a localised sound source in the middle. Now negate one channel:

./pink-phase 1 -1 | aplay -f cd

The sound will come from behind (especially if you have Dolby Surround :) So I though, yeah, I've got the bits to test this, let's do it.

g++ -o pink-phase{,.cc}

Well I did it, and it was measurable, and it was backwards. Sure enough, input wires were swapped on one channel. I can't pick about supertweeters:)