Whispers of a working chip, page-150

Given how it's described in the video I'm almost certain that the pixels don't return to a centered position before they start to create the opposite half of the sine wave. This totally makes sense, as in order to create the greatest pressure (volume) you'd want to utilise the full excursion of the membrane.

"This is graphically depicted on the video as a whole series of pixel actuation sine waves within the audible sine wave"
- Look closely, what is depicted isn't a bunch of sine waves with the audible sine wave, it's a depiction of the pressure being created by each (set of) pixel(s). If you slow the video down you can see they've actually shown this rather accurately too, i.e. the maximum pressure is at the mid point of the pixel's/pixels' movement.