In Video 1 at T=5.4 - 7.14 you can see that groups of pixels, in increasingly larger groups, move "up" to deliver air pressure pulses that emulate the positive part of the sine wave produced by the analog speaker "pushing" air, then the same sequence of groups reverse to emulate the speakers sine wave "pulling" air. But if you look at the depiction of the pixel pulses they go up and then back down to zero amplitude by the end of the clock tick. Then another group pulses, going up and down. This would suggest to me that the pixel piston returns to a centre position with zero amplitude. Then the negative pulse sequence begins and the air movement is in the reverse direction - pulling air backwards to emulate the sine wave's negative half.
As Yuval Cohen then states very clearly, even though the resulting depiction does not appear to be (graphically) identical to the analog sine wave, with a large enough pixel array and optimal signal distribution, the RMS sound generated by the digital speaker array can be made to be identical to the analog speaker - and more importantly, to the original sound. This is DSR in a nutshell.
I think that the development of the speaker pixels since that time has changed from it being a simple electrostatically driven membrane moving between 2 maximal end points of travel (top and bottom) between charged plates, to an electrically driven "stiff piston" driven by piezoelectric lever actuators (see the patents) between stops at top and bottom, so there is also no possibility of the piston moving beyond that position. The movement is clean and very precise. I think this may now give better control over the piston and thus the resultant sound. But I also think that when there is no voltage applied to the pixel the piston sits in the centre position. I think this was also true of the original electrostatic system too. I will grant that I may well have missed something in the explanation. Either way it does not matter that much.
What does matter is that even with the first prototype chip we had proof positive that DSR works as expected. Audible sound can be produced coherently from the array of pixels operating at a 200KHz clock rate.
We have seen the demo of this in the form of that very crude chip playing Johnny Cash's "Ring of Fire". Yes it was crappy sound - it only had 128 pixels - but it played music direct from a digital audio signal file. That was its significance, and that seems to have either been forgotten by many, or never seen by newer investors.
With an array of multiple chips, each with 1028 pixels, AP will be able to play everything from the soaring violins of the Pastoral Symphony to gut thumping heavy metal - direct from the sound files: no speaker stacks, no subwoofers, no tweeters, no fat power cables, no distortion, no carefully crafted cabinet designs needed. Just perfectly reproduced sound, just like standing where the conductor stands.
(20min delay)