1) They're getting effectively nothing above 400Hz. Male speech is 450-650 Hz fundamental and female speech is typically 500-800 Hz fundamental.

2) Their limiting factor is the resolution of the lens, not the resolution of the camera, which is why they've got an extreme telephoto pointed at their bag'o'chips even though it's 6 feet away. The technique is simply not applicable from any kind of useful distance for surveillance.

3) This shit works much better in microwave. Leon Theremin (yes, that theremin) developed a bug for the KGB that had the passive membrane built into a "Great Seal of the United States" (which the US ambassador dutifully hung over his desk). When the KGB beamed microwaves at it, vibration of the membrane altered the capacitance of the dipole circuit, modulating speech over the carrier and allowing the KGB to monitor and record conversations within the embassy from across the street, through walls and everything. That was in 1945.

(the cavity opening was the eagle's mouth - droll, no?)

Espionage has used laser interferometry to detect vibrations on panes of glass for decades. Thermal glass makes it a lot harder... but a polarizing filter and other trickery gives you much more useful results than camming a bag of chips with your iPhone.