Wait, if I'm interpreting the Shannon-Nyquist theorem properly here, doesn't that mean since the signal is 24 Hz that we'd want a 48 FPS frame rate to accurately get the signal?
But in this case they don't want that. They want the illusion that the water is standing still, which you get by sampling the signal at the exact same point in every cycle. So to the sampler/viewer it looks like there is no signal at all.