Fooling humans is much easier when about 50% of the sound is always accurate (in any case it's a wooden drumstick hitting something). The human mind is very for

Fooling humans is much easier when about 50% of the sound is always accurate (in any case it's a wooden drumstick hitting something). The human mind is very forgiving, especially when vision accompanies sound (McGurk effect [1]) Furthermore this fixed variable makes the pool of samples to choose from much much smaller. Still impressive of course ;) just far from dubbing any video there is. [1] https://youtu.be/G-lN8vWm3m0?t=32 https://youtu.be/G-lN8vWm3m0?t=32