I think more likely this was made by some film students for a school project or something then by physicists with a working prototype.

I always thought the only way of doing this would be to have detached microphones positioned between the speaker and the target noise so the system has enough time to compute the inverse of the incoming waves and send them out with enough space to cancel in the correct phase.