http://www.youtube.com/watch?v=iLPIgXCWUWA
Sounds plausible except they threw in the "...two miniscule pauses...each are 2 microseconds long." Even if you consider that the 2 microseconds could be a halfwave and we are looking at a response of 4uS on the tape, it would mean that the recording fidelity of the tape would need to be better than 250kHz. I can't buy that. 2 milliseconds I could buy. Maybe they misspoke.