“Thank you. Obviously no gross physical receiver can respond to a message that brief, or at least that’s what I thought at first. But because there are relay and switching delays, various forms of feedback and so on, in the apparatus itself, the beep arrives at the output end as a complex pulse which has been ‘splattered’ along the time axis for a full second or more. That’s an effect which you can exaggerate by recording the ‘splattered’ beep on a high-speed tape, the same way you would record any event that you wanted to study in slow motion. Then you tune up the various failure-points in your receiver, to exaggerate one failure, minimize aB the others, and use noise-suppressing techniques to cut out the background.”
Thor Wald frowned. “You’d still have a considerable garble when you were through. You’d have to sample the messages”
“Which is just what I did; Robin’s little lecture to me about the ultrawave gave me that hint. I set myself to find out how the ultrawave channel carries so many messages at once, and I discovered that you people sample the incoming pulses every thousandth of a second and pass on one pip only when the wave deviates in a certain way from the mean.
I didn’t really believe it would work on the Dirac beep, but it turned out just as well: 90 percent as intelligible as the original transmission after it came through the smearing device. I’d already got enough from the beep to put my plan in motion, of coursebut now every voice message in it was available, and crystal-clear: If you select three pips every thousandth of second, you can even pick up an intelligible transmission of musica little razzy, but good enough to identify the instruments that are playingand that’s a very close test of any communications device.”