Re: FM radio hiss etc.

[Travis Savo <admin@max-web.com>]

At 02:56 AM 8/3/98 -0000, you wrote:
>Hi,
>Dumb as I am, I have a question about retrieving random stuff out off radio
>hiss:
>
>If it is/could be possible with FM hiss, is it possible too with a
>recording of, for example, a leeking faucet, tree-hiss forced by the wind,
>a sack of beads falling to te ground, even the sound of a shower, etc.? And

I don't see why not. If you got a high enough sampling rate (44khz should
do nicely), and set a peek threshold so anything under a certian volume was
a 0 and every peek over was a 1 I could think of a BUNCH of samples that
with a little playing with the threshold could produce some pretty random
resluts. Furthermore, by adjusting the threshhold you would get a toatly
different number. Using a random sampling with a random threshhold would
bring you closer to a truely random number. Of corse it's still not
possible to get a truely random number because you gotta get the threshold
value somehow... see below for why...

>what is, by the way, TRULY randomness - could that ever be reached? I have
>a bold assumption: it never can be.

That was my understanding, but I'm no expert on the subject. 

>Housenumbers aren't good for that - that *I* know.

Absolutely, because the seed is not truely random. For instance, most
random number generators use the current time from the internal clock to
generate a random number. However if you were to do the exact same thing at
the exact same time (down to the millisecond) it would reproduce the same
number.

If you try this let me know what you find!

>Chris Harwig
>Nieuwegein, The United Netherlands
><chrisharwig@hetnet.nl> (no acces to the Internet; due to lack of money...)
>
>(I am not a native English speaker - as you probably see.)
>
>P.S. I can't help it: I miss Linda Reed fka Linda Reed...
>