[691] in cryptography@c2.net mail archive
Re: Random numbers from the '60's...
daemon@ATHENA.MIT.EDU (Peter Gutmann)
Mon May 5 22:32:05 1997
From: pgut001@cs.auckland.ac.nz (Peter Gutmann)
To: karn@qualcomm.com
Cc: cryptography@c2.net
Reply-To: pgut001@cs.auckland.ac.nz
X-Charge-To: pgut001
Date: Tue, 6 May 1997 12:27:06 (NZST)
>Right, though having a microphone attached certainly won't hurt. My
>point was simply that a microphone isn't *essential*, as it's been my
>experience that the microphone preamp on a typical soundblaster card
>is quite noisy. Just crank the gain all the way up and hash the resulting
>noise. If you do have a microphone connected, so much the better. Just
>make sure the speaker is muted to avoid some annoying screeches.
I looked at this some time ago for use with SFS/cryptlib (or at least I got a
friend of mine to do it for me since I don't have a SB), with no mike
connected all you get is the LSB's toggling in a fairly non-random manner,
with a mike connected it's slightly better, but you really need to stick the
mike in front of the computers fan before you get anything vaguely
trustworthy. In the end I discarded it as a source of no-special-setup-needed
randomness because the only way you could get reasonable results is by having
the user jump through all sorts of hoops, in which case you may as well get
them to drag their fingers across the keyboard or drag the mouse around a
window or any other needs-user-intervention method.
Peter.
