On Sonntag 21 September 2008, Atom Smasher wrote:
> it's probably my interest in crypto that makes me obsess about the
> quality of PRNGs, so feel free to tell me i'm nuts...
Head over to www.ciphersbyritter.com then. :-)
> things were so much simpler when a "sample and hold" circuit was
> just that, derived from transistor noise.
It is actually quite difficult to reduce the correlation of such
signals so that they pass the even the third or fourth test of
randomness. About the only source ouf physical randomness that
doesn't have these problems is radioactive decay (and that is a
problem in itself). If all you need is something that doesn't
overtly repeat and has a white spectrum, things are comparatively
easy.
Achim.
--
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk]>+
Factory and User Sound Singles for Waldorf Q+, Q and microQ:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds