Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:20:12 ADT
Mike Rosing (eresrch@msn.fullfeed.com)
Tue, 7 Jul 1998 17:26:13 -0500 (CDT)
On Tue, 7 Jul 1998, Ben Laurie wrote:
>
> When you say "zero cross overs" I presume you mean something other than
> zero (coz I'm kind of hoping that you can't make resistance negative).
I meant zero voltage cross over. The signal swings +/-. It is a "random
stationary" process with zero voltage average and peaks at +/- 2 volts.
I'll bias it with an accurate 2.5 volt reference and then digitize it.
As the signal goes from + to - (or - to +) that's what I call a "cross
over". I can interpolate the time from several data points because I've
filtered the signal electronicly to eliminate anything below 200 and
anything over 20000 Hz.
> More seriously, the conventional method is to measure them (that is the
> timings, and, presumably, the voltage every so often) and add them to a
> pool, which, once you've decided you've added enough bits of entropy to
> it, you hash down to the number of bits you think you've really got. I
> presume you already knew this, though, so I'm left wondering what you
> are _really_ asking?
One thing I'm trying to do is to compare "real" random with "ideal"
random. Real random data is not the same as a hashed output. That
process can always be added. That's not what I'm interested in. Real
processes have scale limits, so getting random data from a real source
must be done carefully.
Radioactive decay is considered to be "truely random". My purpose is
to compare this to statistics tests to see how well the tests work. If
the tests say a "truely random" source isn't, maybe it isn't and maybe
the tests are screwball. It's a tough problem. And that's the kind of
thing I like to work on :-)
Hopefully by the end of summer I'll have enough data to toss at DIEHARD
and we'll see what "real random" is (or how much more work I'll have to
do on my circuits!)
Patience, persistence, truth,
Dr. mike
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:20:12 ADT