[Cryptography] A TRNG review per day: Turbid

Bill Cox waywardgeek at gmail.com
Wed Oct 29 19:42:19 EDT 2014

On Wed, Oct 29, 2014 at 6:16 PM, Clemens Ladisch <clemens at ladisch.de> wrote:

> Bill Cox wrote:
> > On Wed, Oct 29, 2014 at 8:55 AM, Clemens Ladisch <clemens at ladisch.de>
> wrote:
> >> Bill Cox wrote:
> >>> There will still be significant correlation between samples.  There is
> >>> thermal noise in a band from 9X to 10X below the sample rate
> >>> which will turn into a significant short-term correlation between
> >>> samples 10 away from each other.
> >>
> >> The noise in all the other bands will cancel out these correlations.
> >
> > Not in my experience, but that is somewhat limited.  A simple test would
> be
> > seeing if the zero crossings are correlated between adjacent samples.  My
> > guess is they are highly correlated, as in I have a 70% chance of
> guessing
> > if your next sample is greater or less than zero if you tell me the full
> > value of the previous sample.
> And as it turns out, there are different decimation filters for different
> sample rates (typically, higher rates have smaller passbands).
> I made a quick test (using only the sign of adjacent samples), and there
> are indeed lots of correlations (at any sample rate).
> In any case, higher rates capture more useful noise.
> (And Turbid uses the highest rate by default.)

In this case, it sounds like we're in complete agreement.  This is what
*should* happen.  Turbid is doing the right thing.

> > If you send me some typical Turbid maximum sample rate sound samples, I'd
> > be happy to do that test.
> Knock yourself out:
> <http://cladisch.fastmail.net/cs5381-48000.wav.bz2>
> <http://cladisch.fastmail.net/cs5381-96000.wav.bz2>
> <http://cladisch.fastmail.net/cs5381-192000.wav.bz2>
> <http://cladisch.fastmail.net/vt1708s-48000.wav.bz2>
> <http://cladisch.fastmail.net/vt1708s-96000.wav.bz2>
> <http://cladisch.fastmail.net/vt1708s-192000.wav.bz2>
> <http://cladisch.fastmail.net/ak4520-44100.wav.bz2>
> (The last one is an extreme case where Turbid would refuse to work.
> And it doesn't have much entropy even under optimistic assumptions.)
> Regards,
> Clemens

I'll dive into this in the morning, though I expect to find what you
found.  Again, I am a fan of the the models the Turbid authors put
forward.  I don't know if they coined the term "surprise", but I don't know
how to measure entropy in a sample without it.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20141029/67625e24/attachment.html>

More information about the cryptography mailing list