[Cryptech Tech] First data

[Bernd Paysan]

Am Dienstag, 27. Mai 2014, 07:18:09 schrieb Stephan Mueller:
> Am Montag, 26. Mai 2014, 02:28:51 schrieb Bernd Paysan:
> 
> Hi Bernd,
> 
> > Am Sonntag, 25. Mai 2014, 10:32:03 schrieb Joachim Strömbergson:
> > > Aloha!
> > > 
> > > Attached is a compressed binary data file with the first batch of data I
> > > extracted. I get 18% compression with gzip which shows that there are
> > > some patterns to be found in the data.
> > 
> > Yes, looking at the file shows that the data was stored utf-8-encoded
> > (Python artefact? ;-).  That is easy to compress ;-).  Converted to real
> > binary attached (better compresion than gzip, slightly better than lzma).
> > This one is incompressible, but other tests (histogram over byte values,
> > attached - that test revealed the utf-8 nature quickly ;-) suggest that
> > the
> > individual entropy generators have different biases.
> 
> But even with the real binary data, the Chi-Square result is not really
> convincing for a white noise. Thus we have to assume there quite a bias
> present.
> 
> Maybe some oversampling could help fix the skew?

No, each of the ring oscillators has a bias.  That's not something we can fix, 
nor do we need to.  For more discussion about the difference between entropy 
and white noise, I'll answer to your other, longer mail.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.cryptech.is/archives/tech/attachments/20140527/481114c6/attachment.sig>