Re: A revision of my text stego scheme

On May 20, 4:46 pm, Mok-Kong Shen <mok-kong.s...@xxxxxxxxxxx> wrote:
Am 20.05.2011 18:46, schrieb jbriggs444:





Using line length to encode your stego bits is clumsy.  Why not
compute a hash function based on the line and choose a wording
that makes the [let's say] seventeenth bit carry the stego bit
value?

So you fiddle with the wording, case variations, punctuation,
indentation, spacing or whatever on a line of text until the
seventeenth bit in the hash comes out right.

The advantage is that now you are no longer restricted to
transformations that change line length by an odd number
of characters.  You can now do arbitrary transformations
(still facing the problem that 50% of these transformations
are likely to leave you with an unchanged value in bit 17).

Another advangage is that you are now free to use either
more than or less than a single line of text per stego bit
and are free to encode more than one stego bit per
plaintext unit.

And you are no longer tied to printable text.

Of course, this is all a grade school exercise of
little or no practical value.

Maybe I gravely misunderstood you.

Possibly.

But why do you think that
using a hash function the fiddling is easier than without the
hash function?

My proposal is the obvious generalization of yours.

You: length(line) mod 2.
My example: bit 17 of md5(line)

kg has understood the intent clearly and has
provided a more formal presentation.

I already listed what I think are three advantages
of the general approach over your specific proposal.
Ease of use was not on the list.

1. A larger alphabet of subliminal transformations
that affect the hash.

However, one can make a strong argument that
the set of transformations that fail to preserve,
for instance, md5(line)[17] is roughly equinumerous
with the set that fails to preserve line length
modulo 2.

2. A configurable stego bit density that is not tied
to line length.

3. The ability to use overt plaintext that is not
printable text organized as a series of lines.

Anyway my scheme can be simply done by anyone, without requiring any
piece of software code:

Yes. This is true.
.

Relevant Pages

  • Re: A revision of my text stego scheme
    ... Using line length to encode your stego bits is clumsy. ... compute a hash function based on the line and choose a wording ... So you fiddle with the wording, case variations, punctuation, ... You can now do arbitrary transformations ...
    (sci.crypt)
  • Re: A revision of my text stego scheme
    ... compute a hash function based on the line and choose a wording ... that makes the seventeenth bit carry the stego bit ... So you fiddle with the wording, case variations, punctuation, ... You can now do arbitrary transformations ...
    (sci.crypt)