Real-time sound cyphering algorithm

Hello,
My problem is quite simple:
I'am willing to find a simple algorithm to cypher a sound:
I need the cyphered sound to be audible, but not understandable,
and that we can decypher it from a recorded sound (verbatim copy not needed).

This seems simple, but in fact, i tried a few algorithms and i had no results.
The real problem is that the input data is a 8-bits encoded buffer
(or more, depending on the encoding).
It represents the movement of the speaker membrane.

I tried classical cypher or modifications on the stream:
  -XOR-ing it with a 1 or 4 bytes key. It is or understandable with the 1 byte
keys, or impossible to decypher (because we may not have the same values
due to numerical/analogical conversion, and that we need to know how to sync
the key with the data)
  -permutation of amplitude ranges (still understandable)
  -modifying harmonics after a fast fourier transform. Increasing the level
of each harmonic (still understandable but strong noise), multiplying all
harmonics by a factor (inverse fourier transform goes out of sound range).

The objective is to have an output sound that seems to be a random sound.
And that we can decypher after recording. "real-time" is achevied by a small
buffer (0.5 sec for example).

Thank you.
Received on Thu Sep 29 21:38:34 2005