QUOTE(kennedyb4 @ Mar 29 2003 - 03:44 PM)
I played around with a programme called scramdisk once when I was om a cryptography kick.
It only worked with wav files.
Oh. I tried it once, but quitted because of that 2GB limit (even though that was on NTFS).
They use wav files because it's easiest to read and write many times to/from them. With mp3, it's difficult to replace hidden data once it's put there (the psychoacoustics are not available anymore, etc). So, when you only need to write data once, you can use pretty much any audio/video/picture format

QUOTE
You could choose to use 25% or 50% of the wav file for including encrypted data, and the file would still play OK in a portable. I don't know what effect all this would have on the sound quality though.
The problem in altering lower bits of PCM samples, is not only with sound quality. In fact, if you use 8 bits (in a 16-bit wav file) for hiding data, it's pretty easy to determine (using statistics) that the lower 8 bits behave like pure white noise - and thus that it's unlikely to contain real audio data there. Remember, current ADC's are better than 8-bit. I'd even doubt that the lower 4 bits of a 16-bit converter is pure white noise. So, there goes the "25%" Scramdisk/E4M mode...
QUOTE
The data was encrypted with an algo called "blowfish" It was reported to be "hard" encryption, similar in difficulty to PGP.
You can't compare blowfish with PGP - the former being a (max. 448-bit) symmetric encryption algorithm, and the latter being a PKI solution - that includes both symmetric (ie: private-key) and asymmetric (ie: public-key) algorithms.
Anyhow, as you said, blowfish is believed to be safe - even though I'd prefer twofish a bit (Twofish was also made by Bruce Schneier, but later).