I don't even think plausible deniability would hold in court -- claiming that a large blob of random data on your hard drive is just there for no reason at all is not plausible.
I don't even think plausible deniability would hold in court -- claiming that a large blob of random data on your hard drive is just there for no reason at all is not plausible.
What I don't understand is that in a context of a court (and this group of competent professionals), password disclosure _should_ be considered self-incrimination (although there was at least one case in the UK where a judge came up with some loophole reasoning around that). Disclosure of multiple passwords ("we didn't like what we found, do you have any other passwords?") would certainly be obtained under great duress.
To make sure that you can't distinguish free space from encrypted noise, you have to write random noise everywhere as part of the filesystem creation process.
The one thing Truecrypt is vulnerable to is that you can note what parts changed -- say they raid your house twice and image it between when you used it. Then they'll know that free space isn't really free.
That said -- I would think that a random blob of data sitting around on the hard drive is still highly suspect. Aren't hard drives zeroed from the factory? And wouldn't any true "garbage" data be decidedly not random? (Even if it's compressed... you would still expect to find headers etc. somewhere.)
Take a look at http://www.truecrypt.org/docs/ - it is an interesting read.
When the police demand the key from you, you give them the one that unlocks the clean partition. Now, at this point it doesn't matter if they don't believe you, it doesn't matter if they know all about truecrypt and hidden partitions, there is no way for them to prove in a court of law or otherwise that there is a hidden partition there. You can just keep telling them "I gave you the password! I just wipe my free space with noise every night! It's just noise!" and you have plausible deniability.
As far as I'm aware this is only in theory. I'm not aware of any case of this actually being tested in court. But mathematically, it is apparently sound.
http://www.schneier.com/blog/archives/2009/10/evil_maid_atta...