15.07.2013 05:29, Bruno Cartusia: > you might use a voice distorter A child molester modified pictures by swirling his face.
He thought he would be safe and that his face could not be revealed from the distorted image. Turns out he was not safe (luckily in this case). Law enforcement was able to recover his face from those pictures. I don't know how easy it was to undo the effect. Also depending on what information is lost. If you know what an algorithm outputs for a given input it might be easy to reverse the effect. Voice analysis appears to be successful to me. Bomb threads and kidnapping have been resolved by analyzing how one pronounces different words or even single letters. German law enforcement was able to tell that a caller grew up in a specific state of Germany, just because how he formed different letters with his mouth. Also, like adrelanos, I think that someone that knows how you say things would be able to recognize you, based on words you like to use, phrases you use, words that are not all that common except from your environment, how loud you speak, with what intensity, how you make pauses, where you make pauses, changes in momentum and speed of your speaking. > simply use text-to-speech software to do the talking for you. Anonymous does you text-to-speech in their video, don't they? There are a lot less characteristics to fingerprint on. Best, bastik _______________________________________________ tor-talk mailing list [email protected] https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk
