I present an audio processing algorithm that modifies how a human voice recording sounds, without changing the pitch or speed of the speech, so that the identity of the person speaking is obscured, but the speech can still be understood.
In my implementation, I show how applying amplitude modulation, a “robot” effect based on a fast delay, high-pass filtering and echo/reverb, the resulting audio can effectively achieve this goal.
I present the methodology, results and discussion in the report (report.pdf).