What if the resulting audio output of the digitally converted human vocalization has an inaudible frequency modulation or signature added to it which operates in a range inaudible to the human ear, you could then detect said signal and filtered out its source or merged / manipulated

Now fixing search, that would be truly mind blowing.