>once the microphone is clipping the job is done Try adding a loud ultrasonic sine wave to speech in your favorite audio editor, and then hard clip it. Most of

>once the microphone is clipping the job is done Try adding a loud ultrasonic sine wave to speech in your favorite audio editor, and then hard clip it. Most of the distortion can be filtered out without too much damage to the speech by looking at the spectrogram and notch filtering unwanted frequencies. I imagine modern AI could do the same thing automatically. But high-pass filtered white noise + clipping will make the speech unrecoverable if it's loud enough.