I have video files and I want to edit the sound so you can interpret the tone of a voice but the words are uninterpretable. Is there a way to accomplish this?
Most of the definition in a human voice is usually between 2kHz (attack) and 4kHz (presence/diction).
If you need almost no definition bring the filter down to where most of the speech is presented, which in people is between 60Hz and 255Hz. If you low pass that low, I'd be shocked if you can hear any definition at all.
If you want slightly more, bring the low pass up; maybe to 1kHz or some such.
Low passing is your friend in this case.
You might also try applying some sort of smearing (reverb, delay, ...) to maintain the tonal character of the voice but make the shape of the words less intelligible. Or use multiband compression to mess with the consonants to a point where the words are unintelligible.