The speech synthesizer triggers an OOO sound when the program sees the mouth of the person on the video is shaped to form an OOO. EEEE, AAAA, III etc.
This synthesized correlation of mouth shape with sound that mouth shape would make is played instead of what the person is actually saying resulting in them saying stuff like "Oooo aaaa eeee oo e ii ooo uuu weeee eeeou waaaeee".
You could make an art exhibit of videos of famous people from history making random weird noises, or have a button on your TV to have this activate whenever a video of a politician or a commercial you've seen 100 times comes on the screen.
It would basically make the person on the video look like they've gone nuts and are making random mouth noises.-- doctorremulac3, Jul 17 2020 Good lip reading example https://www.youtube...tch?v=wg3upHE8qJw&tFunny you post this today. [tatterdemalion, Jul 17 2020] // make the person on the video look like they've gone nuts and are making random mouth noises. //
Baked. Just about any elected official.-- 8th of 7, Jul 17 2020 //Couldn’t you just mute the volume and make the noises yourself?//
Wow, great idea.
But seriously, work done on this might be a step towards creating a HAL 9000 style actual lip reading system.-- doctorremulac3, Jul 17 2020 <HAl>
"I'm sorry [doc], I can't do that. And I think you know why ..."
</HAL>-- 8th of 7, Jul 17 2020 Two minute papers has a new video out today about recent improvements in AI lip reading. Yes I know this is opposite of your idea, but nonethewhich interesting. Place a big knob labeled "good" and "bad" on the screen and it would accomplish your task.-- tatterdemalion, Jul 17 2020 // Place a big knob ... on the screen //
Not Monsieur Macron then, because he's quite small...-- 8th of 7, Jul 18 2020 (tatter, re link) No way!! That's friggin' amazing!
Re: sniper bone. Any explanation of why?-- doctorremulac3, Jul 18 2020 random, halfbakery