You mean ADR? It's very possible if you have someone who is audio savvy. There's a program out there called Vocalign used for that very reason. (If I understand your post correctly that is.)
I'm not sure this is the right place to post this question, because it is kind of an unusual idea. But it does involve sound editing in a certain sense. What I want to know is if it's possible to record several minutes of voice, and then to match it up with video of someone saying the same parts but without the audio. In other words take audio recorded separately under different circumstances and apply it to video recorded at a different time. The reason I want to do this is I need to record someone speaking under certain conditions, which are not quite suitable for video recording. Later, this same person would basically be giving the same exact speech, although there would be slight variations of course, such as the timing. I know this kind of sound like a strange idea, but I just wondered if there was a software program that can analyze the audio and video and figure out a way to match them up.
Yes, that's pretty much what I was describing. I should've known that this was probably a pretty common thing in the film industry. I checked out Vocalign, but from the information available on their webpage, I'm not sure if this would actually do the trick. Apparently it tries to sync up two separate audio tracks. However, that doesn't seem to take account of what activity is going on in the video such as lip movement or other facial expressions. Maybe it does, but from what I read it didn't seem to really mention much about using it in video postproduction.
A different idea I thought of which is somewhat related to this topic, is to video record a series of mouth and lip movements that correspond to different syllables or other basic speech units. Then, have a program match up these video clips with an audio track. When it comes to a certain type of speech unit, it would find the corresponding video clip. So, you record the necessary video clips for all of the necessary speech units. That of course would involve repeating most of these clips several times. Is this completely hypothetical, or has this actually been done?
The idea with vocalign is that your location audio has been degraded and/or compromised. You then take the crappy audio dialogue and get your actor to re-record over it with good dialogue (in a studio). Vocalign will then match them both up so when you put the good audio back in, it matches the dialogue on the screen. (mouth movements as well obviously)
I'm not so sure about the second part of your post. Seems pretty complicated for what you're doing... What are you doing anyway?? :P
The way to approach this might be to use a program which manipulates a still photo, based on the sounds on a recording. Take a look at CrazyTalk CrazyTalk - Facial Animation Creator
It's designed for animating drawings but might have limited use on photos.
It sounds more like an animation project to me, like a stop motion thing. If you think of something like Sledgehammer by peter gabriel.
Midnight Blue Productions M.M.Inst.V.