I have partly worked it out - I used a single still picture from file to characterize rather than my own images captured on my video camera. It is because I wanted to characterize a friend not myself. However, if you do that it does not go through the whole set of samples it takes when you characterise youself on your video camera - a talking guide takes you through all the mouth shapes etc. that you need to do for lip sync - but if you just characterize from a still image on file (an option in the drop down menu) it cannot get the mouth shapes and their sounds. Characterize does work for lip sync when you use the camera and audio inputs from your own video feed.
So the question is - is there a way to lip sync when characterizing a still image of a friend, not yourself on the video?
Ummm (which is a hint I have never done this before lol!), if I was doing it then I would say "thank you Characterizer, I will take it from here" and take over adding the extra functionality to the puppet. Characterizer has all the stylization etc. If you only have a single image, I would ...
- take the artwork, I assume its photoshop so I would select the region around the mouth and copy it into a new layer.
- I would then wipe out the mouth on the face (e.g. smudge the surrounding area, or fill it with skin color or something - not too important). This does not have to be very accurate as you are going to put the mouth shape for different sounds in front. The problem is "Oh" with a narrow mouth is going to be smaller, so you don't want the old mouth poking out from behind.
- I would then inside the Head create a Mouth group (is there one already?) and put the copied mouth into that group and call it Neutral.
- I would then clone the Neutral layer and call it "Ah" and use Photoshop to morph it to look like the mouth open using warps etc. (I am not very good with photoshop, but there are various transforms / warps // distortions you can do - I search YouTube for tutorials on this).
- Repeat for all the visemes, doing the best you can. It is fine to duplicate some of the shapes for multiple sounds, at least to get going.
If you want to just get going for quick fun (I would!!!), import one of the default CH puppets that uses a PSD file and copy the mouth out of it and just use it. You will need to resize and reposition it. (Make sure you un-hide all the mouth child layers before moving and resizing to save yourself a lot of pain.) You can then try to get it to work that way as a crude test so you know Lipsync etc works, then follow the instructions above with more confidence your artistics efforts will not be in vain.
AdChoices