Creating Talking head using TBS

Lately I’ve been experimenting with the lip-sync feature to automate certain processes. I want to use the lip-sync feature to create a talking head, newsreader on TV and the like. It looks like it is possible and save a lot of time once the initial preparations are made.

I’ve posted a sample at my site along with the templates i created.

Drawing preparation

The whole process is actually simple, i took out the Roger figure from TBS site and created 8 images with different mouth shapes. And then created another 8 images by slightly tilting the head towards the center. These two sets of images are flipped horizontally and i ended up with 4 templates with 8 images in each of them.

Lip- Synching

Import any sound track and the 4 templates into 4 drawing elements. Enable the thumbnail and show lip-sync. Right click on sound element and select Modify lip-sync mapping, and on the pop-up screen, map to any of the drawing element and repeat the process for all drawing elements.

Now that all the 4 drawing elements are synched, we are free to cut paste frames into a new drawing element. To do this, add an empty drawing element. Cut about 12 frames from left facing element, and then another 12 frames from center facing elements and so on. What we are doing is picking frames randomly and creating a new set of element which is going to be the final output. Once you’ve filled the new drawing element, you can delete all other 4 drawing elements. Care should be taken while cutting the frames, that is go from left to right progressively to have smooth transition.

Instead of 4 elements, one can create 6 or more including up-down movements and then some for eye expressions. Once this set is prepared, it can be repeatedly used with different sound tracks for the same character(Kent Brockman from Simpsons comes to mind).

C & C are welcome


PS: Just noticed this thread;action=display;threadid=3088

where Rob Campbell explained this technique clearly.