Source Filmmaker > 一般的な話題 > トピックの詳細
Nookyava 2013年5月15日 15時26分
Lip Syncing
How can I make it smoother?

http://www.youtube.com/watch?v=KrB3opqzz3U
1-5 / 5 のコメントを表示
< >
raptornx01 2013年5月15日 16時17分 
Good rule of thumb is to have a visual happen a few frames (2 or 3) before the sound. so have each mouth position happen 2 or so frames before the sound in the word. also, I found the best way to do it for me is to mouth the words to yourself and pay attention to the mouth positions.

here's an example I did some time ago, the model is an older, less articulate model, and the overall mouth shape is different, but same process.

http://youtu.be/Tt8O1Ju5c04

I used keyframes in the graph editor, generally in groups of three. I also overlapped so the next mouth shape would start to form before the last one is finished.

so for example (your mileage may vary) to make a "to" you need two positions, a grin or one that shows the teeth together and an "o" shape. say the "t" starts on frame 1 and ends on frame 7. so the middle keyframe would be frame 4. (IE you put a keyframe at frame 1 and 7 then put the head on frame 4 and slide the slider) then for the "o" I might put the start at frame 5 and the end on frame 11, stick the playhead on frame 8 and move the slider for that.
Nookyava 2013年5月15日 16時38分 
So you manually did a lip-sync instead of the quick easy one.
raptornx01 2013年5月15日 16時50分 
I did it manually, yes. You can extract and tweak it to get the same result, but the process is pretty much the same. I believe you just have to make sure it's set to offset mode. though i'd test it to make sure.
Nookyava 2013年5月16日 7時29分 
Hrm alright, I'll have to try offset mode. Here's a :D
Meccanica 2013年5月17日 1時38分 
I prefer the motion editor for lip sync so far, rather than keyframe editor.
But yes, the answer is to do it manually. Using curved falloff in motion editor, spline or flat tangents in keyframe view, and possibly very light application of the 'smooth' procedural.
1-5 / 5 のコメントを表示
< >
ページ毎: 15 30 50
投稿日: 2013年5月15日 15時26分
投稿数: 5