For my full-body animation I'm going to animate after a clip from Shaun of the Dead, where Shaun is speaking in front of the other employees at the place he's working. You can listen to the sound clip below.
"As Mr. Sloman always says, there's no 'I' in team but there is an 'I' in pie. In... There's an 'I' in meat pie. The anagram of meat is team... I don't know what he's talking about. Look, that's it."
AS MSDS-SLOMAN-OVES-SES, DES-NO AI IN TIM BUD-ER-IS-EN-AI IN-PAI. IN... DES-EN-AI IN-MIT-PAI. DE-ANAGRAM-OV MIT-IS-TIM... ADONOISTOKIBA. LUK, DAS-IT.
(Accents marked in red. This sound text actually makes sense when you've listened to the sound clip for a while)
When animating the lip-sync, I'm trying to imaging the clip isn't a language but just some sound made with the mouth - looking away from the actual letters inside the words and just concentrate on the actual sound I'm hearing. This method seems to work out the best for me.