I can only tell you the way I did it, that was okay if you're not too demanding. I made what I called a "YaBa" anim which was the head model mouthing "yaba" over and over. Load the sound file using the head model with the anim.(uses less memory so your computer can almost keep up with the lips while you're editing. Make sure you're using Linear Interpolation. Play your sound track with the yaba anim, you'll be amazed at how much of the time it looks okay. Roll down to the first problem spot (say an f sound), add the lip movement from the library, go to the next until your done. Make small test renders of the head, go back and correct problems. Save anim, load character you're going to use, apply anim. When people say, "hey that was pretty good but what about when he says so-and so" - you mumble while barely moving your lips, "What? I don't see anything wrong".