Speech based LipSync implementation using blend shapes
Minimal LipSync (Redux)
by Chippalrus
Speech based LipSync implementation for Unity
Uses 4 blend shapes. Works with Microphone Input and AudioClips. Can work for 3D models. It isn't perfect or very precise, but for cartoon style characters this works pretty well.
Only 2D Example is provided, but same method and setup can work for 3D models.