Unified Speech Recognition: A Single Model for Auditory, Visual, and Audiovisual Inputs