Abstract: Three-dimensional facial animation has been extensively studied, but the achievement of realistic, human-like performance has not yet been decided. This article discusses various approaches for generating animated facial expressions controlled by speech. Combining the considered approaches for both facial animation, and the identification of emotions and the creation of micro-facial expressions in one system, we get a solution suitable for tasks such as game video, avatars of virtual reality or any scenario in which a speaker, speech or language is not known in advance.