In this article, the author reviews the history of singing voice synthesis and highlights several current (1996) projects in the field. He discusses the benefits and trade-offs of using techniques and models, the attraction of composing with vocal models, and suggests directions for future research.