This paper provides a detailed and complete description of merging virtual actors with animation in a real environment. It describes the tasks involved in each stage of integration, such as video acquisition, extraction of camera parameters, creation and animation of virtual actors, and rendering the final images. The most important problems are discussed: real objects hidden by virtual actors and vice versa, collision detection between the virtual actor and the real environment, correspondence between real and virtual cameras, and casting shadows of the virtual actors on the real world. Case studies are presented, such as the virtual actress Marilyn walking with real people on a real street or sitting down on a real chair