Three-dimensional video is presented on a two-dimensional display. The view of the video is driven by user interaction. In one example, a two-dimensional (2D) video sequence in real time with a first view is synthesized from a video image sequence with depth captured by a first device. The 2D video sequence with the first view is provided to a second device in real time. A command is received from the second device to change a view of the 2D video sequence as the video sequence is rendered. The view of the 2D video sequence is changed as it is rendered to a 2D video sequence with a second view, and the 2D video sequence with the second view is provided to the second device in real time.