A computer-implemented method of configuring a virtual camera. A first and second object in a scene are detected, each object having at least one motion attribute. An interaction point in the scene is determined based on the motion attributes of the first and second objects. A shape envelope of the first and second objects is determined, the shape envelope including an area corresponding to the first and second objects at the determined interaction point. The virtual camera is configured based on the determined shape envelope to capture, in a field of view of the virtual camera, the first and second objects.