Our synthesis model can perform explicit manipulation of style by changing the style code associated to each object. In the middle of the video, we interpolate between different style codes to produce videos with different players and tennis fields.