A method of synchronizing an animation sequence with a video includes placing a virtual camera in a 3D model of a site captured in the video at the same location and the same orientation as those of a video camera that captured the video, generating a set of virtual frames of the animation sequence by projecting the 3D model onto a scene frame from a viewpoint of the virtual camera, for each of a plurality of virtual key frames, placing a virtual object corresponding to a moving object captured in the video in the 3D model at a respective location that matches with the respective position of the moving object in a corresponding key frame of the video, and playing the set of virtual frames by stepping through time such that the plurality of virtual key frames is in synch with the plurality of key frames of the video.