Workflow

    back to services >
  • Prior to shoot
    Gausst calibrates camera rigs and lenses to align the computer graphics world coordinate system to the real world. Then we import the CG elements and make them work in real-time. The CG elements can be anything from 2d photographs, Photoshop layers and 2d moving background plates to fully 3d environments and animated CG characters.
  • On set edit
    The camera tracking data is driving the camera either in Maya or Viz Virtual Set. (Maya for flexibility and Viz Virtual Set for broadcast quality render). The camera data is always real-time and is independent of how many polygons your scene has. Viz Virtual Set uses external chroma keyer while in Maya live video is fed into the foreground layer where it is being keyed with the set (internal chroma or linear key). The synchronized video/CG composite image can be viewed directly on the computer screen or on various monitors that are fed with the computer image. Editing of the scene is possible from any view (including the camera view) in order to fine-tune object positions or animations with relation to a real actor or real set (set extensions). For example the 3D artist can use 4 views. In one view he will have the composite image. In the other views (top, side, front) he might edit object positions or even lighting to match the set with the real world better. The director can use the composite image to position the actors, the actor’s eye lines, and the actor’s movement. Or the director of photography will view the composite image and adjust the camera view or even lights on the set to match the 3D word.
  • Record
    During last rehearsals and the actual takes, the system moves into “Record mode”. The data flow is very much the same as in edit mode and the only changes are:
    1. As soon as the take starts the system operator presses the record button. The "Camera server" starts recording the camera data (logged with the correct time code) and the "Renderer" starts to record the video image. At the end of the take, when the system operator presses the save button, a camera data file is saved. The camera files are logged into the file system according to day/scene/shot/take.
    2. The composite video image is distributed to as many video monitors as needed so director and actors (using "Mirror monitors") can view the composite image. The image can be fed into the camera operator's viewfinder to mix with the real view so the cameraman can view the composite image as well.
  • Playback
    As soon as the take is over the shot can be viewed in Maya, here the data flow changes. The recorded camera file is loaded into Maya and crates camera animation key frames in Maya. In addition the “Foreground layer” in Maya is switching to use the recorded video data. Since the camera file contains the time code for every camera sample as you move along the Maya timeline, the correct Foreground image is displayed and again keyed with the 3D set. The big advantage is that the composite image consists of 2 channels, the foreground channel and the 3D channel, so if for example the actor looks in the wrong direction (for a 3D character) and the director might think he needs another take, the 3D artist can quickly edit the animation and run the scene again, with the change, to see if it works. Loading the recorded video data into Maya is not real-time, so you cannot run the scene in real-time during playback, Maya is interpolating both the video animation and the camera data animation in sync. Therefore every frame you view is “correct” in the way that the correct foreground image is displayed/synchronized with the correct camera data frame, even though Maya is not performing in real time. If the director wants to view the take in real-time all you have to do is to run a play blast and in a few minutes you can have a short movie file with the composite image running in true frame rate.
  • Post-Production
    Gausst records the camera data synchronized to time-code, the recorded video data is slated with time code. Once Post-Production gets the camera data files and the recorded video files (One-Light Telecine, Low-Res Video) the match-moving of each take is very simple. The camera data is loaded into a Maya camera and the video is imported as an image plane onto that same Maya camera, matched to the time code and known film gate. Since the director already saw the composite image on set he/she might already have selected the preferred takes so the post-facility can concentrate on them and the correct Maya scenes can be rendered. As soon as the film is developed and the chosen takes are transferred they can replace the original video image for better image quality. In less then a few days, a high-resolution 3D set can be composited with a high-resolution foreground.
    • back to services >