Stereo Camera Solver

Note: this node is part of the Stereo toolset available with Studio or Enterprise editions

The Stereo Camera Solver node can be used to estimate camera motion and stereo parameters using a set of feature tracks generated by one or more Stereo Auto Track, User Track or Stereo User Track nodes.

The Stereo Camera Solver node has two inputs and two outputs. Individual trackers must be been tracked in both left and right-eye input clips before the stereo camera motion can be solved.

Usage

Most controls and parameters in the Stereo Camera Solver node are described in the documentation for the Camera Solver node, so please refer to that section for further details.

The stereo camera can be solved using two methods: When Separate stereo solve is enabled, the stereo parameters will be estimated after the primary camera motion has been found. Otherwise, the stereo parameters will be estimated during the actual camera solve. This can often give improved results because it allows tracker positions to be estimated more accurately.

The stereo camera model supports constant or variable interocular and convergence distances, as well as fixed offsets between the left and right-eye views. Vertical and depth offsets correspond to having the right-eye camera offset by a constant value along the Y (up) or Z (forward) directions of the left-eye camera. Pitch and roll offsets correspond to rotation around the X (right) or Z (forward) directions of the left-eye camera. A range of convergence distances can also be specified when convergence is allowed to vary throughout the clip.

Controls

The features specific to the Stereo Camera Solver node are as follows:

Separate stereo solve: When enabled, the stereo parameters will be estimated after the motion of the left-eye camera has been calculated. Otherwise, the stereo parameters will be estimated during the camera solve itself, which can often increase the accuracy of the solution.

Interocular distance: The distance between the left and right-eye cameras. Interocular distance can be set to Known, Unknown or Initialised. An initialised distance will be limited to lie between the minimum and maximum values of the Interocular range.

Interocular range: The minimum and maximum allowable interocular distances. Enabling Variable will allow the interocular distance to vary throughout the clip.

Vertical offset: The vertical offset between the centre of projection of the left and right-eye cameras. The vertical offset corresponds to a distance along the Y (up) axis of the left-eye camera.

Depth offset: The depth offset between the centre of projection of the left and right-eye cameras. The depth offset corresponds to a distance along the Z (forward) axis of the left-eye camera.

Pitch offset: The rotation angle between forward axes of the left and right-eye cameras. This corresponds to a rotation around the horizontal axis of the left-eye camera.

Roll offset: The rotation angle between the up axes of the left and right-eye cameras. This corresponds to a rotation around the vertical axis of the left-eye camera.

Convergence distance: The convergence distance between the cameras. Convergence can be set to Known, Unknown, Initialised or Parallel. An initialised distance will be limited to lie between the minimum and maximum values of the Convergence range. Parallel convergence corresponds to the cameras having parallel forward axes that never converge.

Convergence range: The minimum and maximum allowable convergence distances. Enabling Variable will allow the convergence distance to vary throughout the clip.

Default Keyboard Shortcuts

Default keyboard shortcuts are listed in the Keyboard and mouse section for the Camera Solver node.