In the toed-in stereoscopic camera arrangement above the camera axes are set to verge at a point in the scene. Objects in the same plane as the vergence point, such as the yellow square here, appear in the image plane of the final display if no other adjustments are made. In the image seen on the final display other objects appear in-front or behind the image plane depending whether they were in-front (blue square) or behind (green square) the convergence point.
The amount of image disparity (3D depth) captured depends primarily on the camera separation, but is also directly affected by the camera zoom setting, the vergence distance and the object sizes and distances from the convergence point.
The drawback of the method is indicated by the cameras in the diagram above. The film plane at the back of each camera is shown, notice that the film planes are not parallel and that they are rotated in opposite directions. This results in keystone distortions in the left and right images, in particular different height distortions that introduce artificial vertical disparities which can become uncomfortable to view. Also illustrated in stereo in the anaglyph image below you will see that an artificial curvature across the image is introduced by the vertical disparity. The solution is to correct these distortions out in post-production using tools such as Stereophoto-maker's auto-alignment option.
The drawbacks of the toed-in method are often accepted for natural imaging where the physical bulk of photographic and video cameras means it is difficult to bring them close together. In this case verging the camera axes is the only way to ensure the whole field of view of the camera is used. Also because of the difficulties aligning any physical camera system it is routine to use auto-alignment software tools to post-correct geometric and luminance errors.