Any matchmoving software needs horizontal and vertical field of view. This is calculated by using the filmback dimensions of the camera and the focal length of the lens. I'm using a 700D which has a sensor with dimensions of 22.3 x 14.9 mm (3:2) and an 18 mm lens. When I record full HD I know that the camera uses full width of the sensor (22.3 mm) and based on an aspect ratio of 16:9 (1.78) I can obtain the height of 12.544 mm and everything is cool. The problem occurs when I record raw at a reduced resolution of 1600 x 680 (2.35:1) because ML is not using the full width of the sensor. I need to find out what width of the sensor is actually employed by ML (in mm or %) when recording at this resolution.
Thanks for your help.