For stereo cameras on the market, two cameras are always mounted side by side and with a displacement that is perpendicular to the cameras’ optical axes. I take this setup for granted. One idea came to my mind whether this is necessary? If two cameras are not parallel and have different focal length, camera calibration can correct the difference. Why are two cameras mounted in parallel? My guess is that two cameras can have a large overlapping region. Am I correct?
Each camera needs to be defined by 6 variables (3 position, 3 orientation). This would mean that during the calibration process, a solver needs to find 12 variables. As this is done usually with an nonlinear optimization process, the solutions are quite sensitive to the initial guess. By making them parallel and giving them a fixed width, you can give the solver a much nicer initial guess (which can converge to the correct solution).
Aside from this of course, we do get large overlapping areas which become important when creating depth maps.
Making them parallel is beneficial for reducing distortion after a rectification. We usually rectify two images for a fast matching. If speed is not your concern you can skip the rectification stage.