In the stereo camera system, two cameras are needed and should be mounted side by side. I see someone just glues two cameras to a wooden board. However one mobile phone manufacture claimed that the two lens of dual camera modules on phones produced by them are parallel within 0.3 degree. Why do two lens on mobile phones need such high precise assembly? Does this will bring any benefit?
If you glue two cameras to a wooden board, and then write your code with the assumption that your translation matrix is perfectly or really close to [1, 0, 0], once the cameras are slightly displaced, depth errors start creeping up in your stereo applications such as reconstruction. In mobile phones, the drivers and the rest of the API code would be preprogrammed, and most probably cannot be changed later on: which means the geometry needs to be described as accurately as possible.
Also, as you go smaller on your baseline, which is the case for mobile phone cameras, even tiny imprecisions in the setup of the cameras can cause a huge difference in the geometry and depth computations. On the contrary, you can probably get away with a few centimeters of movement on the Z axis when your cameras' baseline is 1 m on the X axis.