The StereoDepth node rectifies the left and right images before performing stereo matching. Does that mean that the depth/disparity at a given pixel (x,y) actually corresponds to the pixel (x,y) in the rectified image (left or right depending on which camera the depth/disparity is aligned to) and not the raw image?
In other words, is it technically wrong to use the depth/disparity map with the raw image (because distortion, etc. would put each depth at a slightly different place)?