Code (Text): X________________ moving | observer | | ---> v (moving speed train) photon | ^________________| X________________ ^ moving | ^ ^ observer | d1 ^ ^ | ---> v ^ ^d2 | ^ ^________________| --- standing observers see d1, passengers see d2 However, if speed of light is to be constant in each and every reference frame and independent of motion of the light source, which means no horizontal momentum for this photon, then it must miss the target. If it was a bullet then it would have this horizontal momentum and it would hit the target, but photon, as fast as it is, actually has to miss it, right? How can it "know" it has to move horizontally (as well) if it can not be impacted with this sideways momentum like a bullet would? Solution, then... Code (Text): X________________ moving | observer | | ---> v photon | ^________________| X________________ ^ moving | ^ observer | ^d1 | ---> v ^ d1=d2 | ^ . _______________| --- standing observers see d1, passengers see d1 ...and so the speed of light is constant and same in all frames. Except that this looks a bit strange, so is this really what experiments measure? Is there any such experiment that tries to measure some lateral momentum of photons like in the scenario above? What is the speed of light relative to?