A common concept used in the explanation of the twin paradox is that upon acceleration, one twin sees the other twin's time speed up. Is this correct? Assuming it is, how would they observe a light clock next to that other twin? The light beam can't bounce faster, right? But (assuming a vertical orientation of a light clock) the clock can't get shorter either, since contraction is only allowed in the direction of the velocity, correct?