The apparent velocity of a filmed person, walking in front of static or moving backgrounds, was estimated in 2 experiments by 18 observers. The camera either followed the walker or remained at the same position (= stabilized vs. mobile observer perspective). A factorial ANOVA was used with the estimate of the walker’s velocity (in km/h) as dependent variable. Based on the number of applicable motion cues and on the role of motion parallax, it was predicted that the mobile observer perspective should lead to a higher estimate of the walker’s velocity. In both experiments, the opposite of this prediction was observed: Stabilized observer perspective produced consistently higher velocity estimates as a main effect and in interaction with the background variables. No velocity increasing effect of motion parallax was found in stabilized observer perspective, presumably because of the ambiguity of motion cues with respect to background distance.