Unconfounding the direction of motion in depth, time to passage and rotation rate of an approaching object.

Observers were presented with a set of 216 simulated approaching textured baseballs in random order. In Experiment 1 each had a different combination of time to passage (TTP), direction of motion in depth (dMID) in the vertical plane and total change in angular size (Deltatheta). In Experiments 2 and 3 each had a different combination of TTP, dMID and rate of ball rotation (RR). When required to discriminate TTP and dMID in separate experimental blocks for a non-rotating baseball (Experiment 1), observers could not discriminate dMID independently of variations in TTP but instead showed a bias towards perceiving objects approaching on a trajectory close to the nose as having a shorter TTP than objects approaching on a trajectory that would miss the face. When required to discriminate TTP, dMID and RR in separate experimental blocks (Experiment 2), TTP judgments were again influenced by dMID but could be made independently of RR. Judgments of the relative dMID were affected by variations in RR and rotation direction: for simulated overspin the (i.e., the top of the ball spins towards the observer) perceived ball trajectory was biased towards the ground whereas for simulated underspin the perceived ball trajectory was biased towards the sky. RR could be discriminated independently of both TTP and dMID. When required to make all three of these judgments simultaneously on each trial (Experiment 3) discrimination thresholds were not appreciably different from those found in Experiment 2. We conclude that TTP, dMID and RR can be estimated in parallel but not completely independently within the human visual system.