Abstract
The angular expansion hypothesis has observed that ground distances are misperceived in a manner consistent with an exaggerated scaling of angular declination (elevation) with a gain of 1.5 and a smaller, but still exaggerated (~1.25) scaling of perceived azimuth direction relative to straight ahead. In the present study we sought to measure whether the explicit estimation of angular direction occurs with respect to the reference frame defined by a visible ground plane (even for sideways observers) or the reference frame defined by the body (e.g., for a sideways observer). In a series of three experiments we tested observers in a stereoscopic virtual environment (back-projection screen) with a FOV extending 70° to the right from straight ahead in azimuth (each direction probe was presented along this physical axis). Upright or sideways-lying participants made angular direction judgments (in degrees) with respect to a ball suspended in space at a distance of 8 m along the same (gravitational azimuth) axis. Directions ranged from 3 to 51°, presented in random order. The straight-ahead was represented by a distant pole in the upright-world conditions, and by the rotated horizon in the sideways-world conditions. Both upright observers and sideways observers showed a gain of 1.3 in visual-world azimuth with a (gravitationally) upright world. When the world was gravitationally sideways (so that judgments were of angular elevation above a depicted ground plane), sideways observers showed a gain of about 1.5, consistent with the expected scaling for elevation. This gain was reliably higher than their 1.3 gain for azimuth in the gravitationally-upright world. These results for explicit estimation of angular direction are consistent with prior studies of perceived distances and heights for sideways observers that also suggest the dominance of the visual reference frame (Li & Durgin, 2016, JOV 16(1):4; Klein, Li & Durgin, 2016, JEP:HPP, 42, 581-593) .
Meeting abstract presented at VSS 2018