Stereoscopic 3D presentation techniques need to deliver both eyes' views separately. To this end, many stereoscopic 3D presentation methods employ interlacing techniques. In temporal interlacing, left-and right-eye views are temporally interleaved: the left-eye image is shown during one period and the right-eye image is shown during other periods. In spatial interlacing, the two eyes' views are presented on every other row of the screen, e.g. the left-eye view on odd-numbered rows and the right-eye view on even-numbered rows.
It was controversial whether these methods have the same effective resolution. We performed a psychophysical experiment that compared the two presentation methods (temporal interlacing S3D and spatial interlacing S3D) in terms of effective resolution. Six subjects participated in the experiment, and the task was a letter recognition task at various viewing distances and letter sizes. The figure below summarizes the experimental results from six subjects. Each panel plots each subject's correct ratio as a function of letter size. Green, red, and blue respectively represent viewing distances of 6, 3, and 1.5 times the picture height; solid and dashed lines respectively represent temporal interlacing and spatial interlacing protocols. The letter size is described in terms of the visual angle, meaning that the same horizontal location for different viewing distances corresponds to the same retinal size. If the performance curve remains closer to 1 at smaller letter sizes, it means the corresponding viewing condition has a higher effective resolution. At the recommended viewing distance for HD TV (3 times the picture height) and shorter viewing distances, temporal interlacing had higher effective resolution than spatial interlacing. It was only at sufficiently long viewing distances that the two presentation methods did not show a significant difference in effective resolution.
Collaborator: Martin Banks