Investigation of 3D projection for qualitative evaluation of daylit spaces

This paper presents the results of a study that investigates 3D projection for lighting quality evaluations. Thirty-nine participants forming groups of three visited eight daylit rooms and rated them through seven-grade bipolar scales. During the visit of each room, a series of pictures were taken at different exposure time. Then, they were combined to reconstitute HDR pictures capturing scene luminances. A global tone-mapping operator mimicking the human visual system was then applied to the resulting HDR pictures to compress the range of real world luminances to that of the projector. In the second place, participants rated through the same questionnaire the pictures projected in 2D and in 3D. Finally, they chose which kind of projection was the more realistic. Comparison of subjective ratings issued from the visit of real spaces, and from 2D and 3D projections of these scenes suggests that projected tone-mapped pictures can be used for lighting evaluation of bright scenes. Dim scenes whose luminances are much lower than those available with the projection presented significant differences of scores. In addition, even if some scenes are judged more realistic in 3D than in 2D, subjective ratings are not significantly different according to the kind of projection.