The effect of perspective on presence and space perception

Abstract

In this paper we report two experiments in which the effect of perspective projection on presence and space perception was investigated. In Experiment 1, participants were asked to score a presence questionnaire when looking at a virtual classroom. We manipulated the vantage point, the viewing mode (binocular versus monocular viewing), the display device/screen size (projector versus TV) and the center of projection. At the end of each session of Experiment 1, participants were asked to set their preferred center of projection such that the image seemed most natural to them. In Experiment 2, participants were asked to draw a floor plan of the virtual classroom. The results show that field of view, viewing mode, the center of projection and display all significantly affect presence and the perceived layout of the virtual environment. We found a significant linear relationship between presence and perceived layout of the virtual classroom, and between the preferred center of projection and perceived layout. The results indicate that the way in which virtual worlds are presented is critical for the level of experienced presence. The results also suggest that people ignore veridicality and they experience a higher level of presence while viewing elongated virtual environments compared to viewing the original intended shape.

Figure 1. The ‘Alberti window’: viewing a virtual reality environment from the center of projection shows the viewer an image as if the viewer looks at the scene through the frame of the screen.

Figure 2. Projector (a) and TV (b) showing a life-size virtual classroom, and TV showing a scaled-down virtual classroom (c).

Four vantage points A, B, C and D were used for the experiment.

Figure 3. Snapshots of the displayed virtual classroom for the projector displaying life-size images.

The four rows represent the four vantage points A, C, B and D, respectively. The three columns represent the three choices for the CoP at each vantage point.

Figure 4. Differences between the three display settings.

(a) Content of the projector displaying the life-size virtual classroom. (b) Content of the TV displaying the life-size virtual classroom. (c) Content of the TV displaying the scaled-down scene, and therefore, with the same content as for the projector. (d) Comparison of the relevant dimensions for the projector and the TV. Note that (a) (b) and (c) are scaled relatively to each other.

Figure 5. Medians and interquartile ranges for presence at four vantage points, with three levels of CoP and in two different viewing modes for all three display settings.

The horizontal line represents median and the distance between the top edge and the bottom edge represents interquartile range.

Figure 6. Displacement between the vantage point and the preferred CoP at four vantage points across three display settings (Medians).

Figure 7. Drawing task interface.

Figure 8. Function of the transformation coefficients in a perspective transformation; the black line is the square in the image plane and the green dashed line is the transformation result.

Figure 9. The 12 classrooms (at four vantage points and three CoPs) as perceived with binocular viewing on the projector displaying life-size content.

Figure 10. Relationships between presence, preferred CoP and perceived distortion (Spearman correlation).