Large, high-resolution displays allow orders of magnitude more data to be visualized at a time than ordinary computer displays. Previous research is inconclusive about the circumstances under which large, high-resolution displays are beneficial and lacks behavioural data to explain inconsistencies in the findings. We conducted an experiment in which participants searched maps for densely or sparsely distributed targets, using 2-million-pixel (0.4 m × 0.3 m), 12-million-pixel (1.3 m × 0.7 m) and 54-million-pixel (3.0 m × 1.3 m) displays. Display resolution did not affect the speed at which dense targets were found, but participants found sparse targets in easily identifiable regions of interest 30% faster with the 54-million-pixel display than with the other displays. This was because of the speed advantage conferred by physical navigation and the fact that the whole dataset fitted onto the 54-million-pixel display. Contrary to expectations, participants found targets at a similar speed and interacted in a similar manner (mostly short panning movements) with the 2- and 12-million-pixel displays even though the latter provided more opportunity for physical navigation, though this may have been because panning used velocity-based control. We are applying these findings to the design of a virtual microscope for the diagnosis of diseases such as cancer.
Performance and interaction behaviour during visual search on large, high-resolution displays