Many things influence a visualization of a dataset. These include the nature of the data itself and its uncertainties, the visual representation space, the visualization algorithms used for implementation, as well as algorithmic simplifications. Another factor impacting the final visualization is how accurately the images are rendered by the display system. We seek to develop a deeper understanding of each part of this process. We are currently studying the impact of the display on stereo visualizations.
Every display technology has its own inherent performance limitations. The non-ideality of the display introduces constraints on what information can actually be observed by the viewer. By characterizing the optical performance of the display system, we can quantitatively determine what visual information may be impaired. The operators of the display facility can use the results to evaluate how the visual characteristics that are important to them are affected, understand the visual limitations of their system, and use the performance data to drive improvements. This also serves as feedback to manufacturers in their work to improved displays. The display performance data also informs the scientists that create the visualization data how much of that data can actually be realized/observed. For example, the contrast ratio of an image feature can be dramatically degraded with the introduction of ambient lighting. Light contamination can be especially problematic for multi-wall CAVE systems, where the images from the adjacent screen can reduce the viewability at the region of interest. In addition, the spatial distribution of the contrast ratio on the front screen of a multi-wall CAVE system changes when the adjacent walls are turned on.