Temporal image capture and display characterization for metrics for moving image quality

Temporal measures of display performance are important for depiction of still images, and are even more critical for depiction of moving images (video). In addition to the flicker and crawling patterns that occasionally appear when certain colors/grayscales and patterns are displayed (e.g. DMD/plasma temporal modulation), moving images are susceptible to a host of time-based artifacts (visible defects) such as motion blur, color breakup, image distortion, and incorrect depiction of object motion. Accuracy of temporal representation can be affected by temporal compression (e.g. motion prediction) in the image data handling, and by the operation of the display system (e.g. scaling algorithm). It is widely known that well-presented motion imagery can offer both improved quality and improved interpretability compared to comparable still imagery (for example a freeze frame of a video), provided that timing characteristics are accurately reproduced. Because of the importance of display characteristics, positional measurement of the physical light output of the display is a necessary addition to measurement of image component motion in the video data file. The NIST Motion Image Quality Project team is adapting its method of sequential image capture with external hardware timestamp (Electronic Imaging 2003), to make precision measurements of the temporal characteristics of displays. Hardware measurements are being combined with subjective evaluation of video content and image file analysis, to produce a new model for moving image quality and interpretability, which will serve as a guide both for content creation and for display design.