Object motion presented in motion picturediffers from that of real moving objects: the former is a time-series of still shots, while the latter features continuous motion. The number of information updates in a motion picture is likewise restricted by the refresh rate of the display, whereas the refresh rate for the real world is infinite. Because of this finite refresh rate, faster motion manifests more obvious motion artifacts when presented on displays. Different types of artifacts can arise depending on the viewer's eye movement and display device characteristics. These artifacts include judder (an appearance of jumpy, rather than continuous, movements), edge banding (multiple repetition of the edge of the object), and motion blur (the edge of the object getting blurred). From left to right, the following three panels illustrate these three types of motion artifacts occurring with a white target circulating on a black background (images courtesy of Paul V. Johnson).
There are many factors that can affect how smoothly motion is presented in motion pictures: object speed, capture rate (how frequently the motion is captured), presentation rate (how frequently the motion picture is presented), duty cycle (the proportion of frame time during which images are presented), and flash number (the number of repetitive presentations of the same frame). Unfortunately, there is no single 'magic' number for either capture or presentation rate that provides smooth object motion at all speeds. Any finite capture or presentation rate will manifest motion artifacts after a certain speed. Therefore, as an object moves faster and faster, you need higher and higher capture and presentation rates.
Using a state-of-the-art Samsung 240Hz OLED TV, we tested how fast object motion could be presented without motion artifacts. Using psychtoolbox and a custom-built video interface (provided by Samsung), we simulated many presentation modes by varying the capture rate, presentation rate, duty cycle, and flash number. Besides the presentation and capture rates, shorter duty cycles and lower flash numbers were also beneficial to reducing motion artifacts. These findings were in line with earlier investigations on motion artifacts.
The visual system has a limited sensitivity to detect temporal variation in brightness. The saliency of flicker depends on many factors, notably the frequency and waveform of temporal modulation, and the intensity. Flicker is less likely to be detected as the temporal frequency gets higher, the duty cycle (proportion of frame time during which images are presented) gets larger, and the intensity gets lower. We tested flicker visibility on a 240Hz OLED display with various presentation modes. The experimental findings were in agreement with theoretical expectations: flicker was less visible with a shorter duty cycle and higher presentation rates.
Collaborators: Paul Johnson [home], David Hoffman, Andy Vargas, Marty Banks