Over the years, a large number of new display principles have emerged from the search for flat, high-quality or low-cost alternatives for the successful CRT. The various properties a good display should have include high static resolution, high peak brightness, high contrast, high lumen efficacy and favorable dynamic behavior. Of these items, it is the last one that sometimes remains underexposed. Nevertheless, the temporal aspects of a video display determine important properties like flicker, dynamic resolution and motion portrayal.
In the early days of television, both the imaging and the display device used an electron beam to scan the scene, and it was only logical that the scanning had the same parameters, i.e., line time, picture rate and interlace factor. In such a system, the delay between the registration of an input pixel and the display of an output pixel is equivalent for all pixels. Over time, the technologies for imagers and displays diverged, and the delay for every displayed pixel no longer had the same value.
In a first category of displays, the delay variation, resulting from the scanning mismatch, can be described as a linear function of the horizontal and vertical position on the screen. Due to motion in the scene, these delay variations are translated into spatial position errors, which are proportional to the local velocity of the objects in the scene. Consequently, for these display and imager combinations, motion results in a geometrical distortion of the moving object. This distortion varies smoothly as a function of the spatial position and therefore is considered mild and acceptable, at least for consumer vision applications.
In a second category of imager and display combinations, positional errors result as a function of the velocity as described above, but the relation between the position error and the velocity is not constant. This relation may vary per color as, for example, in color sequential displays, per picture portion, and for tiled displays like video-wall screens. Or, the relation may depend on the picture number, for example, for broadcasted film material on any known display, and for video in PCs. Since this leads to delay variations that do not necessarily vary smoothly, correction gives a clear improvement that is almost necessary for acceptable image quality.
Recently, correction algorithms reached a level of maturity that allowed introduction on the market of consumer video equipment of dedicated ICs and software packages that enable real-time correction on digital signal processors.
In a third category, positional errors result as a function of the velocity, but the relation between positional error and velocity depends on the displayed data. This is typically the case for displays with an on/off character that realizes gray scale representation with pulse-width modulation, as in PDPs. The artifacts of this category are even stronger than in the previous display categories and repair therefore leads to an obvious quality improvement.
Basically, the CRT is a "stroboscopic" display device. That is, the light for an individual pixel is generated as a pulse, which is very short, compared to the picture time. For nonstroboscopic displays like many of the emerging types such as LCDs and PDPs each image is displayed during a display time.
For an LCD this display time is constant and equals the picture time, while for a PDP it varies with the brightness of the pixel. When there is motion in the image, the viewer will track the motion, and hence integrate the intensity produced by each image, along the motion trajectory. For a given display time, the integration can be described as a convolution of the original image and a motion-tracking/temporal sample-and-hold function. From a mathematical analysis it follows that the perceived effect, for a constant display time, is a blurring of moving-image parts, where the amount of blurring is proportional to the local velocity and to the display time.
For a PDP the display time is smaller than for an LCD. The LCD, therefore, gives more blurring. However, the data-dependent display time of the PDP gives rise to the more annoying "dynamic contours" artifact.
Weaknesses in the temporal behavior of emerging displays results in impairments of moving-image parts. Moreover, the effects are proportional to the velocity of the motion. Knowledge of the motion in the scene can be obtained from so-called motion-estimation algorithms. These devices are designed to calculate whether there is motion at a certain part of the screen, and if so, into which direction and how fast. Evidently, the capability to design high quality true-motion estimators is essential for the design of remedial processing.
The remedy for all artifacts consists of the cascade of a motion-compensated picture interpolation module, a device that applies motion vectors calculated with a motion estimator to compensate for the delay error introduced by the scanning mismatch, together with an inverse filtering module to counteract the effects of integration of nonstroboscopic displays.
An analysis in the Fourier domain shows that this inverse filtering cannot be perfect as the transfer the frequency characteristic resulting from the integration along the motion trajectory has zeros. In addition to the imperfections of the inverse filter due to the nature of the integration filter, the pulse-width modulation techniques for gray-level generation cause data dependencies of the artifacts that can be reduced but not completely eliminated.
Motion estimation plays an enabling role in the reduction of display artifacts. Based on years of research in the motion estimation area, these algorithms have reached a level of maturity that enabled implementation of high-quality true-motion estimators in consumer electronics devices.
The complete version of this article will be presented at ICCE.