Quality Aspects in Multimedia Applications

Multimedia systems and techniques are being applied more and more in various forms in various areas of daily life. As one result of this development media streams are used in different application contexts and for different application goals. Thus, media tailored for one specific application scenario is often transferred into a different context, seldom considering the new and different quality requirements. Consequently, monomedia streams in current multimedia applications are regularly falling short the suitable quality level. To avoid this and to provide high-quality multimedia presentations methods, it is necessary to evaluate media quality with respect to the specific application context. Subjective tests still provide the best and most exact quality rating. However, designing and setting up the test environment and carrying out the test is a very time consuming and expensive task. Moreover, subjective tests alone do not provide a measure of quality necessary for certain applications, such as automatic determination of optimal compression parameters for images and videos. For certain applications, objective quality measurement methods may represent an alternative to subjective tests. Many measures of media quality, escpecially for image and video quality, are based on the Modulation Transfer Function. The Modulation Transfer Function has been used successfully to characterize the resolution properties of analog systems with restriction to linear systems. It might be expected that these measures could be used to evaluate the quality loss in digital media and imaging systems, as well. As will be shown this is not true in general. In the following sections some aspects of objective quality evaluation are discussed and an introduction into quality measurement for still and video images based on the Modulation Transfer Function is provided. We describe the limitations and problems of these quality measures in the context of digital imaging systems. Moreover, we present new approaches for the enhancement of these evaluation methods