Transforms and Subband Decomposition

Abstract

The properties of the audio and video signals, and the digitization process have been discussed in the previous Chapters. When a signal is digitized, further processing of these signals may be needed for various applications, such as compression, and enhancement. The processing of multimedia signal can be done effectively when the limitation of our hearing or visual systems is taken into account. For example, it was shown in Chapter 3 that the human ear is not very sensitive to audio signals with frequencies above 10–12 KHz. Similarly, the eyes also do not respond well above 20 cycles/degree. This dependency of our sensory systems on the frequency spectrum of the audio or visual signals has led to the development of transform and subband-based signal processing techniques. In these techniques, the signals are decomposed into various frequency or scale components. Various components are then suitably modified depending on the application at hand. In this Chapter, we will discuss mainly two types of signal decomposition techniques: transform-based decomposition and subband decomposition.