This project deals with the compression of information for efficient transmission and storage. The information can be of the most general kind, but the approach here takes into account the redundancy in the data which expresses itself as a spectral distribution . This is done with the help of a compression scheme called subband coding. In this scheme, the spectral characteristics as well as the characteristics of the signal, which pertain to the human perceptive systems such as the hearing or visual system, can be exploited. We are addressing several fundamental issues in the design and implementation of subband codes. The optimization of orthonormal as well as biorthogonal subband coders is being considered in this project. We are also addressing the cases of nonuniform subband coders, which are most appropriate for the case of audio and image coding; and the theory of sampling and reconstruction of non bandlimited signals, showing how such a theory can be exploited in quantization and compression of data. This aspect of the research generalizes the family of oversampled analog- to-digital converters to the case of non bandlimited signals. Another topic under investigation is a special class of subband coders called cyclic coders, which are potentially capable of greater coding efficiency. This class is especially useful for the compression of short blocks of independent data.
