Transform coding with integer-to-integer transforms

A new interpretation of transform coding is developed that downplays quanti zation and emphasizes entropy coding, allowing a comparison of entropy codi ng methods with different memory requirements. With conventional transform coding, based on computing Karhunen-Loeve transform coefficients acid then quantizing them, vector entropy coding can be replaced by scalar entropy co ding without an increase in rate, Thus the transform coding advantage is a reduction in memory requirements for entropy coding. This paper develops a transform coding technique where the source samples are first scalar-quanti zed and then transformed with an integer-to-integer approximation to a nono rthogonal linear transform. Among the possible advantages is to reduce the memory requirement further than conventional transform coding by using a si ngle common scalar entropy codebook for all components. The analysis shows that for high-rate coding of a Gaussian source, this reduction in memory re quirements comes without any degradation of rate-distortion performance.