Realization of filter banks by means of wave digital filters

Wave digital filters (WDFs) are derived from classical filters. They have very good properties with respect to coefficient accuracy requirements, stability, and dynamic range. One of the most attractive classes of WDFs are lattice WDFs (LWDFs), which can be used for filtering with sampling rate alteration and multirate filter banks in a very efficient way. The principles of LWDFs and LWDF filter banks are extended for polyphase and multidimensional WDFs and WDF filter banks. The following applications are presented: efficient coding of audio signals, in particular music signals; realization of a filter bank for real-time spectrum analysis; and realization of a 2-D filter bank with sector-shaped regions for efficient coding of images by decomposing these into several directional images