Low bit rate transparent audio compression using a dynamic dictionary and optimized wavelets

The authors propose an audio synthesis/coding method which employs an optimized wavelet-transform (WT)-based adaptive transform coding to exploit perceptual masking. In addition, a dynamic dictionary is used to extract source redundancies. Experiments indicated that transparent coding of monophonic CD quality signals (sampled at 44.1 KHz) is possible at bit rates approaching 64 kbit/s when the WT-based method alone is used, and in the range of 48-64 kbit/s when the dynamic dictionary is also employed.<>