REDUNDANCY REDUCTION IN COMPRESSED AUDIO SIGNALS

Using only psychoacoustic effects for data compression of audio signal s leaves redundant components in the compressed audio signal due to re maining correlations. Prediction-based methods will be presented that allow using these redundancies for further data compression. For such an extension of the standardized MPEG algorithm a parameter optimizati on by measurements with different audio signals is performed. An imple mentation of this new method confirms the measurement results.