An efficient multi-channel spatial audio coding method based on inter channel correlation

Abstract

In this thesis, we propose an efficiently representing method of spatial cues used in multi-channel spatial audio coding. The Binaural Cue Coding (BCC) method introduced recently represents multi-channel audio signals by means of mono signal and spatial cues, e.g., the Inter-Channel Level Difference(ICLD) and the Source Index(SI) . We try to express ICLD and SI information more efficiently based on the Inter-Channel Correlation (ICC) We apply the strategy of varying the type of spatial cues according to the ICC. When the SI is used as spatial cues, audio quality degradation occurs in output channels because the spectrum is empty at other channels that are not indexed. In order to solve the problem, we propose a compensation method for empty spectra created by utilizing the SI. In our experiments, we perform a MOS test and measure the spectral distortion in order to confirm the performance of our proposed method. The results show that bit rate of side information is reduced by applying the strategy of varying the type of spatial cues based on the ICC and that the proposed compensation method is a simple but effective method to reconstruct the empty spectra of output channels from the comparison of the spectral distortion. Consequently we confirm that the proposed method can reduce the bit rate of side information without noticeable degradation in the audio quality.