Abstract
Parametric stereo (PS) audio coding is a stereo version of binaural cue coding (BCC) which is a method for multichannel spatial rendering based on a single downmixed audio channel and spatial information. PS uses three spatial parameters that are similar to spatial cues of BCC: Inter-channel Intensity Difference (IID), Inter-channel Phase Difference (IPD), Inter-channel Coherence (IC). While BCC uses Inter-Channel Time Difference (ICTD) parameter to analysis and synthesis time delay, PS uses Inter-channel Phase Difference (IPD) parameter. IPD analysis is more accurate than ICTD analysis but IPD synthesis using constant IPDs goes through fractional delay.
In this thesis, we propose a new IPDs synthesis algorithm that can reduce the spatial degradation caused by fractional delay. The IPDs is synthesized proportionately to the frequency using a phase unwrapping technique. We also propose a new phase unwrapping algorithm based on the relation between Inter-aural Level Difference (ILD) and Inter-aural Time Difference (ITD). To apply the proposed algorithm to the parameter Overall Phase Difference (OPD), the conventional OPD estimation algorithm is adopted. Subjective listening test was conducted to evaluate the performance of proposed algorithm. It verified the spatial degradation by fractional delay and improvements using proposed algorithm