Primary-ambient Extraction in Audio Signals Using Adaptive Weighting and Principal Component Analysis

Most audio recordings are in the form of a 2-channel stereo recording while new playback sound systems make use of more loudspeakers that are designed to give a more spatial and surrounding atmosphere that is beyond the content of the stereo recording. Hence, it is essential to extract more spatial information from stereo recording in order to reach an enhanced upmixing techniques. One way is by extracting the primary/ambient sources. The problem of primary-ambient extraction (PAE) is a challenging problem where we want to decompose a signal into a primary (direct) and ambient (surrounding) source based on their spatial features. Several approaches have been used to solve the problem based mainly on the correlation between the two channels in the stereo recording. In this paper, we propose a new approach to decompose the signal into primary and ambient sources using Principal Component Analysis (PCA) with an adaptive weighting based on the level of correlation between the two channels to overcome the problem of low ambient energy in PCA-based approaches.