This research proposes a generalized and optimized framework for time–frequency processing of spatial audio using a signal covariance matrix. This framework is relevant for a wide variety of spatial applications, such as perceptual spatial coding, stereo upmixing, decorrelation, and so on. The matrix, which represents interchannel dependencies, is perceptually relevant for the transmission of the listener’s spatial experience. In a typical application, the original time–frequency covariance matrix is transformed into the target matrix, optimizing the sound quality using a least mean square metric. In an example of upmixing stereo music, informal listening tests confirmed the validity of the framework.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.