Minimization of Decorrelator Artifacts in Directional Audio Coding by Covariance Domain Rendering
Directional Audio Coding (DirAC) is a perceptually motivated microphone technique that models the sound field as a combination of a plane wave and a surrounding diffuse field with a time–frequency resolution that approximates that of the human spatial hearing. In this paper a recently proposed covariance domain spatial-sound rendering method was applied to optimize the DirAC reproduction by minimizing the amount of the decorrelated sound energy. When several semi-independent microphone signals were available, this procedure was shown to improve the overall perceived sound quality, especially with audio content that has an impulsive fine structure, such as applause and speech. In all tests, the covariance rendering method performed similarly or better than the legacy rendering method, making it the preferred choice for performing DirAC synthesis.
Download Now (477 KB)