Parametric Time-Frequency Domain Spatial Audio

5 First-Order Directional Audio Coding (DirAC)

Ville Pulkki,¹ Archontis Politis,¹ Mikko-Ville Laitinen,² Juha Vilkamo,² and Jukka Ahonen,³

¹Department of Signal Processing and Acoustics, Aalto University, Finland

²Nokia Technologies, Finland

³Akukon Ltd, Finland

Directional audio coding (DirAC; Pulkki, 2007) is a non-linear time–frequency domain method to reproduce spatial sound. The sound field affects the processing instantaneously in auditory frequency bands, enhancing the quality of reproduction when compared to traditional time domain methods. The method has been developed by starting from assumptions about the spatial resolution of human perception, and designing the reproduction system to follow these assumptions using multichannel microphone signals that have first-order spherical harmonics as their directional patterns.

The name of the system, “directional audio coding,” includes the word “coding,” although the technique is not primarily a method to compress the data rate. The motivation of the name comes from the processing principle of DirAC. The directional properties of the sound field are analyzed, and then used instantaneously in the synthesis of sound. The property values are thus encoded into a parametric stream with human time–frequency resolution, and later decoded into the reproduced sound field. This is done primarily to enhance the quality of spatial sound reproduction, and only in some applications is the system designed to compress the data rate.

This ...