Virtual rendering of object-based audio

摘要

Embodiments are described for a system for virtual rendering of object based audio through binaural rendering of each object followed by panning of the resulting stereo binaural signal between a plurality of cross-talk cancelation circuits feeding a corresponding plurality of speaker pairs. In comparison to prior art virtual rendering utilizing a single pair of speakers, the described embodiments improve the spatial impression for both listeners inside and outside of the cross-talk canceller sweet spot. Also described is an improved equalization technique for a crosstalk canceller that is computed from both the crosstalk canceller filters and the binaural filters and applied to a monophonic audio signal being virtualized. The described techniques improve timbre for listeners outside of the sweet-spot as well as a smaller timbre shift when switching from standard rendering to virtual rendering.

主权项

1. A method for virtually rendering object-based audio comprising:
applying an object signal and a corresponding object signal position to a binaural filter pair to generate a binaural signal, wherein the object signal and the object signal position are associated with an audio object of the object-based audio; multiplying the binaural signal by panning coefficients computed based on the object signal position to generate scaled binaural signals; panning the binaural signal generated from the binaural filter pair to a plurality of crosstalk cancellers, wherein the panning to crosstalk cancellers is controlled by a position associated with each audio object; summing the scaled binaural signals together; and applying a cross-talk cancellation process to the summed scaled binaural signals to generate a speaker signal pair for playback through a speaker.