Binaural Reproduction Over Loudspeakers Using a Modified Target Response
MetadataShow full item record
Crosstalk cancellation (XTC) is a technique that can be used to play binaural content, typically meant for headphone playback, over two or more loudspeakers. Though effective at creating a binaural spatial sound field at the listening position, many XTC algorithms introduce spectral coloration, suffer from spatial robustness issues and create filters that are unrealizable in practice. Past approaches to dealing with this issue rely heavily on regularization. In this work we propose a new topology for loudspeaker binaural rendering (LBR) that performs better than conventional techniques without the need for regularization commonly associated with crosstalk cancellation based binaural renderers (XTC-BR). We then explore the use of a proposed LBR in the context of multiple output channels. A method is investigated to further optimize the filter design process by selecting an appropriate modeling delay and filter length using methods practiced in XTC filter design.