Binaural Cue Coding: Rendering of Sources Mixed into Amono Signal

This paper reviews Binaural Cue Coding (BCC). BCC is a lossy technique for either reducing either a number of source signals or a multichannel audio signal to one audio channel plus side information. In the case when a number of source signals (e.g. separately recorded instruments) are reduced to one audio channel plus side information, the BCC synthesis allows rendering of each source as if the separate source signals were given. The same side information supports different playback setups (e.g. binaural headphone playback, stereo loudspeaker playback, multi-channel loudspeaker playback). In the case when a stereo or multi-channel audio signal is reduced to one audio channel plus side information, the BCC synthesis renders a stereo or multi-channel audio signal with a spatial image similar to the original signal. The BCC representation of audio signals requires only the bitrate for one audio channel, which can be coded with conventional audio or speech coders, plus a few kb/s for the BCC side information. Existing mono broadcasting or communications systems can be upgraded with BCC for spatial audio by transmitting the BCC side information in addition to the existing mono audio transmission.