Forging ahead with spatial audio coding

Spatial audio coding is a term recently coined to describe low bit-rate audio coding that enables the transmission of multichannel spatial audion at fairly low bit rates. The primary channel is often coded using a legacy system for compatibility of the downmixed audio with existing decoders. The trick with these codecs is to code the spatial information in such a way that an approximation to the spatial impression of the original scene can be reconstructed from the downmix plus the parametric information so that it has as few perceptual differences from the original as possible. Among the different approaches include the use of spectral Weiner Filtering (SWF). Another is a format-independent scheme that is based on a time-frequency analysis of the signals to be encoded, deriving cues that describe the scene rather than the channel relationships. Another is based on the argument that a transform Unified Domain Transform can be shown to be equivalent to the process called Binaural Cue Coding used in the MPEG reference model for spatial audio coding. Such a transform is said to be lossless and invertible. Other coding include the representation of interchannel time difference and parametric stereo coding.