A new technology called MPEG Surround allows multichannel audio to be carried over extremely efficient (i.e. low bit-rate) audio channels. The process has been demonstrated over the past year and is now an international standard. Instead of sending five or six channels of discrete audio, the channels are downmixed to stereo and encoded and a small amount of parametric data describing the original multichannel signals is added to the encoded audio data. At the decode side, the stereo audio is decoded and the full surround signal is recreated using the parameterized surround information created during encoding.
Basically, MPEG Surround can be thought of as a matrix-like system that uses a separate data stream to control steering instead of unreliable level and phase detection. However, in reality it is only like a matrix system in that the input channels are downmixed, and the output channels are recreated from the downmix. The major difference is the parametric data, which allows near-perfect recreation of the original discrete audio.
A very attractive side benefit of MPEG Surround is in the downmix. A traditional matrix system requires an LtRt signal that contains intentional phase shifting of certain channels (usually the surrounds). This enables the matrix decoder to extract the phase-shifted material, but it creates a two-channel signal that may not be ideal in all situations. With MPEG Surround, a stereo downmix to LoRo (Left Only, Right Only) can be employed. This signal contains no additional phase shifting and can arguably work better with certain program material such as music.
In reality, MPEG Surround is far more sophisticated than this simple explanation. For more information, visit http://www.codingtechnologies.com/products/mpgsrnd.htm.
