Unified Speech and Audio Coding

Unified Speech and Audio Coding ( USAC ) is an audio data compression format and codec for both music and speech signals or mixed using a very low bit rates from 12 to 64 kbit / s It is currently under development in the MPEG and is used as an international standard ISO / IEC 23003-3 (aka MPEG -D Part 3) and also as an MPEG -4 audio object type in ISO / IEC 14496-3:2009 / Amd 3 defined.

It uses linear prediction in the time domain and tools for the coding of the residual signal ( ACELP similar techniques ) for speech signal segments and transform coding tools ( MDCT - based techniques ) for music signal sections and it can depend on the signal content between the two approaches switch dynamically. Among the objectives for the development include performance improvements over HE- AACv2 and AMR -WB as a specialized format for speech signals. Improved versions of the MPEG-4 spectral band replication ( SBR), and techniques for the parametric audio coding in MPEG -D are incorporated into the MPEG Surround USAC codec.

History

The standardization process began in October 2007 with the call for proposals at the 82nd meeting of the MPEG. For the development of the standard, a proposal from the Fraunhofer Institute for Integrated Circuits and VoiceAge Corporation (see ACELP ) was taken as a basis. In this proposal, the so-called " Reference Model 0 ( RM0 ) ", that choice was determined by a test phase in the 85th MPEG meeting in July 2008. At the 86th meeting in October the creation of a reference encoder has been proposed and agreed upon.