The unified speech and audio coding (USAC) is under the ISO/MPEG standardization, and is supposed to complete the standardization process in 2011. In the perspective of technical merits that the USAC brings in audio and speech compression, this paper highlights stereo algorithms that perform best depending on bitrates and audio content types in various use case scenarios. At very low rates in the ranging from 16 to 48 kbits/sec, MPEG Surround tool with phase parameters can be used to deliver high stereo quality in a bit-efficient way. In the mid range of 48-96 kbits/sec, the unified stereo scheme can be utilized based on residual signals with little burden of complexity. Finally, at higher bitrates, a complex stereo prediction in MDCT domain can be used through the real-to-imaginary Transform in addition to the conventional Mid/Side stereo. Over broad range of bit rates, the newly developed methods in the USAC result in excellent stereo quality encompassing a variety of audio contents in a bit-efficient way with little cost of complexity.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.