Fixed quality layered truncation for scalable lossless audio

Abstract

The paper proposes a fixed quality layered bitstream truncator based on the MPEG-4 scalable lossless (SLS) audio coding system where, in contrast to original SLS truncator, the bitrate for each audio frame after truncation is not constant but instead an attempt is made to create constant quality audio. With the bitrate constraint for a period, the proposed system truncates each audio frame according to a pre-trained bitrate table. Simulation results show that with the same bandwidth, a constant and enhanced quality can be achieved by using the proposed method comparing with the one using the traditional fixed bitrate truncation method.