Fixed bit rate perceptual wavelet packet audio coder

This paper presents a perceptual wavelet packet audio coder incorporating temporal and simultaneous masking models with a scalable bit rate allocation scheme. It is demonstrated that the use of even a relatively simple functional model of temporal masking produces bit rate savings in addition to those given by simultaneous masking. The performance of the coder is evaluated using the PEAQ (ITU-R BS.1387) objective measurement criterion for various audio materials given a fixed data rate constraint. Two methods are evaluated to optimally allocate bits, lower band selection, and signal to mask ratio (SMR) selection. Results reveal that for lower bit rates, the lower band selection method provides higher audio quality than that given by SMR selection