Conventional perceptual coding algorithms do not normally exploit the temporal masking property of the human auditory system. These algorithms rely only on simultaneous masking models to calculate the masking threshold. This work proposes the use of a temporal masking model, combined with a simultaneous masking model, in wavelet packet-based audio coding. The result is a reduction in bit rate of approximately 25 kbps while preserving the transparent perceptual quality of audio signals, at a sampling rate of 16 kHz. This is achieved by a more accurate calculation of the combined auditory masking threshold. Another proposed approach of calculating the masking threshold accurately is by oversampling in the discrete wavelet transform. Most of the current wavelet based perceptual coders use the critically sampled discrete wavelet transform. The problem with this transform is aliasing, resulting from the down sampling process after each decomposition. This aliasing is cancelled in the decoding process, however the masking threshold calculation is done using aliased wavelet coefficients. Oversampling in the discrete wavelet transform is proposed in this work to avoid aliasing. The results show that by oversampling in the discrete wavelet transform, a reduction in a bit rate of up to 16 kbps can be achieved for audio signal, at a sampling rate of 44.1 kHz. The two approaches are then used in scalable audio coding in developing a fixed bit rate audio coder. The bit rate saving from the two approaches is used in scalable audio coding to include additional frequency content at the target bit rate.