Abstract: Using frequency and time masking effects to encode an audio signal. The frequency band of the digital audio signal is divided into a number of subbands. First signal-to-mask ratios for the respective subbands are estimated in response to digital signal samples in each subband of the ith frame included in the audio signal. The first signal-to-mask ratio for the ith frame are stored for a predetermined time period and delayed signal-to-mask ratios for the (i-1)st frame are prestored and synchronized with the first signal-to-mask ratios. Second signal-to-mask ratios are based on the first signal-to-mask ratios and the delayed signal-to-mask ratios. Adaptive bit allocations for each of the subbands are based on the second signal-to-mask ratios Digital signal samples are quantized in each subband in response to the generated bit allocation information for each of the subbands. Finally, the quantized digital signal samples are formatted using the generated bit allocation information.

Abstract: An improved apparatus and method utilizes both frequency and time masking effects to evaluate an audio distortions so that the results obtained thereby have a best match with actual human auditory perception. A power density spectrum is first estimated for an input digital audio signal and a frequency masking threshold is determined based on the power density spectrum for the input digital audio signal. In the meantime, a power density spectrum is estimated for a difference signal, wherein the difference signal represents the difference between the input digital audio signal and an output digital signal. A perceptual spectrum distance is then determined based on the power density spectrum of the difference signal and the frequency masking threshold.