Compression of High-Quality Audio Signals, Including Recent Methods Using Wavelet Packets
✍ Scribed by Mike Sablatash; Todor Cooklev
- Publisher
- Elsevier Science
- Year
- 1996
- Tongue
- English
- Weight
- 143 KB
- Volume
- 6
- Category
- Article
- ISSN
- 1051-2004
No coin nor oath required. For personal study only.
✦ Synopsis
digital formats. The compact disc (CD) has become Sablatash, M., and Cooklev, T., Compression of High-a convenient baseline for comparing the perfor-Quality Audio Signals, Including Recent Methods Using mance of digital audio recording and transmission Wavelet Packets, Digital Signal Processing 6 (1996), systems, although it requires more bits than neces-96-107. sary for high-quality reproduction. In the CD standard each channel of stereo audio signal is repre-Reasons for the importance of compression of high-qualsented using 16-bit samples at 44,100 times per secity audio are advanced. Audio signal analog-to-digital conond, yielding an overall data rate of 1.4 Mbps, version is summarized. Human perception and auditory modeling are briefly discussed. Signal quantization and resulting in a reproduction bandwidth (BW) of 20 block processing are concisely presented. Adaptive trans-KHz and a dynamic range of 96 dB [1,2]. Compresform coding, subband coding, and time domain aliasing sion would result in many audio products if rates cancellation are succinctly reviewed. Advanced physical such as 128, 96, 70, 64, or 48 Kbps-or even lowermodeling compression methods, perceptually based algocould be used for stereo reproduction to permit effirithms, and MUSICAM are compactly described. Several cient utilization of storage capacity and transmisrecent proposals are discussed, and the most promising sion BW. The data rate of 1.4 Mbps is too high for and effective of these, based on adaptive wavelet packets, many computer and broadcast applications with limis described. Performances of many of the techniques are ited storage capacity and bandwidth [1,2]. By 1992 given to provide comparisons. Research directions to enalgorithms based on psychoacoustics had comable improvement of the present state of the art are pro-
pressed high-quality stereo audio to data rates rangvided. It is to be noted that approaches employing adapted wavelet packets show particularly outstanding promise.