Scalable audio coding based on hierarchical transform coding modules

We propose a method that hierarchically quantizes wideband Modified Discrete Cosine Transform (MDCT) coefficients by developing a module that has a transform coding method primarily for audio as the basic structural unit and freely using this module multiple times at the desired frequencies. The major feature of this method is to implement a simple structure having a high degree of freedom in scalable coding to hierarchically quantize MDCT coefficients over a wide band of frequencies by sharing the proposed module and using it multiple times. This paper presents examples using combinations of the module operating at a sampling frequency of 48 kHz and a bit rate of at least 8 kbit/s. In this example, a bit rate of at least 8 kbit/s and a reconstructed frequency band of at least 4 kHz can be selected as the objective. Subjective evaluation tests are performed to verify the effectiveness of the proposed method. © 2001 Scripta Technica, Electron Comm Jpn Pt 3, 84(8): 34–45, 2001