Bit allocation method for AC-3 encoder

Dolby AC-3 is an audio coder selected in the United States high definition television (HDTV) standards and widely adopted for the audio codec in DVD films. For audio codecs, the complexity of the audio encoders is always higher than that of the decoders. The main module leading to the higher complexity is the bit allocation process. The objective of the bit allocation is to allocate restricted bits to encoded information. Since the AC-3 coder has adopted a floating point representation which the magnitude of each spectral line is represented with an exponent and a mantissa, the bit allocation process not only has to decide the suitable quantization method for the mantissa similar to the process in other coding standards such as MPEG 1, but also the exponent strategies and hence the parameters in psychoacoustic models. This coding process has been referred to as a hybrid coding (Todd et al. 1994) and has led to immense complexity for an encoder. This paper proposes an efficient method for the bit allocation process.

[1]  Louis Dunn Fielder,et al.  Parametric Bit Allocation in a Perceptual Audio Coder , 1994 .

[2]  S. Shlien,et al.  Guide to MPEG-1 audio standard , 1994, IEEE Trans. Broadcast..

[3]  Nuria González Prelcic,et al.  A fast noise-scaling algorithm for uniform quantization in audio coding schemes , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.