Audio fine scalable coding method and system based on perception self-adaption bit allocation

Abstract

The invention relates to the technical field of audio coding, in particular to audio fine scalable coding method and system based on perception self-adaption bit allocation. The method comprises the following steps of: preprocessing input signals; carrying out subband division on frequency-domain signals; calculating the perception importance of each subband; uniformly sequencing the subband from small to large according to the perception importance; extracting the subband with the largest perception importance for scalable longitudinal vector quantization; and then carrying out self-adaption adjustment on the vector-quantized subband with the largest perception importance. The system comprises a preprocessing module, a subband division module, a perception importance calculating, sequencing and extracting module, a scalable quantizing and coding module, a self-adaption adjusting module and a scalable coding finish judging module. The invention realizes high-efficient fine scalable audio coding, preferably realizes the unification of quantization precision and quantization efficiency and also satisfies the requirement of high tone quality when improving coding efficiency.

Description

technical field [0001] The present invention relates to the technical field of audio coding, in particular to an audio fine hierarchical coding method and system based on perceptual adaptive bit allocation. Background technique [0002] Scalable audio coding technology divides the code stream into a core layer and multiple enhancement layers. The core layer guarantees the minimum reconstruction quality of the signal, and the enhancement layer gradually improves the reconstruction quality by improving the signal-to-noise ratio or extending the frequency band. The received The more enhancement layers, the higher the decoding sound quality. [0003] Scalable coding methods can adapt to network bandwidth fluctuations by directly discarding the enhancement layer code stream. The finer the grading granularity, the more effective it can adapt to network bandwidth fluctuations; on the other hand, the objective criterion for evaluating the performance of scalable audio coding is that...

AI Technical Summary

Problems solved by technology

The optimal bit allocation method guarantees the improvement of coding quality, but due to the quantization and compression defects of scalar quantization itself, this method limits the improvement of quantization efficiency to a certain extent, and cannot be applied to medium and low bit rates.

[0006] The G.729EV standard enhancement layer divides the signal into 32 sub-bands, adopts the measurement criterion of perceptual importance, sorts each sub-band according to the perceptual importance, and determines the optimal bit allocation scheme according to the sorting results and the number of bits available for allocation. Split spherical vector quantization is performed on the MDCT coefficients of each subband. The bit allocation method used in the enhancement layer coding of the G.729EV standard is not optimal, and the number of bits for encoding each subband is expensive. The allocation method can only guide the encoder to quantize and encode a few subbands, and most of the subband information will be completely lost. Although this method has significant quantization efficiency, it has the phenomenon of uneven bit allocation and bit waste, resulting in some subbands Band bits are seriously wasted, and some sub-band bits are severely lacking, which ultimately affects the improvement of sound quality

[0007] From the above technologies, the current fine scalable audio coding is in two extreme states. The quantization efficiency and grading granularity cannot have a good unified process. The method with high quantization efficiency has low grading granularity, and the grading with low quantization efficiency Granularity is high

Images