An audio coding and decoding method based on human auditory frequency scale signal decomposition

Abstract

The invention discloses an audio encoding and decoding method based on human ear auditory frequency scale signal decomposition, which is used for converting a processing result of a linear frequency domain to a corresponding human ear auditory frequency scale (Bark) to control a quantization error of a linear frequency domain parameter in order to better use a psychoacoustic model. Signal decomposition is directly carried out on the auditory frequency scale of human ears; projection calculation of a time domain signal in a Bark scale is realized; according to the invention, non-uniform sub-band decomposition of the audio signal is carried out based on the human ear auditory frequency scale ; the psychoacoustic model of human ear auditory sense is effectively utilized, and calculation precision and encoding and decoding efficiency of psychological acoustic model are improved According to the method, the quantization object is the sub-band time domain signal, the coding and decoding quantization distortion is controlled within the parameter allowable range of the psychological acoustic model on the premise that the audio coding and decoding code rate is met, the quantization mode isnovel, and the efficiency is high.

Description

technical field [0001] The invention relates to an audio coding and decoding method based on human auditory frequency scale signal decomposition, and belongs to the technical field of speech processing. Background technique [0002] Existing audio codec methods mainly include codec methods for speech and codec methods for general audio signals. In the coding and decoding methods for speech, if the speech transmission with better sound quality is to be achieved at a lower bit rate, the core coding and decoding methods are all based on code-excited linear prediction. The essence is that the speech signal can use an accurate signal The code-excited linear prediction core method can successfully compress the excitation signal and channel model parameters through vector quantization. The basic core of the codec method for general audio is the perceptual codec method. If high-quality audio data transmission is to be achieved, frequency transformation methods with uniform subband ...

AI Technical Summary

Problems solved by technology

The basic core of the codec method for general audio is the perceptual codec method. If high-quality audio data transmission is to be achieved, frequency transformation methods with uniform subband decomposition, such as modified cosine transform (MDCT), etc., are unavoidable. , the reason is that it can realize the real number operation of critical downsampling, and the transformation result is directly related to the signal frequency domain information, and it is easy to apply the psychoacoustic model

Images