【作者】 方贞；

【导师】 余绍军；

【作者基本信息】 中南林业科技大学 ， 计算机应用技术， 2012， 硕士

【摘要】 随着网络信息与多媒体技术的快速发展,人们对于多媒体业务的需求越来越迫切,多媒体与计算机网络技术逐渐融合为一体,已经渗透到社会经济与生活的各个方面。为了方便存储和传输,基于心理声学模型的感知音频编码技术得到了广泛的应用。高级音频编码技术(Advanced Audio Coding,简称AAC)是目前最先进的感知音频编码技术,它的优势主要体现在以下三方面：较高的信号压缩比、模块化的量化编解码过程、完美透明的重建音质。作为一种音频编码标准广泛的应用于各个领域,AAC具有得天独厚的优点以及其潜在的市场价值。但是AAC标准算法同样存在不足之处,由于其算法复杂度很高,消耗大量的运算时间及系统资源,存在一定的编码延时,十分不利于当今感知音频编码技术的实时性要求。根据实验研究分析表明,AAC编码过程中的系统资源消耗及运算消耗主要集中在量化与编码、心理声学模型及滤波器组等编码器的各主要模块上。为实现一个低复杂度、高实现效率的音频编解码器,在具有高效性的同时实现完美透明的重建音质,本文充分利用感知音频核心理论——心理声学模型,用客观的参数指标反映主观的听觉效果,针对量化编码过程中的预回声问题提出一种基于动态分块的自适应窗口切换算法。该算法采用动态分块的理念,针对不同采样率的音频信号采用不同的分块效果,结合心理声学模型中的时域掩蔽效应,更准确的判别出瞬变信号并合理的进行窗口切换,将预回声控制并消除在时域阶段,以达到产生的量化噪声不可感知化,实现高效率、高保真的音频编码效果。该算法在保证音频重建质量的前提下,降低了算法的运算复杂度并减小了编码的耗时。更多还原

【Abstract】 With the rapidly development of network information and multimedia technology, the multimedia business need more and more pressing by people, multimedia and computer network technology already seeped to society economy and all aspects of life. In order to facilitate the storage and transmission, the perceptual audio coding technology based on psychoacoustic model is applied widely.Advanced audio coding (Advanced Audio Coding, referred to as AAC) is the most advanced perceptual audio coding technology currently, its advantages are mainly embodied in the following three aspects:the signal with a higher compression ratio, a modular quantization coding and decoding process, and perfect transparent reconstruction quality. AAC is richly endowed by nature as well as its potential market value, as a kind of audio coding standard is widely used in various fields But AAC standard algorithms also exist deficiencies, such as its complexity is very high, also, consuming a large amount of computing time and system resources exists certain coding delay, so is not conducive to the perceptual audio coding technology of real-time requirements. According to the experimental research, AAC encoding process in system resource consumption and operation cost are mainly concentrated in the quantization with coding, psychoacoustic model, filter and encoder each main module.In order to achieve a low complexity and high efficiency audio codec, also, in high efficiency while achieving perfect transparent reconstruction quality, this paper makes full use of perceptual audio core theory and reflection psychoacoustic model, by using objective parameters reflect the subjective auditory effects, and the quantization process pre-echo problems, the author proposed an adaptive window switching algorithm based on the dynamic block. The new algorithm uses a dynamic block concept, according to different sampling rate of audio signals using different blocking effect, combined with psychoacoustic model in time domain masking effect, so we could get more accurately in the transient signal and get reasonable window switches, also, the echo control and elimination in the time domain, in order to achieve the quantization noise generated can not be perceived, to achieve high efficiency, high fidelity audio coding effect. The algorithm reduced the algorithm complexity and the coding time, under the audio reconstruction in the premise of quality,更多还原