【作者】 彭鹏；

【导师】 褚晶辉；

【作者基本信息】 天津大学 ， 信号与信息处理， 2010， 硕士

【摘要】 AVS音频标准作为我国具有自主知识产权的第二代信源编码标准,它为数字音频设备与系统提供了高效经济的编解码技术方案,是综合算法性能和实现复杂度进行实际应用的适中选择。目前,AVS音频标准的推出时间相对较短,国外对AVS标准仅限于跟踪阶段,主要是国内的研究机构在开展相关研究,且多集中于解码器的实现,对于编码部分的研究并不充分。本文对AVS音频编码算法进行了深入研究和分析,有助于AVS音频编码标准的推广和应用,具有一定的现实意义。本文认真分析了AVS音频标准中的各个模块,从编码框架、算法原理、计算步骤等方面进行了较为详细的说明。接下来,针对暂稳态判决和预回声控制这两个模块,进行了进一步的分析和研究,讨论了目前常见的几种算法。对于暂稳态判决,提出了一种基于峰值均值功率比的暂态信号检测算法,它是基于暂态信号幅度突变的特点,在时域对暂态进行检测的算法。与其它频域检测算法相比,具有复杂度低、时间分辨率高、准确度高等特点。对于预回声控制,提出了一种基于时域平坦化的暂态信号处理方法,它通过对暂态信号加窗,乘以一个下三角函数,使信号突变的位置趋于平坦,然后采用编码平稳信号的方式对其进行编码。实验结果表明,该方法在低码率和不进行频域多分辨分析的前提下,可以获得很好的编码效果,且运算复杂度较低,可以取代AVS标准中采用的频域多分辨率分析技术。更多还原

【Abstract】 AVS audio standard with independent intellectual property rights, as the second generation source coding standards, it provides a highly efficient technical program of coding and decoding for digital audio equipments and systems, it’s a moderate choice during algorithm performance and complexity for practical applications.Currently, AVS audio standard has just released, in other country, the AVS standard is limited to tracking, most of work is carried out by domestic research institutions, and focused on the implementation of decoder, the study for coding part is not sufficient. In this paper, in-depth research and analysis is done for the AVS audio coding standard, and it will contribute to the promotion and application of the standard. So, it has some practical significance.First, this paper carefully analyzes each module of the AVS audio standard from the aspects of coding framework, algorithm theory and calculation steps. Next, for the temporary steady-state decision and pre-echo control, further analysis and research are carried out and the current common algorithms of them are discussed here. Then, a transient signal detection algorithm based on peak average power ratio is proposed, which uses the characteristic of mutation rate in time domain, comparing with other frequency domain detection algorithm, it has the feature of low complexity, high time resolution and high accuracy. At last, the author proposed a transient signal processing method which is based on flattening in time domain, it uses a trigonometric function to multiply with the transient signal, so the signal will tend to flat, then using the method of coding stationary signal to encode the transient signal. The experimental results show that the new algorithm can get a good coding performance with low computational complexity, so it can replace the multi-resolution analysis technique which is used in AVS audio stardard.更多还原