多通道广播源音频处理技术研究

发布时间：2018-06-08 04:20

本文选题：音频广播 + DSP　；参考：《南京理工大学》2015年硕士论文

【摘要】：如今音频广播节目的来源形式众多且制作标准差异较大,这使得各节目源音频信号的电平等参量高低不定,易造成广播中心发射效率降低,播出质量参差不齐。针对此种情况,与某广播设备公司合作研发了此套多路广播源音频处理系统。系统目标任务是对多路音频输入信号的电平和频谱参量进行实时动态处理以达到音频质量和收听响度的完美兼顾。经过对目标任务的研究分析,系统音频处理主体部分通过软件实现,其中使用MATLAB对算法进行仿真设计,C语言完成所有算法编程,VC++语言编写主机控制软件程序。系统解决方案采用DSP+PC,其中核心DSP芯片采用数据处理能力优秀的TMS320VC5509A处理器。通过高效的TLV320AIC23B编解码芯片完成对多路音频信号的采样转换以及5509A自有的USB模块作为数据传输通路,充分保障数据处理的实时性。使得系统具有对多路音频信号参数均衡调整、多频段动态范围控制,实时频率检测功能。音频均衡算法以高/低频shelving滤波器和带通滤波器组合为基础实现,均衡器各段中心频率,增益和带宽均可调整。采用针对性模块客观功能测试及系统联调实际音频测试,证明均衡器对各频段的提升或抑制效果良好,音频改善效果理想。动态范围控制算法采用将其四种工作方式组合形式实现,同时设计了基于LR分割器的具有平坦频响的四频段分割器,通过幅度与时间双向控制,组成多频段可控动态范围处理器,经过模块功能测试及整体实际音频节目客观与主观测试,显示其对音频信号动态范围控制精准,处理后声音自然清晰。频率检测算法是基于复数FFT算法及重心频谱校正算法实现的,可对设定阈值下噪声信息实时检测,进而为噪声抑制提供标准。经仿真及系统实际测试表明,信号频率及幅值测量准确度及精确度高,符合预期设计。
[Abstract]:Nowadays, the source forms of audio broadcasting programs are numerous and the production standards are quite different, which makes the level of audio signals of various program sources fluctuate, which can easily cause the transmission efficiency of broadcast centers to be reduced and the broadcasting quality is not uniform. In order to solve this problem, a multi-channel broadcasting source audio processing system was developed in cooperation with a broadcasting equipment company. The target task of the system is to dynamically process the level and spectrum parameters of multi-channel audio input signal in order to achieve the perfect balance of audio quality and listening loudness. Through the research and analysis of the target task, the main part of the system audio processing is realized by software, in which MATLAB is used to simulate the algorithm and C language is used to complete all the arithmetic programming and VC language to write the host computer control software program. The system adopts DSP PC.The core DSP chip is TMS320VC5509A which has excellent data processing ability. Through the efficient TLV320AIC23B codec chip to complete the multi-channel audio signal sampling and conversion, and the 5509A own USB module as the data transmission path, the real-time data processing is fully guaranteed. The system has the function of adjusting the parameters of multi-channel audio signal, controlling the dynamic range of multi-frequency band and detecting the frequency in real time. The audio equalization algorithm is based on the combination of high / low frequency shelving filter and bandpass filter. The central frequency gain and bandwidth of each segment of the equalizer can be adjusted. The objective function test of targeted module and the actual audio test of system combination are used to prove that the equalizer has a good effect on the enhancement or suppression of each frequency band and the effect of audio improvement is ideal. The dynamic range control algorithm is implemented by combining its four working modes. At the same time, a four-band splitter with flat frequency response based on LR splitter is designed, which is controlled by amplitude and time bidirectional. The multiband controllable dynamic range processor is made up of the module function test and the objective and subjective test of the whole actual audio program. It shows that the dynamic range control of audio signal is accurate and the sound is clear after processing. The frequency detection algorithm is based on the complex FFT algorithm and the centroid spectrum correction algorithm. It can detect the noise information under the set threshold in real time and provide the standard for noise suppression. The results of simulation and system test show that the accuracy and accuracy of signal frequency and amplitude measurement are high and accord with the expected design.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN931

【参考文献】