陈浩 吴煜祺

摘  要： 由于受到音乐专业性强、乐理知识复杂、变化多样等因素影响，导致音乐特征识别难度较大，为此设计基于物联网技术的音乐特征识别系统。系统物理感知层在不同位置布设声音传感器采集音乐原始信号，采用TMS320VC5402数字信号处理器展开音乐信号分析处理;网络传输层将处理完成音乐信号，传输至系统应用层中的音乐信号数据库中;应用层中的音乐特征分析模块，采用动态时间规整算法，获取测试模板和参考模板间最大相似度，实现音乐信号特征识别，并依据识别结果识别音乐曲式和音乐情感对应音乐特征内容。实验结果表明，该系统运行稳定，可采集到高音质音乐信号，且能正确识别音乐曲式特征和情感特征。

关键词： 音乐特征识别; 物联网; 系统设计; 信号采集; 信号处理; 实验分析

中图分类号： TN912.34?34; TP311.13            文献标识码： A                     文章编号： 1004?373X（2020）10?0043?03

Design of music feature recognition system based on Internet of Things technology

CHEN Hao1， WU Yuqi2

（1. Shiyuan College of Nanning Normal University， Nanning 530226， China; 2. Jiaxing University Nanhu College， Jiaxing 314001， China）

Abstract： In allusion to the influence of strong music specialization， complex music theory knowledge， various changes and other factors， it is difficult to recognize the music features. Therefore， a music feature recognition system based on Internet of Things technology is designed. The sound sensors are equipped at the different locations in the physical perception layer of the system to collect the original music signal. The music signals are analyzed and processed by the digital signal processor TMS320VC5402. The processed music signals are transmitted to the music signal database in the application layer of the system through network transmission layer.  The maximum similarity between the testing template and the reference template is obtained with the music feature analysis module in the application layer and by means of the dynamic time warping algorithm， so as to realize the music signal feature recognition， and identify  the musical form and musical emotion corresponding to the music feature contents according to the recognition results. The experimental results show that the system runs steadily， can collect high?quality music signals， and can identify the musical form features and emotional features correctly.

Keywords： music feature recognition; Internet of Things; system design; signal acquisition; signal processing; experimental analysis

0  引  言

音乐特征识别是基于语音识别发展艺术，通过音频信号获取音乐内容，进一步得到音乐曲式、情感等音乐特征[1]。音乐特征识别的研究涉及到许多方面，如心理声学、乐器分析及乐理知识等。目前音乐识别系统未能大范围应用，是由于缺少有利于提升性能信息的系统整体框架设计。物联网技术的出现使音乐特征识别系统的实现成为可能[2]。物联网技术通过有线/无线网络间信息实时传递，实现音乐信号的智能化采集、处理、分析等，其具有感知全面、传递可靠、方便快捷等优势[3]。设计基于物联网技术的音乐特征识别系统，利用物联网技术实现音乐信号的感知、传输及识别。……