《实时伴奏软件系统设计于实现》文献综述
【摘要】
计算机实时伴奏系统，由乐音识别、乐谱匹配和乐音播放三个部分组成。主要目的是实现一个人机互动形式的乐器伴奏系统。由于对实时性要求很高，因此对时间安排要求也很高。第一部分是乐音识别，乐音识别主流方法包括时域分析、频域分析、时频域分析和倒谱分析。同时还需要注意降噪的一些问题，我们使用了Mallat算法。第二部分是乐谱匹配。自Roger B. Dannenberg从上世纪80年代开始提出，但有一些问题仍然困扰着为计算机音乐爱好者。第三部分是乐音播放，我们需要一个能够简单并且体积小的播放系统。最后是时间安排，因为是实时性系统，并且根据我的实验，多线程都无法解决“一边录音，一边匹配，一边播放”的问题，需要一些较高的时间安排技术。
【关键词】
Mallat算法 倒谱分析 乐谱匹配 乐音播放 复合音高估计
【Abstract】
The computer real-time accompaniment system consists of three parts: musical tone recognition, musical score matching and musical tone playing. The main purpose is to realize a musical instrument accompaniment system in the form of human-computer interaction. Because of the high requirements for real-time performance, so also the requirements for timing. The first part is musical tone recognition. The main methods of musical tone recognition include time domain analysis, frequency domain analysis, time-frequency domain analysis and cepstrum analysis. At the same time, we need to pay attention to some problems of noise reduction. We use the Mallat algorithm. The second part is score matching. Since Roger B. Dannenberg has been raised since the 1980s, there are still some issues that still plague computer music lovers. The third part is music playback. We need a simple and small playback system. The last is timing, because it is a real-time system, and according to my experiments, multi-threading cannot solve the problem of 'while recording, matching, and playing', which requires some higher scheduling techniques.
【Key words】：Mallat algorithm ; Cepstrum analysis; Score matching Music playback; Multipitch estimation
1 引言
