摘要
传统的音乐多参数识别方法在记录音符时,常出现多次识别、错误识别问题。为此,在人机协作的环境中,应用人工神经网络设计了新的音乐多参数识别方法,首先对音乐抽象空间进行具体化处理,从而确定音乐创作情感和风格,在此基础上,识别音乐的旋律特征,并在合成旋律的过程中过滤掉多余的音符,然后利用人工神经网络实现对音乐参数的识别。实验结果表明:与传统识别方法相比,应用人工神经网络的识别方法输出结果的音轨质量高,且未出现多识别的音符,充分证明了该方法的应用优势。
When the traditional music multi-parameter recognition method is used to record notes,there are often such problems as multiple recognition and error recognition.Therefore,a new music multi-para-meter recognition method is designed by using the artificial neural network.In the environment of human-computer collaboration,the abstract space of music is first concretely processed so as to determine the emotion and style of music creation.On this basis,the melody characteristics of music are recognized,and the redundant notes are filtered out in the process of synthesizing melody.Then,the artificial neural network is used to recognize the music parameters.The experimental results show that compared with the traditional recognition method,this new method has higher output track quality,and there is no multiple recognition of notes,which fully proves the application advantages of this method.
作者
王瑞
李珊
齐建立
WANG Rui;LI Shan;QI Jianli(School of Music and Dance, Fuyang Normal University, Fuyang 236037)
出处
《常州工学院学报》
2022年第3期34-38,共5页
Journal of Changzhou Institute of Technology
关键词
人工神经网络
音乐参数识别
旋律特征
人机协作
artificial neural network
music parameter recognition
melody characteristics
human-computer collaboration
