摘要
针对混合型电子音乐信号重叠时频自动化分离中存在的中心频率求取不准确、分离性能欠佳的问题,文中设计了一种新型分离系统。通过硬件部分的音乐信号采样传感器和混频器,结合音乐信号的字典表示,引入双曲正切函数降维滤波信号,并精确计算信号中心频率。基于特征信号库,系统计算时频幅度调制系数和状态参数,有效提取源信号中的有用信息,实现自动化分离。系统性能测试显示,所设计的系统可以有效分离混合信号,得到的PESQ分数较高,分离性能更好,为电子音乐信号处理提供了有效解决方案。
In response to the problems of inaccurate center frequency calculation and poor separation performance in the overlapping time-frequency automatic separation of hybrid electronic music signals,this study designs a new separation system.By using the music signal sampling sensor and mixer in the hardware part,combined with the dictionary representation of the music signal,a hyperbolic tangent function is introduced to reduce the dimensionality of the filtered signal,and the center frequency of the signal is accurately calculated.Based on the feature signal library,the system calculates the timefrequency amplitude modulation coefficient and state parameters,effectively extracts useful information from the source signal,and achieves automated separation.System performance testing shows that the designed system can effectively separate mixed signals,resulting in higher PESQ scores and better separation performance,providing an effective solution for electronic music signal processing.
作者
闫雪姣
陈娟
YAN Xuejiao;CHEN Juan(The Art Institute of North University,Taiyuan 030051,China)
出处
《电子设计工程》
2024年第15期180-184,共5页
Electronic Design Engineering
关键词
混合型电子音乐信号
重叠时频
自动化分离
系统设计
hybrid electronic music signal
overlapping time-frequency
automated separation
system design
