摘要
针对现实环境中,大多数声音都存在于背景声源的干扰影响中,为了获得需要的声音,在频谱技术分析现有的声信号的基础上,开发出了基于Matlab GUI界面的混合声信号的可视化频谱分离系统。该系统软件可读取并存储采集到的声信号,然后通过分析,做出采集信号或者读取信号的时域图以及该信号的频谱图,通过对频谱图的分析,选择合适的滤波器,设定合理的频率范围,对混合信号进行分离,最终分离出需要的信号。通过对一段wav格式的音乐进行分析分离处理,验证了该系统的可行性以及正确性。
In real-world environment,most speech signals are interfered by noises originated from various sources. The authors developed a Matlab GUI-based visualized spectral separation system based on the existing spectral separation method of acoustic signals. The Matlab system first reads and stores the collected sound signals,presents the time- and frequency-domain spectrums,then compares the frequency spectrums,and elects an appropriate band filter to separate the required acoustic signals. The validity and feasibility of the system are verified by error analysis and an example of processing the measured data of a song in WAV format.
出处
《实验室研究与探索》
CAS
北大核心
2016年第1期68-71,共4页
Research and Exploration In Laboratory
