Marxim Rahula Bharathi B, Balaji N S, Meena R, Raja Chandra Sekar M, Sivalingam Krishna Moorthy, Yafang Yan, Elumalai P V
Department of Mechanical Engineering, Aditya University, Surampalem, 533437, Kakinada, Andhra Pradesh, India.
Department of Mechanical Engineering, SRM Institute of Science and Technology, Tiruchirappalli Campus, Tiruchirappalli, 621105, India.
Sci Rep. 2025 Apr 23;15(1):14044. doi: 10.1038/s41598-025-97823-5.
Wavelet analysis is a prominent time-frequency analysis method in investigating various signals such as speech, vibration, acoustic signals, ultrasound, and underwater acoustic signals. Throughout the coronavirus pandemic, people have adopted diverse face shields and face masks, which have caused difficulties in understanding speech. To address this issue, the wavelet transform (WT), a proven effective method, can be implemented. Time-frequency analysis serves as a standard approach since it combines useful information between time-domain observations and frequency-domain data. However, the selection of an appropriate mother wavelet represents the main obstacle when using WT. The same signal produces different outcomes when analyzed with various mother wavelet selections. In this research, speech signals were obtained under various conditions of face masks and face shields. This work proposes the COPRAS (COmplex PRoportional ASsessment) technique to select the appropriate mother wavelet function. Maximum Cross-Correlation Coefficient (MCC) and Maximum Energy to Shannon Ratio (MEER) evaluation criteria are utilized to rank the better mother wavelet function. From the results, the proposed methodology establishes a comprehensive protocol for selecting mother wavelet for the speech signal in various conditions.
小波分析是一种重要的时频分析方法，用于研究各种信号，如语音、振动、声学信号、超声波和水下声学信号。在整个新冠疫情期间，人们使用了各种各样的面罩和口罩，这给语音理解带来了困难。为了解决这个问题，可以采用已被证明有效的小波变换（WT）方法。时频分析是一种标准方法，因为它结合了时域观测和频域数据之间的有用信息。然而，选择合适的母小波是使用小波变换时的主要障碍。用不同的母小波选择分析同一信号会产生不同的结果。在本研究中，在各种面罩和口罩条件下获取了语音信号。这项工作提出了复比例评估（COPRAS）技术来选择合适的母小波函数。利用最大互相关系数（MCC）和最大能量与香农比（MEER）评估标准对更好的母小波函数进行排序。结果表明，所提出的方法建立了一个在各种条件下为语音信号选择母小波的综合方案。
