// Effect of Type of Speech Equalization and Averaging Method on the Long-Term Average Speech Spectra of Five Indian Languages and British English
本研究比較均方根均衡(RMSe)與響度均衡(Le)兩種方法，測量不同語言的長期平均語音頻譜(LTASS)｜包括英國英語、印度英語及多種印度語言，透過響度模型驗證Le方法響度的準確性｜結果顯示Le方法能捕捉更廣泛的語言差異，並在印度語言中獲得較高的語音清晰度指數(SII)，對助聽器調校更具感知相關性。
// This study compared two methods—root-mean-square equalization (RMSe) and loudness equalization (Le)—for measuring the long-term average speech spectrum (LTASS) across different languages. Speech samples from British English, Indian English, and several Indian languages were analyzed using both methods, with a loudness model confirming accurate loudness equalization. Results showed that Le captured broader cross-language spectral differences and produced higher speech intelligibility index (SII) values for Indian languages than RMSe. The study concludes that Le-derived LTASS offers a more perceptually relevant basis for hearing aid fitting across languages.
Published 2026年6月3日
// 正在尋找相關文獻...