修改語音基頻曲線與低通濾波處理影響聽障學生國語聲調聽辨成效之研究

A Study of Assisting Hearing-Impaired Students inIdentifying Mandarin Tones by Using Modified Pitch Contours and Low-Pass Filtering

本研究旨在透過電腦，採用擴大語音基頻曲線與低通濾波處理兩種方法，將語音訊號修改成「原始語音基頻曲線」、「設定語音基頻曲線」、「擴大語音基頻曲線」、「原始語音低通濾波」、「濾波設定基頻曲線」、「濾波擴大基頻曲線」六種方式，並分析聽障學生透過六種方式聽辨國語聲調的差異及錯誤類型。研究對象為26 名中重度聽力損失且配戴助聽器的國小高年級與國中學生。本研究主要結果如下：一、藉由擴大語音基頻曲線與低通濾波處理兩種方法，能顯著地提升聽障學生的國語聲調聽辨表現。二、六種電腦語音修改方式下，聽障學生均對於第四聲最容易聽辨，最難聽辨之聲調則因不同電腦語音修改方式而有所差異。主要的聲調聽辨錯誤類型為第一聲混淆為第二聲；第二聲混淆為第三聲；第三聲混淆為第二聲或第四聲；第四聲混淆為第一聲。最後提出有關利用電腦語音修改系統進行聽障學生聲調教學與未來研究之建議。

This study explores the effects of using a computerized speech modification system to expand pitch contours and low-pass filtering in hearing-impaired students' perception of six different types of modified Mandarin tones. Twenty-six students with moderate to severe hearing impairment participated in this study. The results of this study were that hearing-impaired students' ability to identify Mandarin tones was significantly improved by expanding pitch contours and low-pass filtering. Among the six types of computer modified speech, the easiest tone for identification was tone 4, and the most difficult tone to identify varied with the computer speech modification type. Frequent error patterns were that tone 1 was identified as tone 2; tone 2 was identified as tone 3; tone 3 was identified as tone 2 or tone 4; and tone 4 was identified as tone 1. Suggestions for teaching hearing-impaired students Mandarin tones and future studies were discussed based on the findings.