頁籤選單縮合
題 名 | 以巴克頻譜失真為基礎之音高週期估測=Pitch Detection Based on Bark Spectrum Distortion |
---|---|
作 者 | 王德譽; 劉啟昇; | 書刊名 | 暨大學報 |
卷 期 | 7:1 2003.06[民92.06] |
頁 次 | 頁169-180 |
分類號 | 312.23 |
關鍵詞 | 高音估測; 巴克頻譜失真; 正弦語音模型; Pitch detection; Bark spectral distortion; Sinusoidal speech model; |
語 文 | 中文(Chinese) |
中文摘要 | 音高估測在語音信號處理中是一個相當重要的問題,目前研究不論是時域或頻域之估測,基本上都是藉由比較預估波形與原始波形的關聯性或信號雜訊比來決定音高週期。對於完美有聲語音,簡單的波形比對即可達到正確之音高估測,然而實際語音是時變信號,若音高估測不當,將使合成之語音品質嚴重下降。巴克頻譜分析包含頻率扭曲、臨界頻帶積分、等響度預強調及主觀響度轉換,可有效將人耳對語音頻率及響度等非線性響應等化。故本論文以巴克頻譜失真為評估標準,比較原始語音及預估諧波頻譜,求得最佳之音高週期,並決定有聲語音機率。模擬結果顯示以巴克頻譜失真為基礎之音高估測,配合正弦語音模型,可有效合成高品質之語音。 |
英文摘要 | Pitch detection is an important issue in a variety of speech applications. Many pitch detection algorithms (PDAs), both in time and frequency domains, have been proposed for the voiced/unvoiced detection and pitch abstraction. During highly voiced stationary sections of speech, the pitch period is easily observed using PDAs based on peak detection in the time domain, such as auto-correlation function, zero crossing rate and average magnitude difference function (AMDF). In the frequency domain, the PDAs utilize the harmonic structure of the speech spectrum or the spectral auto-correlation property. All of the proposed algorithms have their limitations, and no presently available PDAs can be expected to give perfectly satisfactory results across a wide range of speakers, applications, and operating environments. In this paper, we provide a pitch detection algorithm based on Bark Spectral Distortion (BSD), in which several known features of the perceptual processing of speech sounds by the human ear are emulated. The experimental results show that the proposed method is more accurate than the sinusoidal speech model, and the reconstructed speech sounds more natural. |
本系統中英文摘要資訊取自各篇刊載內容。