音訊壓縮的位元分配方法與品質量測比較

Abstract

這篇論文提出一個新的音訊壓縮位元分配準則，依據這個準則我們設計了一個有效率的位元分配方法在 MPEG-1層次三進行實做。我們以大量的客觀量測實驗來輔助編碼器的設計，證明提出的演算法有助於編碼器品質的改進。實驗顯示，在效率，新的位元分配方法比起傳統的兩層迴圈位元分配方法有更低的計算複雜度；在聲音品質的客觀量測，新的位元分配方法能夠獲得較佳的品質。

This thesis presents a new bit allocation criterion for audio coding. Based on this criterion, we design an efficient bit allocation method and implement on MPEG-1 Layer III. Lots of objective measurements are used to help design of the encoder and verify quality improvement from the proposed algorithms. From experiments, the new bit allocation method gains lower computation complexity and better quality of objective measurements compared with traditional two nested iteration bit allocation method.