行動通訊網路之混合音訊編碼技術   免費試閱

A Hybrid Audio Coding Scheme for Mobile Communications Network

新世代行動裝置具備多樣化之功能，因此各種混合著語音與音樂內容之音訊，以下載或串流方式傳送到行動裝置上進行播放，讓市場上對能夠編碼混合音訊內容及對於不同音訊能具備一致性品質的新編解碼器有強烈之需求。本論文對此一複合式音訊編碼器進行研究，提出一新型複合音訊編碼架構，整合「高效率先進音訊編碼」（HEAAC）與「延伸型寬頻調適性多位元率」（AMRWB+）兩種編碼器。由於語音與音樂訊號特性之差異，對訊號進行分類以維持其解碼訊號品質極為重要，藉由基於訊號分類之開迴路編碼選擇機制，讓所提出之編碼器能對輸入訊號選擇合適之編碼核心，使其效能達到最佳化。研究者將所提方法、HEAAC 編碼器與AMR-WB+ 編碼器之輸出，以客觀品質量測進行效能評估，實驗結果顯示，本論文所提之方法在相同之位元率與音訊類型，其品質不會低於現有的MPEG 技術。

New generation mobile devices become multi-functional. Various types of mixed contents can be played on or streamed to mobile devices. It makes a strong market requirement for a codec that is able to provide consistent quality for mixed speech and music content. In this project, we investigate a new hybrid audio codec. The proposed codec combines techniques from both HEAAC and AMRWB+ by means of switching between the core coders of the two standards. Owing to the different property between different signals, it is important that signal classification must be made to retain the quality of decoded signal. The switched coding paradigm consists of appropriately selecting the coding mode, which is designed to perform best for a certain category of signal, by using an open-loop audio coding selection scheme controlled by a signal classification module. An objective quality measurement for the proposed method and two reference systems (HEAAC and AMR-WB+) are presented. The experimental results shows that the quality of proposed scheme is not worst than the performance of current MPEG technology both are operated at the same bit rate and audio category.