MPEG環繞音訊編碼之效率及品質改進

Abstract

MPEG環繞(MPEG Surround)是由ISO/IEC MPEG所新制定的一套標準，它能夠在非常低的位元率下壓縮多聲道的音訊訊號。在本篇論文中，我們設計並實現了4個MPEG環繞標準中所制定的工具，可以用來增進編碼效率以及音效品質。 我們首先設計一個方法選擇適當的子參數集(parameter subset)編碼模式，此模式利用了時間軸上的相關性(correlation)來減少資料量；此外我們也設計了一個流程用來決定參數頻帶步幅(parameter band stride)，利用了頻率軸上的相關性，並且和子參數集配對(pair)結合。此兩模組可以同時運作並減少資料量。我們的實驗結果顯示，在允許一些誤差的情況下，可以使位元率(bitrate)減少25~55%。除此之外，參數的時間解析度也可以根據訊號的內容做動態的調整；最後，利用適應性參數平滑化(adaptive parameter smoothing)工具可以減少因為粗略量化所造成的影響，此兩模組對於某些測試曲子，可以增進其輸出音效品質。

The MPEG Surround is an efficient audio coding standard recently defined by the ISO/IEC MPEG committee. It is able to compress the multi-channel audio signals at a very low bit-rate. In this thesis, four tools in the MPEG Surround specifications are designed and implemented to enhance the coding efficiency and quality. We first design a method that selects the best parameter subset coding mode that exploits the correlation along time axis. Then, combined with pairing of subsets, the parameter band stride mechanism exploiting the correlation along frequency axis is proposed. These two methods can work together to reduce the compressed data. With a small error tolerance, our experimental results show that the bitrate reduction can reach 25~55%. Furthermore, the time resolution of MPEG Surround parameters can be adjusted dynamically, depending on the contents of signals. Finally, to reduce the artifacts caused by coarse quantization, the adaptive parameter smoothing tool is also employed. These two modules can enhance the output audio quality particularly for certain audio sequences.