在压缩域对音频信号进行处理与分析，避免了传统方法中的音频解码环节，降低系统复杂度和系统延迟，对媒体大数据时代的海量音频处理与分析具有重要意义。本项目研究在音频变换域MDCT/ODFT中进行参数估计的关键技术，并探讨其应用于压缩域音频处理与分析、压缩编码/参数分析一体化音频处理系统时的核心技术内容。研究内容：复杂模型下MDCT/ODFT域音频参数估计方法；MDCT/ODFT域音频特征参量提取方法；ODFT域音频多通道与高频带编码参量提取；压缩音频处理与分析仿真验证平台。创新点：结合实际音频信号特征，实现复杂模型条件下的参数估计方法；结合压缩域音频处理分析需求，实现压缩域音频特征参量提取；结合ODFT域分析的特征，实现压缩编码与信号分析一体化的低复杂度音频系统方案。本项目研究将提供有效的变换域信号参数估计方法，为压缩域音频处理与分析提供技术支持和实现方案，并探索新的音频处理平台框架。

Compressed domain audio signal processing and analysis are of vital significance for large scale audio processing because it can avoid the decoding modules required by conventional methods, helping to reduce the system complexity and latency. We plan to do research on the key techniques in parameter estimation methods in MDCT/ODFT domain that are used in audio compression. We also plan to study the key techniques that are used in compressed domain audio processing and analysis, and in building a unified audio compression and parameter extraction system. To be specific, our research is composed of the following three facets: 1) parameter estimation methods of audio signals with complex models in MDCT/ODFT domain; 2) characteristic parameter extraction for audio signals in MDCT/ODFT domain; 3) Multi-channel and high-frequency bands coding parameters extraction in ODFT domain; 4) simulation and verification platform for compressed audio processing and analysis. Innovation points are: 1) Propose parameter estimation methods for complex models that take the actual audio signal characteristics into account; 2) Propose compressed domain audio characteristic parameter extraction methods; 3) A solution for low-complexity audio processing platform with unified compression and signal analysis functions. Out research will provide effective transform domain signal parameter estimation methods, and it will provide the necessary technical supports and a valuable reference for audio signal processing and analysis in the compressed domain. Further more, it will explore a new audio processing platform.