Perceptual coding, distribution, and enhancement of audio

音频的感知编码、分配和增强

• 批准号: 356668-2007
• 负责人: Kabal, Peter
• 金额: $ 7万
• 依托单位: McGill University
• 依托单位国家: 加拿大
• 项目类别: Strategic Projects Supplemental Competition
• 财政年份: 2007
• 资助国家: 加拿大
• 起止时间: 2007-01-01 至 2008-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=364637
• 关键词: Perceptual; coding; distribution; enhancement; audio

The coding of audio is increasingly important for both storage and streaming of audio content. Taking accountof human hearing (psycho-acoustics) can be beneficial in reducing the bit rate needed to reproduce audiosignals with good quality. Only the frequency components that are audible need be coded. This project willexamine new methods that make even more efficient use of the bits when coding an audio signal, therebyallowing for greater reductions in the bit rate required to represent the signal. The coding will employ arepresentation of the audio signal using the "modulation spectrum". This representation dissects the audiosignal frequency-by-frequency to reveal the perceptually important fluctuations of the frequency components.This project also examines scalable and multiple description coding for audio streaming from, for instance, theInternet. This strategy uses a layered approach to coding. A base layer will allow usable quality, whileadditional layers give successively better renditions. In this way, a single audio stream can be used to serviceterminals with differing quality requirements. For instance, a hand-held wireless terminal might need only abasic quality, while a premium quality is appropriate for a home theatre setting. Multiple description codingprovides assurance of the quality delivered to the end user over error-prone media such as the Internet andwireless links.

音频编码对于音频内容的存储和流媒体来说越来越重要。考虑人类听觉（心理声学）可以有利于降低再现高质量音频信号所需的比特率。只有可听到的频率分量需要编码。该项目将开发新的方法，在编码音频信号时更有效地利用比特，从而允许更大程度地降低表示信号所需的比特率。编码将使用“调制频谱”来呈现音频信号。该表示法逐频地剖析音频信号，以揭示频率分量的感知上重要的波动。该项目还研究了来自例如互联网的音频流的可伸缩和多描述编码。该策略使用分层编码方法。一个基本层将允许可用的质量，而额外的层提供连续更好的再现。通过这种方式，单个音频流可以用于服务具有不同质量要求的终端。例如，手持无线终端可能只需要基本质量，而优质质量适合家庭影院环境。多描述编码通过易出错的媒体（如互联网和无线链路）向最终用户提供质量保证。