Quality of service evaluation for spatial audio coding and processing systems

这个合作项目将进行基础研究，以一种客观的方式测量声音处理设备的感知空间质量。与未受损的参考相比，这将使空间质量退化的自动评估成为可能。这样的系统需要能够通过任意渲染格式和扬声器布局来评估空间声音的再现质量，因此本研究基于对聆听位置的再现声场进行建模和评估的思路，而不是对再现的音频信号本身进行测量。这是必要的，因为当前的音频质量测量标准没有考虑到空间因素。测试信号和感知相关测量将被设计用于评估被测设备的空间性能，然后预测参考空间音频表示和受损表示之间的感知“差异等级”。这就需要将客观测量结果与听力测试结果联系起来，并开发考虑到所涉节目材料类型的对上下文敏感的统计模型。

This collaborative project will undertake basic research leading to a means of measuring the perceived spatial quality of sound processing equipment in an objective fashion. This will enable the automatic evaluation of degradations in spatial quality, as compared with an unimpaired reference. Such a system needs to be able to evaluate the quality of spatial sound reproduced by arbitrary rendering formats and loudspeaker layouts, so the research is based on the idea of modelling and evaluating the reproduced sound field at the listening position rather than measuring the reproduced audio signals themselves. This is needed because current standards for audio quality measurement do not take spatial factors into account.Test signals and perceptually relevant measurements will be designed to evaluate the spatial performance of the device under test followed by the prediction of a perceptual 'difference grade' between a reference spatial audio representation and an impaired representation. This requires the mapping of objective measurements to listening test results and the development of context-senstive statistical models that take into account the type of programme material involved.