【摘要】 随着三维(Three Dimension,3D)音频技术的发展,与之相配合的三维音频质量评价方法也越来越受到广泛的关注。当前的三维音频质量评价方法主要包括两大类:主观评价方法与客观评价方法。主观评价由于其直观性和可靠性而成为音频质量评价的黄金标准,主要包括ITU国际标准中的带隐藏参考和基准的多激励测试(MUSHRA,ITU-R BS.1534)以及带隐藏参考的三次听音双盲听评法(ITU-R BS.1116)。主观评价方法认可度高,但操作耗时耗力,研究者也希望开发一种快速方便的客观评价工具。目前国际标准中针对音频信号的客观模型主要是感知音频质量客观测试方法(PEAQ),但相较于主观评价方法,它的准确度仍不是很理想,尤其是在针对多声道的三维音频质量进行评价时缺少很好的解决方案。本文对目前主要的三维音频质量主客观评价标准进行了阐述和分析,并提出一种改进的基于双层结构和数据挖掘方法的三维音频质量客观评价模型,其准确度优于现有的PEAQ方法。更多还原

【Abstract】 With the development of 3 D audio technology,the corresponding 3 D audio quality evaluation methods have been gaining widespread attention. Currently,3 D audio quality evaluation methods mainly including: subjective evaluation method and objective evaluation method. The subjective evaluation methods consist of MUSHRA and ITU-R BS. 1116,they are intuitive and reliable which make them become the "gold standard". However,due to its time-consuming,researchers hope to develop the objective evaluation method to make it easier. At present the international standard objective method is PEAQ,but its accuracy isn’t good enough,especially for the 3 D audio quality evaluation. In this paper,we analyzed the main 3 D audio subjective quality evaluation standards and proposed a novel two-layer model based on Multiple Linear Regression( MLR) and Neural Network( NN). The experimental results show that the accuracy of the method is preferable to that of the existing PEAQ method.更多还原