三维声双耳渲染算法的主观评价分析

Subjective Evaluation and Analysis of 3D Sound Binaural Rendering Algorithms

以耳机重放三维声的形式越来越多的应用于移动终端,如何有效地评价三维声双耳渲染算法成为关键问题。本文针对三维声双耳渲染算法设计了主观评价实验,将评价术语进行归类和分层,采用MUSHRA测试方法对不同评价术语进行0到100的打分。利用方差分析和回归分析,对比不同渲染算法的特点和信号适用性,并探究不同层级评价术语的权重关系。实验结果表明,对于强调移动声源定位及环绕感的声音信号,双耳渲染会产生显著优化作用;对于强调音质和厅堂感的古典音乐信号,双耳渲染算法并不会带来明显的优化。音质是影响渲染算法总体评价最重要的影响因素,其次是空间感。因此双耳渲染算法在权衡音质和声音外化感时,应首先考虑避免明显的音质劣化。

The form of three-dimensional headphone sound reproduction is increasingly applied to mobile terminals,and how to effectively evaluate the three-dimensional sound binaural rendering algorithm has become a key issue.In this paper,a subjective evaluation experiment was designed for the three-dimensional sound binaural rendering algorithm,and the evaluation terms were classified and layered.The MUSHRA method was employed to score different evaluation terms from 0 to 100.The characteristics of different rendering algorithms and signal applicability were compared based on ANOVA(Analysis of Variance)and regression analysis,and the weight relationships of different levels of evaluation terms were explored.The results show that binaural rendering algorithm can produce significant optimization effects for sound signals that emphasizes positioning and spatial perception;whereas the binaural rendering algorithm does not bring significant optimization for classical music that emphasize timbre.Timbre is the most important factor affecting the overall evaluation of rendering algorithms,followed by spatial perception.Therefore,it should first consider avoiding obvious timbre degradation when binaural rendering algorithm weighs timbre and sound externalization.