A multi-criteria subjective evaluation method for binaural audio rendering techniques in virtual reality applications

In state-of-the-art Virtual Reality (VR) technology, the audio experience tends to be dynamic and interactive, which is not feasible to be evaluated with conventional audio quality evaluation methods. For audio-visual VR scenarios, significant results can be drawn when overall perceived quality method is adopted, but the technical details become difficult to be located and analyzed. This paper proposes a multi-criteria subjective evaluation method to evaluate the binaural audio rendering techniques utilized in VR applications by setting up disparate influencing indicators in multiple scenarios. Subjective experiments are conducted on the test platform loaded with two VR audio renderers. The test results show that this method can be used to evaluate the performance of the renderers from different perspectives.