Evaluation of recording techniques for three-dimensional audio recordings: Comparison of listening impressions based on difference between listening positions and three recording techniques

Abstract

Various microphone techniques for three-dimensional audio (abbreviated as 3D audio) with playback channels along the height direction, such as NHK 22.2 multichannel sound, Dolby Atmos, and Auro 3D have been proposed. In this study, we compared three microphone techniques for 22.2 multichannel sound, the spaced microphone array, near-coincident microphone array, and coincident microphone array (Ambisonics). First, the evaluation attributes were extracted by referring to the repertory grid technique. Then, using these attributes, participants compared the differences between these microphone techniques, including the difference in the listening position through two experiments. From the results, we observe that the difference, depending on the listening position, was the smallest in the spaced array. In addition, it was estimated that Ambisonics gives the impression of ``hard,'' the near-coincident array gives ``rich'' and ``wide,'' and the spaced array gives ``clear'' and ``presence.'' Furthermore, ``presence'' was evaluated from the viewpoints of clarity and richness of reverberation, with a negative correlation with the spectral centroid and a positive correlation with the reflection from lateral and vertical sides.