Real rooms vs. artificial reverberation: An evaluation of actual source audio vs. artificial ambience in the rear height channels of immersive audio systems

Abstract

Many spatial audio researchers and content producers agree that the best source material for height channels in immersive audio is provided by the capture of actual elevated channels in the room. Particularly for music recording, this technique is preferred as opposed to signal processing, providing a more natural and realistic impression of immersion. While previous work has proven this to be the case in the front channels of various 3D playback systems such as 22.2, the content of the rear height channels has not been specifically evaluated in this respect. Multichannel audio recording, specifically 3D recording, can be a cumbersome task as the channel counts expand—and so the question arises—is it really necessary to capture discrete rear height information? This research compares four height channel capture points are compared to two capture points applied to the front height channels in conjunction with artificial reverberation in the rear channels. A two-part study is employed—the first is a simple ABX test to determine discriminability between the real rooms and the artificially generated version. Part two is a preference test, based on several standard acoustic/perceptual descriptors, revealing the subtle differences between real and artificial rear height channel information.