The paper describes the results of a subjective evaluation experiment that compare two methods for recording a three-dimensional sound field and for reproducing it through loudspeakers in a proper listening room. The first method is the binaural method known as the stereo dipole, based on digital filtering of binaural recordings or binaurally synthesized sound tracks. The digital filters allow reproduction over a pair of closely located loudspeakers. The second method is a software implementation of the well-known ambisonics methodology in which a B-format recording made with a Soundfield microphone, or a synthesized B-format sound track, is reproduced over a 3-D array of eight loudspeakers. The subjective comparisons were made in a listening room fitted with 10 loudspeakers, and the listeners did not know which of the two systems they were hearing. The subjective results show that the stereo dipole system gives a more accurate localization, but that the ambisonics system is more natural and less fatiguing, and that the obtained 3-D effect is not position-dependent, allowing many simultaneous listeners. Both reproduction systems were employed for blind evaluation of the sound field generated by the audio systems of different cars.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.