Perceptual equalization in near-speaker panning

In recent years, extensive work has been done on spatial sound reproduction techniques involving multiple loudspeakers. Techniques like wavefield synthesis or crosstalk cancellation aim to simulate the soundfield at the listener’s ear for sounds from different directions. In the simple approach of amplitude panning sounds are played simultaneously from two loudspeakers while variations in level between speakers alter perceived direction. Previous studies investigated the perceived direction (e.g. [5]), but we ask if panning can truly replace a loudspeaker at the panned location by not permitting audible panning errors. This requires not only that panned direction is accurate, but also that source width, loudness, timbre, and temporal aspects are equally reproduced. Obviously, for large loudspeaker spacings panning errors will be audible, but if both speakers are spaced only 7.5 or 15 apart, which is large compared to requirements for wavefield synthesis, errors might be very small. Informal listening convinced us that errors are nevertheless clearly audible. In the first part of the manuscript these errors are characterized, while the second suggests three methods for panning equalization. An in-depth evaluation of the equalization techniques found that all three approaches lead to similar results, that a filter derived from a spherical head model is sufficient to render panning errors inaudible, and that errors will be inaudible in all practical applications. The simplicity of the equalization approach ensures that results are also valid for listeners wearing hearing devices.