DISTRIBUTED LOUDSPEAKER ARRAY MEASUREMENT AND CORRECTION TECHNIQUES

It has been common practice in cinema calibration (and some other areas of audio) to use a distributed-source loudspeaker-array for the ambient sound-field reproduction. As a part of the cinema loudspeaker-alignmaent process, systems engineers have been required to analyse and equalise these arrays to a recommended target response by means of the averaging of between four and eight microphones distributed within the calibration area.1, 2 Ostensibly, this procedure would make the array subjectively match the sound of an individual screen-channel that had been calibrated to the same response, but the evidence that this would actually be the case was, at best, tenuous. Other experts had argued that more-compatible results could be achieved if the individual loudspeakers in the arrays were timbrally matched to the screen channel. This paper will present evidence of the finer details and pitfalls encountered when attempting to analyse a complex soundfield created by what is often more than five individual distributed sources, and whether such analysis for the purpose of system correction is even possible.  The system under test was a working, ‘Dolby Features’-certified dubbing theatre, with a low decaytime and a high degree of acoustic control. While not exactly anechoic, the room was defined as ‘non-environment’, with excellent acoustic control, even at very low frequencies. A fundamental reason for having such a low decay-time is so that the natural acoustics of the room will never dominate the intended ambience of the soundtrack, such as in a scene in an open field, for example. The decay time (RT60) in the test room was nominally around 120 ms, as shown in Figure 1 and verified by the waterfall decay-plot shown in Figure 2.