Reducing the variability of loudspeaker preference ratings through digital equalization

It has long been recognized that loudspeaker location and room geometry are sources of variability in listeners' ratings of loudspeakers. When the intent of a listening test is to compare different loudspeakers, the effects of these variables are reduced through proper experimental design. In this manner, subtle differences in sound quality can be revealed. Unfortunately, the environments in which consumer audio products are used do not always conform with those of a controlled listening test. As a result, the sound quality delivered to the consumer may not reflect the fidelity of the audio product. This paper reports some key results of the recently completed Athena project, a collaboration between NRC and the Canadian Audio Research Consortium. The purpose of Athena was to investigate the interaction of loudspeakers and rooms and to develop a means to lessen the deleterious effects on subjective assessments