Adaptive grouped paired comparison: Theory and selection of seeds

In the subjective sound quality evaluation based on paired comparison method, time cost grows with square of the number of samples, which makes the evaluation of large quantity of samples difficult. An Adaptive Grouped Paired Comparison (AGPC) design was proposed to greatly decrease time requirement and difficulties of subjective comparison test, and to avoid the reduction of the precision when improper reference samples were chosen in common grouped paired comparison by adapting the seeds. The mathematic theory and implementation method were given, including the experiential formula. It was shown from computer simulations and the subjective sound quality evaluation that the results of the proposed AGPC method were in good agreement with the true merits of samples and those obtained from traditional paired comparison.