MODELING REFRIGERATOR NOISES

摘要

In this paper, we propose objective methods which predict perceptual refrigerator noise levels. Eight home refrigerators are chosen and their noises are recorded in an anechoic-chamber and a real-life apartment. Then, we compute 12 features which are used to predict perceptual refrigerator noise levels. In order to obtain subjective data, subjective scores were obtained from 100 evaluators using the paired comparison (PC) method. Finally objective models are developed, which provide high correlations with the subjective data.