Low-frequency Noise Subjective Response To Objective Evaluation

In this thesis, in order to look for effective methods for predicting subjective reactions to low-frequency noises (LFNs), quantitative evaluation models, based on proposed theoretical assessment frame and a large scale of subjective tests, are accomplished.Firstly, taking into account of sound quality theory and its assessment techniques, a LFN subjective assessment composed of two new psychological metrics: Unpleasantness (up) and Unpleasant Probability (P_up), is proposed for the first time. Secondly, a novel objective measurement composed of 3 psycho-acoustical quantities: loudness, sharpness, roughness; and a synthetical metric 'unpleasantness' is presented. Thirdly, 3 large-scale jury tests, adopting Paired Comparison method and Semantic Differential method with the magnitude estimation technique: "5-points, 4-ranks" rating scale, for 200 low-frequency pure tones, 15 and 60 varied low-frequency noise samples, have been carried out, respectively. Finally, two classified misjudgment and the principle for eliminating invalid data, are introduced in detail. Inspecting the obtained data more carefully, the author has found that:1) Unpleasantness can be suitable for the general reaction to LFN perfectly and the LFN subjective assessment system is reasonable and effective.2) There exist some crucial frequencies or bands, which could make LFN unpleasant. For varied low-frequency pure tones, the trend lines of subjective response changing with sound level differ from each other.3) Loudness plays a significant role in shaping subjective responses to LFN; more attention should be paid to sharpness and the whole sound environment. Temporal and spectral characteristics, the low-frequency-range energy ratio influence the LFN's unpleasantness more than overall sound energy itself.