轎車車門聲音品質評估與分析

Abstract

本文主要在探討坊間四門房車車門關閉時，車門碰撞聲的聲音品質評估。文中首先回顧國內外有關聲音品質主觀評價的研究方法及成果，顯示當今多改採用心理聲學的方式來評估聲音品質，取代傳統A-weghting的方式，使評估結果更貼近於人耳聽覺的特性。在國外的研究中，對於車門影響最大的兩項心理聲學指標為響度(Loudness)與尖銳度(Sharpness)。本文則另加計算波動度(fluctuation strength)以及粗糙度(Roughness)兩個參數，探討其在車門聲音品質的影響。 本研究首先將隨機取樣的聲音樣品分成A(22輛)、B(15輛)兩組，以樣品的心理聲學參數作統計學上的T-test。在99%的信心水準下，顯示兩筆資料具有一致性。然後，以線性迴歸分析，找出A組中的主觀評審成續與四個參數的量化關係，定出聲音品質優劣的標準。最後以B組的樣品、驗證上述標準的可靠性。 本研究的結果顯示，讓人有較佳感受的房車關門聲，要符合下列原則﹔ 響度須少於 sone；尖銳度須少於 acum；波動度須少於 vacil； 而粗糙度則與音質優劣無明顯關係。 符合上述準則下，音質優劣判定的正確率分別為；響度80%；尖銳度67%；波動度60%。

In this paper, the evaluation of the sound quality due to the closing vehicle door is discussed. Firstly, the research works and results concerning subjective estimation were reviewed. It shows that the psychoacoustics analysis, instead of traditional A-weighting method, were used in sound quality study recently, so that the human auditory characteristics can be simulated more closely. According to the foreign research, the most significant two parameters of sound quality are loudness and sharpness. However, the effects on the sound quality due to the roughness and the fluctuation strength are also discussed in this study. The randomly sampled sound data are divided into two groups. Group A and group B include 22 and 15 sound records respectively. T-test (of 99% confidence level) was used to check the consistency between these two groups. In group A, Linear relationships among those scores of subjective assessments and the sound quality parameters are computed. These criteria are then tested by using the sound record in group B. The result of this experiment shows that when the following conditions are satisfied. The sound will have a “better” quality. These conditions are: Loudness < 16.4 sone ; Sharpness<2.4 acum ; Fluctuation strength <5.8 vacil , however, the relationship between roughness and sound quality is not obvious. The above-mentioned criteria will have judgment correctness of a 80%, 67% and 60% respectively.