The Influence of Lombard Effect on Speech Recognition
The Influence of Lombard Effect on Speech Recognition
复制标题
DOI：
10.5772/17520
复制
发表时间：
2011-06
期刊：
影响因子：
--
通讯作者：
D. Vlaj;Z. Kacic
中科院分区：
文献类型：
--
作者：
D. Vlaj;Z. Kacic
伦巴第效应的起源可以追溯到一百年前。1911年，Etienne Lombard发现了噪音环境下言语的心理效应（Lombard，1911）。隆巴德效应是一种现象，其中说话者在存在大的背景噪声的情况下增加他们的声音水平，并进行几次声音变化以提高语音信号的可懂度（Anglade & Junqua，1990; Bond等人，1989; Dreher &奥尼尔，1957;埃根，1971; Junqua，1996; Junqua & Anglade，1990;货车萨默斯等人，1988年）。在当今语音识别应用中，Lombard效应会出现在许多领域，在这些领域中，自发的和会话式的语音通信会发生在不受控制的声学环境中。对伦巴第效应有两种主要的解释。第一种观点认为，这种效应是一种生理性的听觉-发声反射（Lombard，1911），第二种观点认为，Lombard的变化是由说话者对可懂度降低的补偿引起的（Lane & Tranel，1971）。一些作者还认为，这两种机制都可能有助于说话者在嘈杂环境中所做的改变（Junqua，1993）。关于伦巴第效应现象的文献的详细调查在（Lane & Tranel，1971）和最近的（Junqua，1996）中进行。进行的研究表明，伦巴第语在许多方面与正常语言不同。伦巴第语语音特征的主要变化可以在语音水平、基频和元音持续时间的增加以及F1和F2的共振峰中心频率的移位中看到（Anglade & Junqua，1990; Applebaum等人，1996; Junqua，1996; Junqua & Anglade，1990）。在（汉利& Steer，1949）中还报道了当在噪声环境中产生语音时，说话速率可能降低。汉森，1988年也对不同类型的重音（包括伦巴底效应）下的语音进行了详细的声学和语音分析。研究表明，在隆巴德效应下，元音的时长增加，而清音塞音和擦音的时长减少。此外，频谱倾斜减少意味着在隆巴德效应下的高频分量的增加。在这两种情况下，音高和第一共振峰位置也会增加。此外，能量迁移从低，高频率的元音，并从低到高频段的清音塞音和摩擦音的运动范围的中间。除上述情况外，在（Junqua，1993年）中还注意到男女讲话者之间的差异。另一方面，成人的隆巴德变化大于儿童，自发言语的隆巴德变化大于阅读任务的隆巴德变化（Amazi & Garber，1982; Lane & Tranel，1971）。
The origin of Lombard effect dates back one hundred years. In 1911 Etienne Lombard discovered the psychological effect of speech produced in the presence of noise (Lombard, 1911). The Lombard effect is a phenomenon in which speakers increase their vocal levels in the presence of a loud background noise and make several vocal changes in order to improve intelligibility of the speech signal (Anglade & Junqua, 1990; Bond et al., 1989; Dreher & O'Neill, 1957; Egan, 1971; Junqua, 1996; Junqua & Anglade, 1990; Van Summers et al., 1988). In nowadays speech recognition applications appearance of Lombard effect can be expected in various domains, where spontaneous and conversational speech communication will take place in uncontrolled acoustic environments. Two main interpretations of the Lombard effect have been proposed. The first argues that the effect is a physiological audio-phonatory reflex (Lombard, 1911), the second that Lombard changes are motivated by compensation on the part of the speaker for decreased intelligibility (Lane & Tranel, 1971). Some authors have also argued that both mechanisms may contribute to the changes made by the speaker in noisy environments (Junqua, 1993). Detailed surveys of the literature on the Lombard effect phenomenon was made in (Lane & Tranel, 1971) and more recently in (Junqua, 1996). The conducted research showed that Lombard speech is different from normal speech in a number of ways. The main changes of characteristics of Lombard speech can be seen in increase in voice level, fundamental frequency and vowel duration, and a shift in formant center frequencies for F1 and F2 (Anglade & Junqua, 1990; Applebaum et al., 1996; Junqua, 1996; Junqua & Anglade, 1990). It was also reported in (Hanley & Steer, 1949) that speaking rate may be reduced when speech is produced in a noisy environment. A detailed acoustic and phonetic analysis of speech under different types of stress including the Lombard effect was carried out also in (Hansen, 1988). The studies showed that under the Lombard effect, duration of vowels increase while that of unvoiced stops and fricatives decrease. Also, spectral tilt decreases implying an increase in high-frequency components under the Lombard effect. An increase in pitch and first formant location also occurs in both cases. Also, energy migration from low and high frequency to the middle range for vowels, and movement from low to higher bands for unvoiced stops and fricatives was observed. In addition to the above, differences between male and female speakers was noted in (Junqua, 1993). Lombard changes are on the other hand greater in adults than in children and in spontaneous speech than in reading tasks (Amazi & Garber, 1982; Lane & Tranel, 1971).