Visual Speech Articulation Training Aid

本项目提出的研究是研究和优化提取独立于说话人的基于声学的语音参数的方法，这些参数表示语音的语音内容，并确定可以在听力受损的发音训练辅助设备中显示这些特征的转换。拟议的项目是先前由美国国家科学基金会资助的研究的延续和扩展。这项工作的一个主要组成部分将是协调收集来自听力正常和听力受损听众的语音样本的大型数据库，这不仅是拟议工作所需要的，而且也将促进该领域的其他研究工作。我们将进行大量的实验来优化语音参数到显示参数的转换，从而使在孤立语境或音节语境中产生的音素和音素的独特特征能够很容易地根据其显示特征进行区分和识别。非线性/线性变换的组合将听觉刺激的声学测量转换为视觉显示表示。声学特征将从全局频谱形状、基频和短时能量中提取出来。特征提取/分类过程将针对每一对英文手机进行单独优化。＊＊＊

9411607 Zahorian The research proposed in this project is to investigate and optimize methods for extracting speaker-independent acoustically- based speech parameters that signal the phonetic content of speech and to determine transformations which can display these features in an articulation training aid for the hearing impaired. The proposed project is a continuation and extension of research previously funded by the NSF. A major component of the effort will be to coordinate the collection of a large data base of speech samples from both normally-hearing and hearing- impaired listeners, which will be required not only for the proposed work, but which will also facilitate other research efforts in the field. Extensive experiments will be conducted to optimize the conversion of speech parameters to display parameters so that phonemes and distinctive features of phonemes, produced either in isolation or syllabic contexts, can readily be discriminated and identified based on their display characteristics. A combination of nonlinear/linear transform will convert acoustic measurements of auditory stimuli to a visual display representation. Acoustic features will be extracted from global spectral shape, fundamental frequency, and short-time energy. The feature extraction/classification process will be individually optimized for every pair of phones in English. ***