针对语音变换的语音篡改检测

针对使用语音变换技术的语音篡改,提出一种自动检测方法。在分析语音变换基本模型和变换语音失真的基础上,提取语音信号的声道参数以及相关的信号统计量,并通过支持向量机递归特征消除法,选择出对语音变换比较敏感的特征作为分类特征,使用支持向量机进行语音变换检测和变换语音的说话人性别判别。对于一种语音变换软件的实验结果表明,该方法具有较高的检测准确率,其中语音变换检测的平均准确率为94.90%,变换语音的说话人性别判别平均准确率为92.09%。     

Poetry I: Teaching verse with CAI

The teaching of literature through CAI raises problems of both a linguistic and instructional nature; student involvement and creativity in studying literature, and especially poetry, is difficult to build into a computer-based lesson. We have confronted these difficulties in the lessonPoetry I, which introduces undergraduates to basic concepts of poetic verse in a design using screen display, speech synthesis, and verse processing to maximize interactivity and student involvement.The lesson contains instructional modules which include the student's composition of a limerick. Computational processing of the limerick's text enables the program to offer guidance as the student composes and revises the verse and gains first-hand experience with metrical language. The significant problems inherent in processing verse are addressed by adapting Digital Equipment Corporation's DECTalk speech synthesizer as an engine converting natural language verse text into accessible strings of phonemic symbols. Although the verse processor cannot scan verse, it can interpret DECTalk's symbols reliably enough to successfully stimulate students' thinking about their verse compositions.Preliminary responses to the program have been favorable and demonstrate its effectiveness in involving students more deeply in learning about poetic verse. This suggests that the techniques of verse processing prototyped inPoetry I might be usefully extended to other types of verse and levels of study.W. Webster Newbold is an Assistant Professor of English at Ball State University, Muncie, Indiana, USA, who has recently served as English Department Computer Coordinator. His research interests include computer-assisted instruction and computer-based composition.Herbert F. W Stahlke is a Professor of English, specializing in linguistics, and currently serves as Associate Director for Academic Computing, Ball State University Computing Services. His research interests include computer-assisted instruction, natural language processing, and lexical databases.     

Effects of real‐time visual feedback on pre‐service teachers' singing

This pilot study focuses on the use real‐time visual feedback technology (VFT) in vocal training. The empirical research has two aims: to ascertain the effectiveness of the real‐time visual feedback software ‘Sing & See’ in the vocal training of pre‐service music teachers and the teachers' perspective on their experience with VFT. Forty participants from an undergraduate music teacher education programme were randomly assigned to three groups. The first two groups were assigned the teacher and student versions of Sing & See, respectively, and the third group served as a control group without access to the software. The experimental groups were given 12 weeks to use the software for vocal training on a self‐regulated basis. The technique of complex selections (TCS) was designed to assess the pitch accuracy and richness of vocal timbre. Pre‐ and post‐test singing tasks showed that the participants from the experimental groups significantly improved their vocal timbre. A questionnaire survey conducted after the training period found that most of the participants in the experimental s were positive about the effectiveness of VFT software in vocal training. As future leaders in the implementation of music education technology, pre‐service music teachers expressed a positive attitude towards the use of VFT as a technology‐enhanced learning (TEL) component in music education.     

高职院校非音乐专业声乐教学方法新探

针对高职院校非音乐专业学生的声乐教学应该采取与传统声乐教学不同的教学方法,让学生既掌握了一定的演唱技能又感受到美,同时又经历了愉快而轻松的学习过程。通过掌握歌唱的基本原则;咬字、吐字训练要点的处理;在歌唱中对歌曲进行艺术处理等方法,取得事半功倍的效果,实践证明这是高职院校素质教育声乐课程教学中的成功经验。     

针对构音异常辅助治疗的声道仿真研究

针对构音异常,本文提出了使用声道仿真来实现辅助治疗的方法。基于声道是一个弯曲的、三维的具有慢时变特性的声学管道,并且在声道中的声波传播是平面波的特性,可以把声道等效于一个具有不同截面的圆柱体或者椭圆体管道。使用极点形式,在牛顿插值的基础上得到共振峰。对声道进行了60段分段,通过经验公式得到声道在不同部位的面积。定义了描述声道特性的9个参数,进而对这9个参数使用Corana算法进行优化。使用辐射模型描述声音从嘴唇辐射出去以后的特性。最后进行声音的合成,这个声音可用于反馈治疗。经过实验证明,这种声道仿真模型可以为制定合适治疗方法提供参考。     

多信息融合的汉语方言辨识

文章提出了一种融合声学、音素配位和韵律特征等多信息融合的汉语方言辨识系统,分析了将语言信息转化为这些特征的实验方法,在此基础上,根据汉语方言辨识的特点,提出了一种基于概率模型的多信息辨识机制,实验结果表明,韵律特征对于短时语音具有很好的辨识效果,而音位配列特征对于长时语音更加有效。对于汉语三种方言的辨识,融合这三种特征的辨识率达95％。     

Finger Braille teaching system-teaching of prosody of finger Braille

Finger Braille is one of tactual communication media of deaf-blind people. Deaf-blind people who are skilled in Finger Braille can catch up with speech conversation, because of prosody of Finger Braille. Features of prosody are： （1） characters at the end of clauses are dotted long; （2） characters of voiced sounds and double consonants are dotted shortly. In this paper, we designed three teaching patterns which taught prosody of Finger Braille （dot conditions about duration of dotting）. Teaching pattern 1 indicated a dot pattern with colored solid background. Teaching pattern 2 indicated a dot pattern with colored empty background. Teaching pattern 3 indicated a dot pattern with colored arrow （long or short）. Every teaching pattern includes 6 colors （red, orange, yellow, green, blue and purple）. In the questionnaire, the yellow pattern of teaching pattern 3 was selected as the most visible pattern. In the evaluation experiment, the yellow pattern of teaching pattern 3, the red and purple patterns of teaching pattern 1 （the least visible pattern） and the existing patterns without teaching of dot conditions are compared. As a result, every subject could dot long at the end of clauses and could dot shortly at the voiced sounds and double consonants. The answers of questionnaire indicated that the yellow pattern of teaching pattern 3 was also the most visible and easiest teaching pattern.     

Analysis of Written and Spoken Form of Hungarian Numbers for TTS Applications

Correct reading of numbers is an essential feature of today's text-to-speech systems (TTS). This article discusses the problems of converting numbers into letters for Hungarian speech synthesis. In the first part of the article, numbers as an organic part of the text are analyzed. Different representations of ordinals, cardinals, telephone numbers, and date and time formats are examined. Special characters, such as comma, period, and plus or minus sign, and various suffixes attached to the numbers may modify the written and spoken form of the number. Numerous conjugated forms are introduced, and examples are shown for these cases. Next, the prosody of spoken numbers is analyzed; intonation, word stress, and pause assignment rules are discussed. Last, implementation questions of number-to-letter (NTL) conversion are addressed with respect to application in TTS and number-to-speech systems.     

Close Shadowing Natural Versus Synthetic Speech

Close shadowing experiments involving natural and synthetic stimuli are described. Preliminary results show that speakers are able to follow natural stimuli with an average delay of 70 ms whereas this delay typically exceeds 100 ms for stimuli produced by text-to-speech systems. A complementary experiment shows that this contrast is mainly due to the inappropriate or impoverished prosody generated by actual text-to-speech systems.