2008年5月29日 星期四

Silent Speech Translation by interACT

http://tw.youtube.com/watch?v=aMPNjMVlr8A

This prototype demonstrates the electromyographic (EMG) silent speech recognition research in InterACT. The EMG research focuses on making use of the EMG signal generated by the articulatory muscles to recognize silent speech.Th e demo person will speak in silent Mandarin speech and the system translates it to English or Spanish.

Electromyography

Area Functions and Articulatory Modeling

Chp.5
5.2 Background

5.2.1 The F-pattern
1. F-pattern: The filter function of the VT contains only resonance, labeled F1, F2, F3, F4, and F5.
2. Main gestures of those formants
3. Spectrograms of oral vowels and consonants
4. Spectrograms of stops, voiceless fricatives, voiced fricatives

5.2.2Maeda’s Articulatory Model
1. Guided Principal Component Analysis of Midsagittal X-ray tracings
2. Tongue dorsum position, tongue shape, tongue apex position, lip aperture, lipper, and larynx height
3. The three different types of input used in the model: seven articulatory parameters, X-ray or Magnetic Resonance Imaging data, area fuction table

Etienne Jules Marey

http://en.wikipedia.org/wiki/%C3%89tienne-Jules_Marey
http://www.ctie.monash.edu.au/hargrave/marey.html

2008年5月15日 星期四

Questions and Sub-questions in Phonology

How phonology involve into its current state?
How did language and speech arise or evolve in our species? Why is the vocal apparatus different as a function of the age and sex of the speaker? What is the relation, if any, between human speech and non-human communication?
What are phonological universals?
How is language and its parts, including words and morphemes, represented in the mind of the speaker; how is this representation accessed and used? How can we account for the variation in the phonetic shape of these elements as a function of context and speaking style?
How, physically and physiologically, does speech work---the phonetic mechanisms of speech production and perception, including the structures and units it is built on?
Why and how does pronunciation change over time, thus giving rise to different dialects and language and different forms of the same word or morpheme in different contexts? How can we account for common patterns in diverse languages, such as segment inventories and phonotactics?
How is speech acquired as a first language and as a subsequent language?
How is sound associated with meaning?
How can we ameliorate communication disorders?
What is language disorder
What factors contribute to phonological changes?

What new methods can we use in phonology?

Where do phonologists acquire information?

What is the way phonology going to be changed?

Important Questions in Phonology for Me

8 questions in our textbook
1. How is language and its parts, including words and morphemes, represented in the mind of the speaker; how is this representation accessed and used? How can we account for the variation in the phonetic shape of these elements as a function of context and speaking style?
2. How, physically and physiologically, does speech work---the phonetic mechanisms of speech production and perception, including the structures and units it is built on?
3. Why and how does pronunciation change over time, thus giving rise to different dialects and language and different forms of the same word or morpheme in different contexts? How can we account for common patterns in diverse languages, such as segment inventories and phonotactics?
4. How can we ameliorate communication disorders?
5. How can the functions of speech be enhanced and amplified, for example, to give permanency to ephemeral speech, to permit communication over great distances, and to permit communication with machines using speech?
6. How is speech acquired as a first language and as a subsequent language?
7. How is sound associated with meaning?
8. How did language and speech arise or evolve in our species? Why is the vocal apparatus different as a function of the age and sex of the speaker? What is the relation, if any, between human speech and non-human communication?

My 8 questions ( 1.2.3 for the questions in our textbook, (1)(2)(3) for my questions)
(1) What are phonological universals?
1. 2. 3. 6. 7. are relevant to my question (1) because
Asking questions 1, 2, 3, 6, 7 in our textbook enables one to draw inference of language universal.
8. is relevant to my question (1) because
Language universal could be a sub-question of 8.

(2) What is language disorder?
4 is relevant to (2) because
One should ask (2) first before move to 4.

(3) How phonology involve into its current state?
All questions are relevant to (3) because
The changes of phonology have much to do with our pursuing answers of all questions in our textbook.

(this one is relevant to (8))

(4) What new methods can we use in phonology?
This one is relevant to all questions because
New method could be applied to get those answers.
(It is also relevant to (7))

(5) What factors contribute to phonological changes?
This one is relevant to 1. 2. 3.

(6) Where do phonologists acquire information?

(7) What other field can phonology apply?
This one is relevant to all questions because
New method could be applied to get those answers.
(It is also relevant to (4))

(8) What’s the way phonology going to be changed?
All questions are relevant to (8) because
The changes of phonology have much to do with our pursuing answers of all questions in our textbook.
(this one is relevant to (3))


Questions (1) (2) (3) (5) (8) overlap those in our textbook. (4) and (7) are relevant to questions in our textbook, but they are not overlapping.

2008年5月8日 星期四

知人知面不知心?專家指未必

中新網報導:人類臉部會露出一種「微表情」(「微小」的「微」),它通常露出只有二十分之一秒的時間,一般人肉眼根本察覺不到,當事人自己也沒感覺。這種「微表情」就像腦波一樣,在不同的心情下呈現出不同的表情,加拿大最新科學研究發現,在各種「微表情」中,「說謊」的表情特別容易被辨識出來。
由於肉眼無法在瞬間抓到別人臉上的細微表情,也無法辨識出微表情的意義,所以經常愛撒謊的人暫時不用擔心被人抓包,不過一旦發展出「微表情電腦辨識儀器」,「知人知面不知心」這句話恐怕就得改寫了。

reference:http://tw.news.yahoo.com/article/url/d/a/080507/1/yp54.html

2008年5月1日 星期四

Area Functions and Articulatory Modeling&Phonetically based phonology

On Phonetically based phonology
http://www.vanoostendorp.nl/pdf/phoneticallydriven.pdf
相關書籍
http://www.linguistics.ucla.edu/people/hayes/PBP/index.htm

Area Functions and Articulatory Modeling

Website :

http://www.fon.hum.uva.nl/praat/manual/Fant__1960_.html
Vocal Tract Length during Speech Production
http://www.caip.rutgers.edu/~sdusan/d_euro2007a.pdf
The study of the French Vowels constriction of the vocal tract using an acoustic-to-articulatory inversion method
http://www.loria.fr/~slim/publications/ICPhS-inv.pdf

Key words:

1. AM: articularoty modeling
2. VT: vocal tract
3. Formant: a frequency range where vowel sounds are at their most distinctive and characteristic pitch
4. Amplitude: largeness in size, volume, or extent
5. Midsagittal: of imaginary midline plane through body: relating to or situated along an imaginary plane that passes through the midline of the body or an organ
6. Semipolar grid
7. Semipolar
8. Grid: reference lines on map: a network of evenly spaced horizontal and vertical lines on a map, used as a basis for finding specific points
9. Spectrogram:record of spectrum: a photograph or representation of a spectrum