Đào tạo

TTLV: Research on phonetic and lexical characteristics of some dialects in Hanoi, application in speech recognition

Friday - June 26, 2015 05:23

MASTER'S THESIS INFORMATION

1. Student's full name:Tran Quang Thanh          

2. Gender: Male

3. Date of birth: February 21, 1980

4. Place of birth: Hanoi

5. Decision on recognition of students No. 1883/2010/QD-XHNV-SDH dated October 21, 2010 of the President of the University of Social Sciences and Humanities, Vietnam National University, Hanoi

6. Changes in training process: None

7. Thesis topic name:Research on phonetic and lexical characteristics of some dialects in Hanoi, application in speech recognition

8. Major: Linguistics Code: 60.22.02.40

9. Scientific instructor:Assoc.Prof.Dr. Vu Kim Bang,working at the Institute of Linguistics.

10. Summary of thesis results:

This is the first work to study phonetic-lexical features to distinguish Hanoi dialects applied in speech recognition. The thesis uses phonetic analysis, auditory perception and specialized software to accurately and scientifically identify the dialect features of Hanoi. The thesis has achieved the following results:

The thesis has studied the characteristics of segmental phonemes, characteristics of the tone system (supra-segmental phonemes), some lexical characteristics of the dialects in the districts of Thach That, Soc Son, Dan Phuong, Phuc Tho, Hoai Duc - Hanoi. Through the survey in the above 5 districts, we found the differences expressed in the entire syllable components. Therefore, the zoning of dialects in the identification of the origin of the voice needs to have both the vowel and consonant systems and tones. In addition, vocabulary reference is also a necessary condition that can supplement the tracing.

The project has initially tested the construction of a phonetic feature system to help quickly identify dialects in Hanoi. The consonant system of the 4 districts of Soc Son, Dan Phuong, Phuc Tho, and Thach That is basically the same as the Northern dialect, including: 19 consonants, the missing consonants are:/ʈ/, /ş/,/ʐ/ is gradually moving to completely merge with the consonants /z/,/s/,/c/. In which, the consonants /z1/,/s1/,/c1/ are similar to /z/,/s/,/c/, in terms of pronunciation, the only difference is that they have greater intensity to emphasize, creating a small difference. In particular, the consonant system of Hoai Duc district, in addition to the above pronunciation, has the conversion of tr-t, s, x-th (that change is not in all sounds). All districts have the conversion of "o"-oo, ô", "e"-"ee" but the difference in the vowel system of the districts is not large, only Hoai Duc district has the vowel "o" converted to "ơ" in most syllables. This is also a distinct feature compared to other dialects. The tone system has the most distinct features of the above dialects, we model the contours of the tones to help quickly trace.

In the identification of some local words: Like most other rural areas in the Northern Delta, most districts still preserve ancient words such as: gioi- troi, dua-nhua, nay-nay, cu-con trai... However, these words only appear sporadically, unsystematically and mainly among the elderly. Therefore, in the near future, they will no longer have any traceability value. We only note that Cat Que commune, Hoai Duc district has a rather different word system. It is the differences in terms of address or local words that make the language in Cat Que richer and more diverse than some localities in Hoai Duc district.

With the goal of quickly identifying Hanoi dialects, we have built clear and concise distinguishing criteria to help accurately identify each dialect through the characteristics of vowels, consonants, tones, and vocabulary. The thesis has achieved the goal of building a diagram of the dialect's characteristic system along with a spectrum analysis template to help quickly trace and quickly identify dialects in Hanoi.

Regarding voice assessment activities, the thesis has the following outstanding contributions:

+ The thesis initially delves into the analysis of dialects. Currently, the determination of dialect regions in judicial appraisal activities has only stopped at the dialects of the three regions of North, Central, South and some provinces and cities; there has not been any tracing to each dialect region of districts and communes.

+ The initial spectral analysis method has entered the structure of phonemes in syllables. Currently, the main method of speech recognition is to measure and analyze the 3 average spectral formants in the entire syllable.

+ Combining auditory perception and analysis of pitch and formant parameters for accurate and scientific data is a new direction in phonetic analysis for voice identification, overcoming the current limitation that auditory perception analysis is not associated with spectrogram analysis, reducing objectivity in identification.

11. Practical application:

The thesis has application value in voice identification, serving the scientific-criminal work. In criminal investigation activities, collecting the voice of the perpetrator or related people to serve the identification activities, on that basis determining the perpetrator, criminal responsibility is important and necessary. Voice identification identification activities depend on many factors such as dialect, phonetics, tone, syllables, regional dialect... of the comparison sample as well as the collected voice sound. To improve the effectiveness of voice identification work, it is necessary to study and clarify these contents.

12. Further research directions:

Research on the phonetic and lexical characteristics of some dialects in the Red River Delta (compared with Hanoi dialect) to serve the work of voice identification in the field of science and forensics.

13. Published works related to the thesis:

1). Tran Quang Thanh,Survey of some phonetic and lexical characteristics of the local dialect of Canh Nau commune, Thach That district, Hanoi, Proceedings of the 3rd International Conference: Research and Teaching of Chinese-Vietnamese Language and Culture, from page 82 to page 93.

2). Tran Quang Thanh (co-author),Phonetic and lexical characteristics of the local dialect of Cat Que commune, Hoai Duc district, Hanoi, National Linguistics Yearbook 2013, from page 183 to page 192.

 

INFORMATION ON MASTER'S THESIS

1. Full name: Tran Quang Thanh 2. Sex: Male

3. Date of birth: February 21, 1980 4. Place of birth: Hanoi

5. Admission decision number: 1883/2010/QD-XHNV-SĐH Dated: October 21, 2010

6. Changes in academic process: No

7. Official thesis title:Research on phonetic- vocabulary characteristics of some dialects in Hanoi, application in the assessment of speech recognition.

8. Major: Linguistics 9. Code: 60.22.02.40

10. Supervisors: Associate Professor Doctor Vu Kim Bang, Institute of Linguistics.

11. Summary of the findings of the thesis:

This thesis is the first research about phonetic - vocabulary characteristics to find out different characteristics of Hanoi dialects, applied in expertising to recognize speech. The thesis uses some methodology such as: phonetic analysis, auditory perception, dedicated software to identify features Hanoi dialects correctly.

The thesis research on characteristics of Vietnamese's consonants system, vowel system, 6 tones, vocabulary of dialects in Thach That, Soc Son, Dan Phuong, Phuc Tho, Hoai Duc- Ha Noi. The difference is based on all of syllabes.

The theme was initially tested building featured phonetic's system to rapid assessment dialects in Hanoi. Consonants system in Soc Son, Dan Phuong, Phuc Tho, Thach That is almost the same as the North's dialect: 19 consonants, didn't have/ʈ/, /ş/,/ʐ/. There is a change in Hoai Duc district: tr-t, s,x- th (not in all of syllables). There are some changes in 5 districts: “o”-oo, Ô”, “e”-“ee” and in Hoai Duc has “o” -> “ê”. Tone system touches most obviously localized on dialects (we described in graph).

In the assessment of some local words: Cat Que Commune, Hoai Duc district has the distinction of terms of local dialect ors.

Thesis has built a criteria clear, concise, help to identify each dialect correctly through characteristics of vowels, consonants, tones, vocabulary. Thesis has achieved the goal of building the system diagram characteristic of dialects with spectrum analysis templates to help trace the fast, rapid assessment in Hanoi dialect.

For voice assessment activities:

Thesis initially goes into dialect analysis. Currently, the identification of regional dialects in judicial assessment activities just at three dialects: North dialect, Central dialect, South dialect and some provinces; no trace of the dialect in districts, communes.

Spectral analysis method initially went into the phonological structure of syllables. Currently, methods for detemining speech recognition is mainly applied analytical measurements of 3 formant spectrum in average in both syllables.

The combination of auditory perception and analyzing the parameters pitch, formant for data accurate, science is a new direction in the work of analytical expertise phonetic voice service, ensuring objectivity.

12. Practical applicability, if any:

Thesis valuable in the application of voice assessment, catering for the science- crime.

13. Further research directions, if any:

Research phonetic- vocabulary characteristics of some dialects in the Red river delta (compares with dialects in Hanoi) to serve the voice expertise in the field of science and crime.

14. Thesis-related publications:

1). Tran Quang Thanh, Survey some phonetic - vocabulary characteristics in dialect of Canh Nau commune, Thach That district, Hanoi, Proceedings of the International Workshop: Research and teaching language, culture China- Vietnam III, from page 82 to page 93.

2). Tran Quang Thanh- Nguyen Thi Phuong Thuy, Characteristics of phonetic - vocabulary of dialect in Cat Que commune, Hoai Duc district, Hanoi, National Linguistics Proceeding 2013, from page 183 to page 192.

Total score of the article is: 0 out of 0 reviews

Click to rate this article

Newer news

Older news

[LANG_MOBILE]
You have not used the Site,Click here to stay logged inWaiting time: 60 second