000 03708nam a2200433 4500
005 20260818113227.0
008 180698 th eng
035 _a.b11489388
099 9 _aAIT Thesis no.TC-95-5
100 1 _aRamalingam, H.
245 1 0 _aExtraction of tones of speech :
_ban application to the Thai language
260 _aBangkok :
_bAsian Institute of Technology,
_c1995
300 _a85 leaves :
_bill.
490 1 _aThesis ;
_vno. TC-95-5
502 _aThesis (M.Eng.) - Asian Institute of Technology, 1995
500 _aA thesis submitted in partial fulfilment of the requirement for the degree of Master of Engineering.
520 _aThe pitch vanat10n that convey lexical infonnation about the meaning of a word is commonly refeITed to as a tone. Automatic speech recognition of tonal languages requires the recognition of tones associated with the syllables in addition to the phonemes for proper identification of the syllable. Here, in this thesis work, a general model for the automatic recognition of tones of syllables of tonal languages is developed. The pitch contour of the syllable is initially estimated using Subharmonic Summation as the Pitch Determination Algorithm and given as input to vector quantizer for correct recognition of tones associated with the syllable. The input vector is time aligned and pitch n01malised to remove inter and intraspeaker variations before being applied to the vector quantizer. A codebook is implemented containing reference vectors corresponding to each of the tones of the tonal language. A distortion measure is computed between the test vector and each of the reference vectors. The reference vector c01Tesponding to the least distortion is identified as the tone. The perfo1mance evaluation is done in MATLAB environment. Thai language is chosen as the tonal language for the pe1formance evaluation. Four isolated syllables, uttered by four speakers for all the five tones, are used for simulation. For noise-free speech the system gave 100% correct recognition of tones for the four speakers. Recognition rates of 98, 97, 97, 98, 93 and 93 % were obtained for signal-to-noise-ratios of 40 dB, 30 dB, 20 dB, 10 dB, 5 dB and 2 dB respectively for the worst case speaker.
650 1 0 _aAutomatic speech recognition
700 1 _aMakeHiinen, Kimmo,
_eChairperson
700 1 _aAhmed, Kazi Mohiuddin,
_eExamination Committee
700 0 _aChindakorn Tuchinda,
_eExamination committee
710 2 _aNORAD (Norwegian Agency for International Development),
_eScholarship donor
810 2 _aAsian Institute of Technology.
_tThesis ;
_vno. TC-95-5
856 _3Full-Text
_uhttp://203.159.5.9/ait-thesis/detail.php?q=B15296
907 _a.b11489388
_bmnait
_cs
902 _a240717
998 _b0
_c960603
_dm
_ea
_fs
_g0
945 _lmnait
945 _lmnait
945 _lmnarc
945 _lmnarc
942 _c20
942 _c40
909 _aBarcode : 30050120813877
_bCREATED : 1994-12-18
_cRECORD # : i11753833
_dLPATRON : 1019606
_eLCHKIN : 2007-08-28
_f# RENEWALS : 0
_g# OVERDUE : 0
_hIUSE3 : 0
_iTOT CHKOUT : 3
_jTOT RENEW : 5
909 _aBarcode : 30050120813885
_bCREATED : 1994-12-18
_cRECORD # : i11753845
_dLPATRON : 1012987
_eLCHKIN : 2000-03-07
_f# RENEWALS : 0
_g# OVERDUE : 0
_hIUSE3 : 0
_iTOT CHKOUT : 6
_jTOT RENEW : 4
909 _aBarcode : 30050160046081
_bCREATED : 2016-02-03
_cRECORD # : i12887523
_dLPATRON : 0
_eLCHKIN : -
_f# RENEWALS : 0
_g# OVERDUE : 0
_hIUSE3 : 0
_iTOT CHKOUT : 0
_jTOT RENEW : 0
909 _aBarcode : 30050211018741
_bCREATED : 2022-03-17
_cRECORD # : i13373742
_dLPATRON : 0
_eLCHKIN : -
_f# RENEWALS : 0
_g# OVERDUE : 0
_hIUSE3 : 0
_iTOT CHKOUT : 0
_jTOT RENEW : 0
999 _c42685
_d42685