Title: ????? ???? Teaching Language Prosody - The role of speech recognition -
1????? ????Teaching Language Prosody- The role
of speech recognition -
- ????? ???? ???
- KT ??????
- ??????? ???
- 2008. 2. 29(?) ??2?
2??? ??
- ??? ??3?
- ??? ?? 9?
- ??? ?? 14?
- ???? 16?
- ???? 35?
- ???? ??50?
- ????? ??? ?? 66?
-
31. ??? ??
41. ??? ??
51. ??? ??
- ???
- ???(?,??) ????(??)????
61. ??? ??
- ?? ?? ???? ?????
- ?? ???, ?? ??, ?? ??? ???? ?? ??? ? ? ????
- ??, ??, ??? ??? ?? ??? ??? ??? ?? ?? ??
71. ??? ??
- ??? ??? ??? ??? ?? ? ? ????
- ????????? ?? ??? ?? ??? ??? ?
8??? ???/?????
92. ??? ??
102. ??? ??
- ??(stress)? ??? ???? ?????? ???
- ?? ???? ??
- ?? ??? ??
- ?? ??? ??
- ?) banana b?n?n? A What did you have
for lunch? B A banana. A A
banana?
112. ??? ??
- ??? ???? ????? ????
- ?, ????(?????)? ?? ????
- ???? ??, ?? ??? ????? ??? ????? ?????? ???
- ?) ??? ???? ???? ????? ????
- ?) ??? ???? ????? ????? ????
122. ??? ??
132. ??? ??
- ?? ? ??? ????
-
- ?? 1
- ?? 2
- ?? 3
143. ??? ??
153. ??? ??
- ????? ?? ? ????? ???? ?? ? ???? ??,
?? ???? ?? ? ???? -
164. ?? ??
- Cloning Prosody
- ???? 10-0701338
- ??? ??? ????? ??? ?????
17Acquiring Prosody in Language Learning
4.1 ?? ?? ??
- One of the critical tasks in language learning
- Prosody as non-segmental features of speech1.
phrase breaks2. intonation (F0) contour3.
segmental durations4. intensity contour
18Previous Approaches
4.1 ?? ?? ??
- Explicit teaching of prosodic features such as
the intonation contours, segmental durations,
etc. - Audio aidListen and repeat!
- Visual aid in computer softwareDr.Speaking F0
contour comparison between native speaker and
non-native speaker
19A New Approach
4.1 ?? ?? ??
- A new kind of audio aidin the form of a
non-native speakers utterance with the prosodic
features of a native speakers utterance - How this works1. Software presents a native
speakers utterance2. A non-native speaker
repeats the utterance3. Software records the
non-native speakers utterance4. Software
imposes the native speakers prosody onto the
non-native speakers utterance5. Software
presents the processed non-native utterance
20Flowchart
21Technical Details
4.1 ?? ?? ??
- Manipulation of1. segmental durations, including
phrase breaks 2. F0 contours 3. intensity
contours - For 1 and 2
- PSOLA (Pitch Synchronous OverLap and Add),
developed by Moulines Charpentier,
1990implemented in Praat - For 3Intensity swap in Praat
22Moulines Charpentier, 1990 1
4.1 ?? ?? ??
Technical Details
original waveform
windowed waveform
1 2 3 4 5 6 7 8 9 10
11 12 13 14 15 16 17 18 19
shortened waveform
waveform with lower F0
1 4 7 10 13 16 19
1 3 5 7 9
11 13 15 17 19
234.1 ?? ?? ??
Technical Details Segmental durations (1)
- Segment alignment PSOLA processing of
durations Alignment can be manual or
automatic (with the help of speech recognition)
k
eI
m
i
n
came in
native
stretch
shrink
k
eI
i
n
m
non-native
244.1 ?? ?? ??
Segmental durations (1) F0 contour (2)
- PSOLA processing of F0 on duration-treated
utterance
native F0
k
eI
m
i
n
native
k
eI
m
i
n
non-native
non-native F0
254.1 ?? ?? ??
Segmental durations (1) F0 contour (2)
Intensity contour (3)
- Mathematically neutralize non-native speakers
intensity contour and transfer native speakers
intensity contour in Praat Holger Miterer
(p.c.)
native intensity
k
eI
m
i
n
native
k
eI
m
i
n
non-native
non-native intensity
264.1 ?? ?? ??
Segmental durations (1) Intensity contour (3)
- Segment alignment PSOLA processing of duations
followed by intensity contour transfer
native intensity
k
eI
m
i
n
native
stretch
shrink
k
eI
i
n
m
non-native
non-native intensity
274.1 ?? ?? ??
F0 contour (2) Intensity contour (3)
- Reverse segment alignment PSOLA processing of
F0 followed by intensity contour transfer
native F0
native intensity
k
eI
m
i
n
native
stretch
shrink
k
eI
i
n
m
non-native
non-native F0
non-native intensity
284.1 ?? ?? ??
- Weakness1. Voiceless segments can be made
voiced in the windowing process
(pitch-synchronous technique)2. Excessive
handling results in unnatural synthesis (One
solution pitch rescaling 3) - Segment alignmentshould be fine-tuned according
to the voiced/voicless status of the
(sub-)segments for better results
294.2 ?? ?? ??
native utterance
non-native utterance
synthetic non-native(durationsF0intensity)
synthetic non-native(durationsintensity)
synthetic non-native(F0intensity)
30Comparison before synthesis duration, F0
intensity
4.2 ?? ?? ??
(See blue yellow lines)
native utterance
non-native utterance
31Comparison after synthesis duration, F0
intensity
4.2 ?? ?? ??
(See blue yellow lines)
native utterance
synthetic non-native
32Comparison after synthesis duration intensity
4.2 ?? ?? ??
(See blue yellow lines)
native utterance
synthetic non-native
33Comparison after synthesis F0 intensity
4.2 ?? ?? ??
(See blue yellow lines)
native utterance
synthetic non-native
344.3 ?? ?? ?? ??
- The technique could be used (1) In second
language education to facilitate/motivate
acquisition of the target language prosody
to emphasize the importance of prosody in
achieving native speaker fluency(2) For
patients with vocal disorders to help
achieve the prosody of a normal voice - Auto-segmentation via ASR (Automatic Speech
Recognition) or DTW (Dynamic Time Warping) 3
can be employed to automate the segment
alignment.
355. ?? ??
365.1 ??? ?? ????
- ??? 90? ??? ?????(1) 89?? ???(2) 90?? ???(3)
90? ???? ???? ?? ?? 90?? ??? - ?? ??? ? ??????? ???? ????
375.2 ??? ??? ??
- MOMEL algorithm(Hirst Espesser, 1991)? ???
???/??? ?? ?? - Praat implementation by Cyril Auranhttp//stl.rec
herche.univ-lille3.fr/sitespersonnels/auran/englis
h/index.html - ?? PSOLA algorithm(Moulines Charpentier, 1990)?
????, ???/???? ????? ?? ??????? ???? ???? ???? - MOMEL algorithm? ???? ?? ?? Manipulation object?
Stylize pitch (2 st) ??? ??? ?? ??
385.2 ??? ??? ??
395.2 ??? ??? ??
405.2 ??? ??? ??
415.2 ??? ??? ??
425.2 ??? ??? ??
435.2 ??? ??? ??
445.2 ??? ??? ??
455.2 ??? ??? ??
465.2 ??? ??? ??
475.2 ??? ??? ??
485.3 ??? ??? ??
PSOLA algorithm(Moulines Charpentier, 1990)?
??, ???? ??????? ???? ???
495.4 ??? ?? ??
506. ???? ??
516.1 ??? ???? ??
- ??? ???... Listen and repeat!
-
526.1 ??? ???? ??
- ?? ???... Repeat after me!
-
536.1 ??? ???? ??
- ?? ???/??? ???? ???? ?? ???? ??? ???.
- ?, ??? ???? ?? ???? ??.
- ? ? ???? ??? ??? ?? ???? ???? ????? ??? ??/???
???.
546.2 ??? ????
- ??? ?? ???? ????? ????? ??? ??? ??? ?? ???? ???
?????, - 1. ???? ?????? ??? ???? ?? ????.2. ???? ??????
??? ???? ?? ????.3. ???? ?????? ??? ???? ??
????.?? ??? ?? 1,2? ??? 1,3? ??? 2,3? ??? ???
1,2,3? ?? ??? ????.
556.2 ??? ????
- ??? ?????,??? ??? ?? ?? ??, ? ?? ??? ??? ??? ??
?? ????. ?,1. ???? ?????? ??? ?????,2. ????
?????? ??? ?????, 3. ???? ?????? ??? ????? ??
??? ?? ??? ? ??? ???? ???? ?? ??? ???? ?? ??? ?
??? ????.
566.2 ??? ????
- ??? ???? ?????? ?? ??? ????? ???? ?????? ?????
?? ?? ????? ???. - ?? ???? (?????)1. ???? ???? ???? ??2. ???? ??
?? ?? ??3. ?????? ???? ??? ????? ????4. ????
????? ??? ???? ??
576.2 ??? ????
?????? ??
586.3 ????? ??? ????
- ??? ????? ???? ???? ???? ?? ??? ?? ? ??? ??
- ???? ???? ???? ????? ??? ????
- ???? ??? ???? ??? ?? ???? ???? ? ??? ????
- ???? ??? ??? ?? ?? ??? ?? ??? ???? ???? ???? ???
?? ??
596.4 ??/??? ?? ??
- ??/??? ??
- ?? ??/??? ?? ???
- ??? ?? ??? ?? ???
- ?? ??
- ?? ??? ?? ???? ???? ???
- ??/??/??
- ?? ??? ??? ????? ???
- ???? ?? ??? ???/???? ?? ???? ???
- ?
- http//www.tellmemore-online.com/
- http//www.uiowa.edu/acadtech/phonetics
606.5 ?????? ?
616.5 ?????? ?
626.5 ?????? ?
636.5 ?????? ?
64???? ?? ??
65???? ?? ??
667. ????? ??? ??
677.1 ????? ??
- ??? (superfluous segments) ?? ??
- ?? ??? ?? (??? ??) ??
- ??? ????? ???? ??
- ???? ???? ?? ??
- send paper ?? d? ???? ??
- ??? (deficient segments) ?? ??
- ?? ??? ??? ???? ??? ??
- Ive told ?? v? ?? ??
687.2 ????? ??
- ???/??/??? ??? ??
- ??? ????? ??? ??? ??
- ?????(sub-segments) ??? ??
- ??? ????? ?? ??(?)
- ?) ???? ????? ??? ??? ??? ??
- ?) ???? ????????(VOT) ?? ??
- ?) ???? ???? ??
- ?) ??? ?? ? ?????? ?? ?
69References
1 E. Moulines and F. Charpentier (1990) Pitch
synchronous waveform processing techniques for
text-to-speech synthesis using diphones Speech
Communication 9, 453-467. 2 P. Boersma (2005)
Praat, a system for doing phonetics by
computer, Glot International, Vol.5(9/10), pp.
341-345.3 S. Yi (2007) Perception of English
prosody by Americans and Koreans and its
pedagogical implications, Ph.D.
Dissertation, Busan Pusan National
University. 4 K. Yoon (2006) Imposing native
speakers prosody on non-native speakers
utterances, Proceedings of the 9th Western
Pacific Acoustics Conference (WESPAC9), Seoul,
South Korea.