PhD thesis - School of Informatics - University of Edinburgh

More documents

Recommendations

Info

Chapter 6. Other Potential Applications 155 this effect both for German speakers where the similar English sound is℄and the new English sound is℄(Bohn and Flege, 1992) as well as for French speakers, the similar English sound beingÙ℄and the new English soundÝ℄(Flege, 1987). In Flege et al. (1997), the work on German learners of English (Bohn and Flege, 1990, 1992) is extended to native speakers of Spanish, Mandarin and Korean, three typologically diverse languages. Although the conclusions are similar, the paper stresses that further research is required to determine the degree of perceived relatedness between vowels found in English and in each of the native languages examined in the experiment. Although similar in some respects, SLA research does not deal with the exact same issues involved in the production and perception of embedded foreign inclusions. It addresses the way in which speakers of a language L1 approach a language L2 with the intention of mastering it fluently, while sometimes living in a country where L2 is spoken. The research presented in this thesis, on the other hand, focuses on the occurrence of lexical items from a foreign language embedded within utterances in a native language and on how such foreign inclusions are dealt with by native speakers in their own language surrounding. In fact, the finding of SLA studies that similar L2 sounds are pronounced with properties of L1 and are therefore adapted is likely to be even more severe for embedded foreign inclusions as a result of co-articulation and a series of other factors determined in Section 6.1.1.1. 6.1.1.3 Implications for the Perception of Polyglot TTS In the context of polyglot TTS synthesis, first and foremost the type and degree of adjusting foreign inclusions to the native language must be addressed. As the studies described in Section 6.1.1.1 have illustrated, pronunciation of foreign inclusions by native speakers varies depending on a series of factors, including age and educational level of the speaker. Other aspects that play a role in the production of foreign inclusions involve the orthography, the context of the particular foreign inclusion, the effort of production as well as the expectations of speakers about their listeners. As pointed out earlier, the degree of tangibility of all of these issues varies which can present a serious difficulty when it comes to modelling or controlling them. Moreover, the assumption that a speaker is capable of producing a near authentic pronunciation of a particular foreign inclusion does not necessarily imply that the listener will actually understand, let alone like and accept what they hear. Wijngaarden
Chapter 6. Other Potential Applications 156 and Steeneken (2000) show with a combination of speech reception threshold (SRT) 3 tests and letter guessing procedure (LGP) 4 experiments that non-authentic, i.e accented pronunciation of foreign words increases intelligibility to inexperienced learners. Ex- perienced L2 learners, on the other hand, perceive non-accented L2 speech as more intelligible. Their experiments involves native Dutch speakers listening to German and English speech produced by German and English native speakers as well as Dutch native speakers. If the same holds true for the perception of polyglot speech, where the foreign items are spoken in the context of the listeners own language, is still an open question. Even if a TTS synthesis system is capable of producing authentic pronunciations of embedded inclusions or can at least approximate them to some extent in the foreign language, it is unlikely to be the accepted or preferred rendering compared to one that adjusts at least some foreign speech sounds to the acoustic characteristics of the listener’s own language. As Lindström and Eklund (2000) point out, too much ap- proximation to the foreign language may not result in the desired effect for listeners. They may consider it exaggerated or inadequate. The other extreme, producing the pronunciation of embedded foreign words merely by means of letter-to-sound rules of the base language of the text, will result in bad mispronunciations or overly accented speech which listeners are likely to deem uneducated or may not even understand. Besides speaker- and word-related factors such as age, experience of the foreign language, awareness of the task or frequency of the foreign word, which can heavily influence the results of a mixed-lingual speech perception experiment, a perception study of mixed-lingual TTS synthesis output also depends on the actual synthesis approach of foreign inclusions. Although the chosen approach may well be modelled on the behaviour of subjects in production experiments, the perception study subjects 3 An SRT test involves adding masking noise to test sentences. If subjects listening to the test sentence repeat all its words correctly, the noise is increased, otherwise it is decreased. The SRT value is the average speech-to-noise ratio over a set of sentences and provides a reliable measure for sentence intelligibility in noise. 4 LGP was devised by Shannon and Weaver (1949) as a way of estimating lower and upper bounds of linguistic information content of a language. Subjects are asked to guess a string of text one letter at a time without receiving any prior information. After each guess, the correct letter is either revealed (single-LGP) or kept undisclosed until the guess is correct (multiple-LGP). Subjects can make use of the letters guessed up to the current point for predicting the next letter. The more letters a subject guesses randomly, the less redundant the language is to them and the more linguistically skilled they are in that language (van Rooij and Plomp, 1991). LGP scores are expressed in terms of linguistic entropy L (in bits): L = −log2(c) where c is the fraction of correctly guessed letters.
Page 1 and 2:
Automatic Detection of English Incl
Page 3 and 4:
these parsers with the annotation-f
Page 5 and 6:
Declaration I declare that this the
Page 7 and 8:
3.3.5 Post-processing Module . . .
Page 9 and 10:
A.2.2 Kappa Coefficient . . . . . .
Page 11 and 12:
5.6 Average relative token frequenc
Page 13 and 14:
3.16 Most frequent English inclusio
Page 15 and 16:
Chapter 1. Introduction 2 siderable
Page 17 and 18:
Chapter 1. Introduction 4 Chapter 3
Page 19 and 20:
Chapter 1. Introduction 6 1.1 Relat
Page 21 and 22:
Chapter 2. Background and Theory 8
Page 23 and 24:
Page 25 and 26:
Page 27 and 28:
Page 29 and 30:
Page 31 and 32:
Page 33 and 34:
Page 35 and 36:
Page 37 and 38:
Page 39 and 40:
Page 41 and 42:
Page 43 and 44:
Page 45 and 46:
Page 47 and 48:
Page 49 and 50:
Page 51 and 52:
Page 53 and 54:
Page 55 and 56:
Page 57 and 58:
Page 59 and 60:
Chapter 3 Tracking English Inclusio
Page 61 and 62:
Chapter 3. Tracking English Inclusi
Page 63 and 64:
Page 65 and 66:
Page 67 and 68:
Page 69 and 70:
Page 71 and 72:
Page 73 and 74:
Page 75 and 76:
Page 77 and 78:
Page 79 and 80:
Page 81 and 82:
Page 83 and 84:
Page 85 and 86:
Page 87 and 88:
Page 89 and 90:
Page 91 and 92:
Page 93 and 94:
Page 95 and 96:
Page 97 and 98:
Page 99 and 100:
Page 101 and 102:
Page 103 and 104:
Page 105 and 106:
Page 107 and 108:
Page 109 and 110:
Page 111 and 112:
Page 113 and 114:
Chapter 4 System Extension to a New
Page 115 and 116:
Chapter 4. System Extension to a Ne
Page 117 and 118: Chapter 4. System Extension to a Ne
Page 129 and 130: Chapter 5 Parsing English Inclusion
Page 131 and 132: Chapter 5. Parsing English Inclusio
Page 159 and 160: Chapter 6 Other Potential Applicati
Page 161 and 162: Chapter 6. Other Potential Applicat
Page 167: Chapter 6. Other Potential Applicat
Page 187 and 188: Chapter 7 Conclusions and Future Wo
Page 189 and 190: Chapter 7. Conclusions and Future W
Page 191 and 192: Appendix A. Evaluation Metrics and
Page 199 and 200: Appendix B. Guidelines for Annotati
Page 205 and 206: Appendix C TIGER Tags and Labels C.
Page 207 and 208: Appendix C. TIGER Tags and Labels 1
Page 209 and 210: Appendix C. TIGER Tags and Labels 1
Page 211 and 212: Bibliography 198 Andersen, G. (2005
Page 213 and 214: Bibliography 200 Bresnan, J. (2001)
Page 215 and 216: Bibliography 202 Damashek, M. (1995
Page 217 and 218: Bibliography 204 Finkel, J., Dingar
Page 219 and 220:
Bibliography 206 Hachey, B., Alex,
Page 221 and 222:
Bibliography 208 Kirkness, A. (1984
Page 223 and 224:
Bibliography 210 and Technology (In
Page 225 and 226:
Bibliography 212 Poplack, S. (1988)
Page 227 and 228:
Bibliography 214 Sokol, D. K. (2000
Page 229:
Bibliography 216 Yang, W. (1990). A
show all

PhD thesis - School of Informatics - University of Edinburgh

Create successful ePaper yourself

Delete template?

Save as template?