創新科技產品往往被視為企業成長的重要關鍵，有許多創新 科技產品成功的案例讓企業獲利大幅成長，但是並非所有創新科技產 品都被市場接受，也有許多創新科技產品在市場上曇花一現，或是經 歷一小段時間的成長之後又被市場淘汰。 許多不同的因素都會影響創新科技產品成功與失敗，在複雜的因 素中，創新科技產品公司的初始經營模式是其中一項重要關鍵。因為 初始的經營模式代表著如何看待自身所處的產業環境與產業鏈的位 置，由此延伸出公司本身應該重視與培養的核心能耐就會不同，獲利 模式也會不同。 如果只是針對現有科技改良的創新，開發新產品時可以遵循先前 產品的基礎來預估新產品的可能市場，但是如果是不連續性創新或是 破壞式創新要進入市場時就沒有前例可循，所以要選擇目標市場時會 非常困難。 因此不連續性創新科技要進入市場之前要先選定經營模式，然後 才能進一步開發新產品，接著遵循行銷學中的新產品開發步驟，STP 與4P將產品推進市場中。 紙喇叭此項不連續創新科技要進入市場時，既可以當成零組件也 可以做為成品來銷售，而企業在經營時也要選擇是否要成立品牌還是 從事代工。 在比較各類經營模式之後，創新科技，尤其是不連續創新或是破 壞性創新的產品，最佳的經營模式是成立品牌，自行銷售成品到市場 上是最合理的選擇。 / All companies embrace innovation as the key for growth. But not all innovation brings fortune to company who owns it. Initial business model is one of the most critical factors that will affect the success of innovating product. If the innovating product follow its predecessor with continuous innovation, it is easy to predict new product market size and revenue. But it is very difficult for a discontinuous innovating product to predict market size without any predecessor as base. Chose a right business model is the basis before discontinuous innovating product start marketing. There are 4 types of business model analyzed in this article; OBM or ODM, selling parts or final product. In the case of flexible speaker, making consumer production with own branding would be the right choice.

<p>Humans have the ability to recognize other humans by voice alone. This is important both socially and for the robustness of speech perception. This Thesis contains a set of eight studies that investigates how different factors impact on speaker recognition and how these factors can help explain how listeners perceive and evaluate speaker identity. The first study is a review paper overviewing emotion decoding and encoding research. The second study compares the relative importance of the emotional tone in the voice and the emotional content of the message. A mismatch between these was shown to impact upon decoding speed. The third study investigates the factor dialect in speaker recognition and shows, using a bidialectal speaker as the target voice to control all other variables, that the dominance of dialect cannot be overcome. The fourth paper investigates if imitated stage dialects are as perceptually dominant as natural dialects. It was found that a professional actor could disguise his voice successfully by imitating a dialect, yet that a listener's proficiency in a language or accent can reduce susceptibility to a dialect imitation. Papers five to seven focus on automatic techniques for speaker separation. Paper five shows that a method developed for Australian English diphthongs produced comparable results with a Swedish glide + vowel transition. The sixth and seventh papers investigate a speaker separation technique developed for American English. It was found that the technique could be used to separate Swedish speakers and that it is robust against professional imitations. Paper eight investigates how age and hearing impact upon earwitness reliability. This study shows that a senior citizen with corrected hearing can be as reliable an earwitness as a younger adult with no hearing problem, but suggests that a witness' general cognitive skill deterioration needs to be considered when assessing a senior citizen's earwitness evidence. On the basis of the studies a model of speaker recognition is presented, based on the face recognition model by V. Bruce and Young (1986; British Journal of Psychology, 77, pp. 305 - 327) and the voice recognition model by Belin, Fecteau and Bédard (2004; TRENDS in Cognitive Science, 8, pp. 129 - 134). The merged and modified model handles both familiar and unfamiliar voices. The findings presented in this Thesis, in particular the findings of the individual papers in Part II, have implications for criminal cases in which speaker recognition forms a part. The findings feed directly into the growing body of forensic phonetic and forensic linguistic research.</p>

Humans have the ability to recognize other humans by voice alone. This is important both socially and for the robustness of speech perception. This Thesis contains a set of eight studies that investigates how different factors impact on speaker recognition and how these factors can help explain how listeners perceive and evaluate speaker identity. The first study is a review paper overviewing emotion decoding and encoding research. The second study compares the relative importance of the emotional tone in the voice and the emotional content of the message. A mismatch between these was shown to impact upon decoding speed. The third study investigates the factor dialect in speaker recognition and shows, using a bidialectal speaker as the target voice to control all other variables, that the dominance of dialect cannot be overcome. The fourth paper investigates if imitated stage dialects are as perceptually dominant as natural dialects. It was found that a professional actor could disguise his voice successfully by imitating a dialect, yet that a listener's proficiency in a language or accent can reduce susceptibility to a dialect imitation. Papers five to seven focus on automatic techniques for speaker separation. Paper five shows that a method developed for Australian English diphthongs produced comparable results with a Swedish glide + vowel transition. The sixth and seventh papers investigate a speaker separation technique developed for American English. It was found that the technique could be used to separate Swedish speakers and that it is robust against professional imitations. Paper eight investigates how age and hearing impact upon earwitness reliability. This study shows that a senior citizen with corrected hearing can be as reliable an earwitness as a younger adult with no hearing problem, but suggests that a witness' general cognitive skill deterioration needs to be considered when assessing a senior citizen's earwitness evidence. On the basis of the studies a model of speaker recognition is presented, based on the face recognition model by V. Bruce and Young (1986; British Journal of Psychology, 77, pp. 305 - 327) and the voice recognition model by Belin, Fecteau and Bédard (2004; TRENDS in Cognitive Science, 8, pp. 129 - 134). The merged and modified model handles both familiar and unfamiliar voices. The findings presented in this Thesis, in particular the findings of the individual papers in Part II, have implications for criminal cases in which speaker recognition forms a part. The findings feed directly into the growing body of forensic phonetic and forensic linguistic research.

This thesis considers the entire automated speech recognition process and presents a standardised approach to LVCSR experimentation with HMMs. It also discusses various approaches to speaker adaptation such as MLLR and multiscale, and presents experimental results for cross­-task speaker adaptation. An analysis of training parameters and data sufficiency for reasonable system performance estimates are also included. It is found that Maximum Likelihood Linear Regression (MLLR) supervised adaptation can result in 6% reduction (absolute) in word error rate given only one minute of adaptation data, as compared with an unadapted model set trained on a different task. The unadapted system performed at 24% WER and the adapted system at 18% WER. This is achieved with only 4 to 7 adaptation classes per speaker, as generated from a regression tree.

Grice (1982) and Bar-On and Green (2010) each provide 'continuity stories' which attempt to explain how a human-like language could emerge from the primitive communication practices of non-human animals. I offer desiderata for a proper account of linguistic continuity in order to argue that these previous accounts fall short in important ways. I then introduce the recent evolutionary literature on non-cooperative communication in order to construct a continuity story which better satisfies the proposed desiderata while retaining the positive aspects of the proposals of Grice and Bar-On and Green. The outcome of this project is a more tenable and empirically investigable framework chronicling the evolution of human-like language from communicative abilities currently found in non-human animals.

This study focuses on the pragmatic transfers that emerge in the English emails produced byChinese L2 English speakers. Despite doubts about taking Chinese English as a new variety, the study believes there are some common and unique pragmatic features existing in the English text produced by Chinese L2 English speakers. 104 emails written by 13 subjects with different English proficiency and different English exposure were collected. Questionnaires were sent out to the same subjects, trying to find out the factors that affect their pragmatic performance. The study accomplished the following main findings: 1) There are differences in the extent to which pragmatic transfer occurs among different subjects. 2) The individual subject’s pragmatic performance in English is not necessarily decided by the subjects’ English proficiency. 3) The extent of pragmatic transfer in the individual subject’s case seems to be much more complex situation than depending on any single factor of the following: English proficiency, exposure to English, or confidence in using English. 4) Some email writers have different extent of pragmatic transfer in the situations with different levels of tension. 5) However, whether the subjects have different extent of pragmatic transfer or not seems again to be too complex a situation to decide which of the factors (English proficiency, exposure to English, or confidence in using English) plays a decisive role.

This essay will explain some of the key concepts of my theoretical interests in my artistic practice and point out how these are adapted in my recent work, the Satumaa triology. The trilogy consists of films concerning my own status as a Swedish speaking Finn and the history of processes of exclusion in a national community, in this particular case Finland. These processes are in my home country working both outside and within the national borders and are parallel to the French structuralist Michel Foucault’s investigations of a discourse.

This thesis considers the entire automated speech recognition process and presents a standardised approach to LVCSR experimentation with HMMs. It also discusses various approaches to speaker adaptation such as MLLR and multiscale, and presents experimental results for cross­-task speaker adaptation. An analysis of training parameters and data sufficiency for reasonable system performance estimates are also included. It is found that Maximum Likelihood Linear Regression (MLLR) supervised adaptation can result in 6% reduction (absolute) in word error rate given only one minute of adaptation data, as compared with an unadapted model set trained on a different task. The unadapted system performed at 24% WER and the adapted system at 18% WER. This is achieved with only 4 to 7 adaptation classes per speaker, as generated from a regression tree.

This exploratory study examined the relationship between public speaking anxiety levels indicated by scores on the Personal Report of Confidence as a Speaker questionnaire (PRCS: Paul, 1966) and evaluation probability on a wide domain of evaluation items reflected by scores on the Audience Attention Allocation questionnaire (devised for the purpose of this study). A large student sample (n=220) completed the PRCS as well as the AAA questionnaire. The AAA assessed the perceived allocation of the attentional resources of the audience members during a speech by asking respondents to rate how probable it is that a speaker is evaluated on a set of domains. The results of regression analyses indicated that AAA scores, Gender, and Study year were significant predictors of PRCS scores accounting for 8.5% of the variance. More interestingly, the nature of results obtained was contrary to the hypothesis of the study. It was in fact revealed that subjects scoring low on the AAA questionnaire, indicating less likelihood that audience members make evaluations about the speaker on a variety of items, tended to have higher anxiety scores. The results are discussed in terms of defense mechanisms and response bias.

A four-session text independent, TV-recorded audio-video database for speaker recognition is collected in this thesis. The speaker data is used to verify the applicability of a design methodology based on Mel-frequency cepstrum coefficients and Gaussian mixture model. Both single-session and multi-session problems are discussed in the thesis. Experimental results indicate that 90% correct rate can be achieved for a single-session 3000-speaker corpus while only 67% correct rate can be obtained for a two-session 800-speaker dataset. The performance of a multi-session speaker recognition system is greatly reduced due to the variability incurred in the recording environment, speakers¡¦ recording mood and other unknown factors. How to increase the system performance under multi-session conditions becomes a challenging task in the future. And the establishment of such a multi-session large-scale speaker database does indeed play an indispensable role in this task.