Taylor P. and Black A. W., "Speech Synthesis by Phonological Structure Matching", in Proceedings of Eurospeech'99, pp. 623-626, 1999  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the advent of using real human speech, rule based formant methods as shown in Figure 1 1 were popular. Unit selection was trivial with a database of very few (one) examples, while much manual tuning was required in the modification and generation stages. Hard decisions made by decision trees [13, 59, 148, 29] or omission of units from the corpus [9, 38] are typically made before the unit selection search to prune the search space. Intonation contours can be applied by waveform modification as part of a post processing stage [12] or it can be modelled in the unit selection search [29, 19, 21] Speech ....

....the size of the search space. In the AT T s NextGen synthesizer [7] over 99 of the transitions are determined to be rarely used after profiling over a training set and are simply omitted to make their unit selection algorithm tractable [9] In other works that make explicit use of decision trees [13, 148, 38], only the exemplars at the leaf nodes are licensed to populate the trellis in the Viterbi search. Although a decision tree could be incorporated into L decision trees can be represented by FST s [134] the benefit of using the proposed constraint kernel is that soft decisions are employed and ....

P. Taylor and A. Black, "Speech synthesis by phonological structure matching," in Proc. Eurospeech '99, Budapest, Hungary, Sept. 1999, vol. 2, pp. 623--636.

.... in many recent conferences reveal a global quest for unit selection systems providing highly intelligible and natural speech [1 5,19] More recently, efforts have also been devoted to achieving this goal with manageable computational cost and storage requirements [6,16] Phonologically grounded [7,17] or acoustically grounded [8] all such selection algorithms also need a synthesis engine, if only for properly concatenating speech segments. In practice, such engines range from almost nothing (as originally proposed by Nick Campbell [9] to frequencydomain synthesizers (as in proposals by the ....

Taylor, P. and Black, A.W. "Speech Synthesis by Phonological Structure Matching", Proc. of EUROSPEECH, Budapest, Hungary, Sept. 1999.

....and clustered acoustic data. The attribute set comprises a subset of the classical distinctive features (Chomsky and Halle, 1968) articulatory features, features related to syllable structure, and other suprasegmental features. Segment identity is assigned higher priority than context matching. Taylor and Black (1999) present a similar approach, phonological structure matching, where phonological information, such as canonical pronunciation, positional factors and accentuation, is used for unit selection instead of narrow phonetic transcriptions and absolute duration and F 0 values. In contrast to the BT ....

....with the highest node level and resorting to daughter nodes whenever no candidate 99 is found. As a result of the search, candidate units will appear at various positions and levels in the tree, and they will correspond to units of arbitrary length in the database. One potential drawback of Taylor and Black s (1999) approach is that word boundaries appear to represent hard boundaries in the phonological tree. The authors argue that coarticulation has been found to be stronger within constituents than across constituent boundaries at all levels in the tree. However, the claim that the word boundary is a ....

Taylor, Paul and Alan W. Black. 1999. Speech synthesis by phonological structure matching. In Proceedings of the European Conference on Speech Communication and Technology (Budapest, Hungary), volume 2, pages 623-626.

....the database as a consequence of the selection algorithm. Some system however explicitly allow for mixed sized units. Bonn s HADIFIX system was more explicit in its varying unit length including consonant clusters sized units as well as single phone units [10] Phonological Structure Matching [11] is explicit in its selection of non uniform lengthed units. The database is labelled with tree structures. An utterance to be synthesized is also labelled with a tree structure. The database is then searched top down for the largest sub trees that are contained within the desired utterance. Thus ....

P. Taylor and A. Black, "Speech synthesis by phonological structure matching," in Eurospeech99, Budapest, Hungary, 1999, vol. 4, pp. 1531--1534.

....use. Due to the sheer scope of the text to speech conversion process, it is commonplace in speech synthesis for a variety of techniques to be used throughout the system. While important work has been done on trying to unify certain stages of the conversion process (Local, 1992) Sproat, 1996) (Taylor and Black, 1999), for the foreseeable future a variety of techniques will need to be used within synthesis systems. The baseline requirement for the formalism is that it can represent the required input and output information for any linguistic processing module, no matter what type of linguistic process the ....

Taylor, P. A. and Black, A. W. (1999). Speech synthesis by phonological structure matching. In Eurospeech 99.

No context found.

Taylor P. and Black A. W., "Speech Synthesis by Phonological Structure Matching", in Proceedings of Eurospeech'99, pp. 623-626, 1999

No context found.

Taylor P. and Black A. W., "Speech Synthesis by Phonological Structure Matching", in Proceedings of Eurospeech'99, pp. 623-626, 1999

No context found.

Taylor, P. and Black, A. W. "Speech Synthesis by Phonological Structure Matching." Proceedings of the 6 th European Conf. on Speech Communication and Technology, Budapest, Hungary, 1999, vol. II, p623626.

No context found.

Paul Taylor and Alan W. Black, "Speech synthesis by phonological structure matching," in Proc. European Conference on Speech Communication and Technology (Budapest, Hungary), 1999, vol. 2, pp. 623--626.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC