Unlimited Vocabulary Grapheme to Phoneme Conversion for Korean TTS

Byeongchang Kim,WonIl Lee,Geunbae Lee,Jong-Hyeok Lee
DOI: https://doi.org/10.48550/arXiv.cmp-lg/9806008
1998-06-10
Abstract:This paper describes a grapheme-to-phoneme conversion method using phoneme connectivity and CCV conversion rules. The method consists of mainly four modules including morpheme normalization, phrase-break detection, morpheme to phoneme conversion and phoneme connectivity check. The morpheme normalization is to replace non-Korean symbols into standard Korean graphemes. The phrase-break detector assigns phrase breaks using part-of-speech (POS) information. In the morpheme-to-phoneme conversion module, each morpheme in the phrase is converted into phonetic patterns by looking up the morpheme phonetic pattern dictionary which contains candidate phonological changes in boundaries of the morphemes. Graphemes within a morpheme are grouped into CCV patterns and converted into phonemes by the CCV conversion rules. The phoneme connectivity table supports grammaticality checking of the adjacent two phonetic morphemes. In the experiments with a corpus of 4,973 sentences, we achieved 99.9% of the grapheme-to-phoneme conversion performance and 97.5% of the sentence conversion performance. The full Korean TTS system is now being implemented using this conversion method.
Computation and Language
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to explore and understand the different evolution paths of strange non - chaotic attractors (SNA) in quasi - periodically forced systems and how they eventually transform into chaos. Specifically, the author studied the formation mechanism and different evolution paths of SNA in a specific quasi - periodically forced velocity - dependent non - polynomial oscillator system in the two - parameter (f - ε) space. Through this research, the author hopes to reveal the following points: 1. **Formation mechanism of SNA**: Determine how SNA evolves from the quasi - periodic state of two frequencies. 2. **Different evolution paths**: Show and explain the paths through which SNA can evolve into chaos, including but not limited to: - Through torus doubling to chaos - Through torus breakdown to chaos - Through torus doubling, shrinking, SNA and then to chaos 3. **Discovery of new phenomena**: Specifically pointed out some new bifurcation phenomena, such as torus bubbling, which play an important role in the formation process of SNA. ### Summary of main content #### Research background and problem description The paper focuses on strange non - chaotic attractors (SNA) in quasi - periodically forced dynamical systems. The author explored several possible paths through which SNA can appear in the same model and eventually evolve into chaos. Specific paths include: - Quasi - periodicity of two frequencies → torus doubling → chaos - Quasi - periodicity of two frequencies → torus breakdown → chaos - Quasi - periodicity of two frequencies → torus doubling → shrinking → SNA → chaos → SNA → inverse torus doubling → torus → torus bubbling → torus breakdown to chaos #### New findings The author particularly emphasized a new phenomenon - torus bubbling, that is, under certain parameter conditions, the torus doubling sequence will merge in certain parameter intervals, suppressing the transformation of the torus doubling path to chaos. This re - merging bifurcation phenomenon makes the orbit stable again, and the motion in these regions is regular and predictable. #### Experimental verification To confirm the existence of SNA, the author calculated multiple characteristic quantities, such as Lyapunov exponents, winding numbers, power - spectrum analysis and dimensions. These characteristic quantities are helpful to distinguish quasi - periodic, strange non - chaotic and chaotic attractors. ### Conclusion Through the detailed study of the above system, the author showed multiple paths to chaos and revealed the key role of SNA in it. In particular, the author discovered the new phenomenon of torus bubbling, providing a new perspective for understanding non - chaotic strange attractors in complex dynamical systems.