Click here for full text:
South Asian Languages in Multilingual TTS-Related Database
Shalonova, Ksenia; Tucker, Roger
Keyword(s): text-to-speech technology
Abstract: In this paper we overview possible problems to be overcome in building TTS systems for different languages, in particular for the languages of South Asia. We do this by an analysis of both script and language features (presented in the Multilingual TTS- Related Database), and observe that all languages (not just in South Asia) have a limited number of these features. We briefly describe possible TTS problems in accordance with the described script and language categories (several examples from South Asian languages are provided). By attributing scores to the features, we can rank the languages in order of difficulty. On this basis, Bengali is one of the easiest languages and Pashto is one of the most difficult.
Back to Index