Text-to-Speech Synthesis for Mirandese
The Mirandese language (autonym: mirandés or lhéngua mirandesa) is a Romance language belonging to the Astur-Leonese linguistic group, sparsely spoken by around 15 000 citizens in a small area of northeastern Portugal. The Portuguese Parliament granted it co-official recognition (along with the Portuguese language) for local matters on 17 September 1998. (http://en.wikipedia.org/wiki/Mirandese_language)
The project already developed several basic linguistic resources for speech technology in Mirandese, namely:
The language first reference large text corpus.
A complete proposal for the language phone set.
Language resources for TTS:
The first large phonetic lexicon (100 K entries).
POS tagger, inflectioner, text normalization module, letter-to-sound rules, automatic stress and syllable marker.
High quality voice talent data base with 5000 utterances.
Preliminary results of the project were published in Phonetics and Phonology in Iberia, Univ. of Tarragona, held in June, 2011 in Spain:http://download.microsoft.com/download/A/0/B/A0B1A66A-5EBF-4CF3-9453-4B13BB027F1F/PaPI2011_MLDC_poster.pdf