Oriental COCOSDA — Country report 2016 language resources developed in Taiwan
Sin‐Horng Chen, Chiu-yu Tseng · 2016
NCTU — Speech Corpora • Multi-Speaker Mandarin Speech Corpus — Application: speaker adaptation for Mandarin TTS — Size: 5K∼10K syllables/speaker; 9 male and 6 female speakers. — Format: 20 kHz sampling rate/16 bits depth — Speaking rate: ranging from 0.15 to 0.3 sec/syl (or 3.4 to 6.8 syl/sec) • Taiwan Min and Si-Xian Hakka Dialectal Speech Corpus — Application: cross-dialect speaker adaptation for Chinese dialect TTS — Size: 23K and 20K syllables for Taiwan Min and Si-Xian Hakka — Format: 20 kHz sampling rate/16 bits depth — Text: 255/200 short paragraphs selected from story books for Min/Hakka.