论文信息 - An evaluation of Mongolian data-driven Text-to-Speech

An evaluation of Mongolian data-driven Text-to-Speech

This paper presents a first attempt to evaluate data-driven speech synthesis of Mongolian trained on 1500-sentence female speech corpus. The speech corpus contains nearly 6 hours of Mongolian female speech that is designed to cover all Mongolian phones. The evaluation is done on two levels. In overall quality evaluation, we generated 25 sentences and asked raters about their quality based on Mean Opinion Score (MOS). The second evaluation uses Phoneme confusion test, which contains all possible phoneme set in Mongolian.

Chatchawarn Hansakunbuntheung | Altangerel Chagnaa | Purev Jaimai | Kerey Yesyenbyek

[1] Purev Jaimai,et al. Corpus building for Mongolian language , 2008, IJCNLP.

[2] Hui Zhang,et al. Mongolian Speech Recognition Based on Deep Neural Networks , 2015, CCL.

[3] Multilingual Text – Speech Corpus of Mongolian , 2006 .

[4] Peter Kulchyski. and , 2015 .

[5] Kuldip K. Paliwal,et al. Diphone-Based Concatenative Speech Synthesis System for Mongolian , 2008, IMECS 2008.

[6] Munkhtuya Davaatsagaan,et al. A New Approach for the Mongolian Text-To-Speech Conversion , 2006 .

[7] Ausdang Thangthai,et al. Mongolian speech corpus for text-to-speech development , 2011, 2011 International Conference on Speech Database and Assessments (Oriental COCOSDA).