Vaja is a Thai text-to-speech software developed by Human Language Technology Laboratory (HLT). Vaja has been extensively researched and developed since 1997. The most recent version Vaja 6.0 utilizes the latest speech synthesis technique HTS which uses a statistical model called Hidden Markov to produce synthesized sound. This technique helps eliminate the problem of uneven sound that occurred in the previous versions. The new synthesis technique together with a prosody prediction module which predicts phrase boundaries and the duration of each phone make the synthesized speech sounds more natural and increase users’ satisfaction. Furthermore, Vaja 6.0 is able to synthesize all Thai words since it has a text analysis module which can generate the pronunciation of every word even the one not found in a dictionary.
Vaja 6.0 Home Edition under the code name “jRaja” can be downloaded free of charge for personal use and evaluation purpose. For advanced users or developers who would like to use enhanced features, such as Vaja API, editable dictionary for new words and pronunciations, can contact us to license Vaja 6.0 Professional Edition.