Voice building for a new language

1. Download xml dump of wikipedia in your language

2. Extract clean text and most frequent words

3. Transcribe most frequent words

1. Create pronunciation dictionary and train letter-to-sound rules
2. Minimal NLP components for the new language

4. Run feature maker with the minimal nlp components

5. Database selection

select a phonetically/prosodically balanced recording script

6. Manually check/correct transcription of all words in the recording script [Optional]

7. Record script with a native speaker using our recording tool "Redstart"

8. Build an unit selection and/or hmm-based voice with Voice import tool