Universiti Teknologi Malaysia Institutional Repository

Statistical parametric speech synthesis of Malay language using found training data

Tan, Tian Swee and Yong, L. C. (2014) Statistical parametric speech synthesis of Malay language using found training data. Research Journal of Applied Sciences, Engineering and Technology, 7 (24). pp. 5143-5147. ISSN 2040-7459

Full text not available from this repository.

Official URL: http://www.maxwellsci.com/print/rjaset/v7-5143-514...


The preparation of training data for statistical parametric speech synthesis can be sophisticated. To ensure the good quality of synthetic speech, high quality low noise recording must be prepared. The preparation of recording script can be also tremendous from words collection, words selection and sentences design. It requires tremendous human effort and takes a lot of time. In this study, we used alternative free source of recording and text such as audio-book, clean speech and so on as the training data. Some of the free source can provide high quality recording with low noise which is suitable to become training data. Statistical parametric speech synthesis method applying Hidden Markov Model (HMM) has been used. To test the reliability of synthetic speech, perceptual test has been conducted. The result of naturalness test is fairly reasonable. The intelligibility test showed encouraging result. The Word Error Rate (WER) for normal synthetic sentences is below 15% while for Semantically Unpredictable Sentences (SUS) is averagely in 30%. In short, using free and ready source as training data can leverage the process of preparing training data while obtaining motivating synthetic result.

Item Type:Article
Uncontrolled Keywords:smoothing technique, fourier series
Subjects:Q Science > QH Natural history
Divisions:Biosciences and Medical Engineering
ID Code:62664
Deposited By: Fazli Masari
Deposited On:01 Jun 2017 10:59
Last Modified:01 Jun 2017 10:59

Repository Staff Only: item control page