Syllable-Based Concatenative Speech Synthesis for Marathi Language

Abstract

The objective of speech synthesis is to convert text into speech, using the syllable got best the results in term quality speech. Speech processing plays important role in human–machine interaction. Syllable segmentation is done using the different parameters like cutoff frequency, amplitude, the magnitude of speech input signal. The proposed paper suggests a method of syllable segmentation; it converts the syllable representation to modified syllable waveform clips that can be combined together to produce as sound [1]. Syllables are a better choice for speech synthesis in Marathi language. We do manually identify, segment, and label syllable units from speech data for experiment purpose [2]. The concatenative speech synthesis methods provide highly understandable speech utterance.