Estimation of Spectral Mismatch for Joint Cost Evaluation in Marathi TTS

Among different methods of speech synthesis, Concatenative Speech Synthesis is widely used due to its naturalness and less signal processing requirement. But concatenative TTS has problems like requirement of large database and resulting spectral mismatch in output speech. In concatenative TTS position of syllable plays very important role while carrying out segmentation. If proper position syllable is used while forming new words from existing syllables, resulting spectral mismatch is less. If position of syllable is not considered during concatenation of speech units, resulting synthesis end up in more concatenation cost. This paper presents different techniques like PSD, Wavelet and DTW to find spectral mismatch in concatenated segments. In all these three techniques PSD results are more superior who shows spectral mismatch in graphical form. With direct formant modification we can overcome spectral mismatch and smooth some of the frames which helps to reduce glitch type of sound at concatenation point. Wavelet based audio results shows more naturalness compare to other two methods. In proposed work the discontinuities at the cutting point are smoothed by changing the spectral characteristics before and after the cutting point so that the spectral mismatch is equally distributed over the number of adjacent frames. This work throws light on how spectral mismatch calculation and reduction increases naturalness of concatenative Marathi TTS.