• Neteti Aswani

Most of the current literatures focus primarily on presence of single noise in corrupted speech which is far from real-world environments. In this paper the model was proposed for improving quality of text to speech using Convolutional Neural Networks(CNN) algorithm. It consists of three layers input layer, output layer and hidden layer. It takes text as input which is processed in the input layer now text is converted to speech using TTS(Text to Speech) system and noise is added to speech then CNN process this noisy speech using Mel cepstral coefficient(MCEP) it measures noise error rate ,Short Term Objective Intelligibility(STOI)it represents the speed of reducing noise and Perceptual Evaluation of Speech Quality(PESQ) it gives the quality of speech these all are done in hidden layers of CNN then clear speech without noise is given by output layer of CNN . Based on the three parameters noise , speed of reducing noise and quality of speech CNN as shown better results than recurrent neural networks(RNN).

Text to Speech, Neural Networks, RNN, CNN

"Improving quality of text to speech using neural networks", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.7, Issue 2, page no.643-648, February-2020, Available :http://www.jetir.org/papers/JETIR2002305.pdf

"Improving quality of text to speech using neural networks", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.7, Issue 2, page no. pp643-648, February-2020, Available at : http://www.jetir.org/papers/JETIR2002305.pdf