Natural-and-Linguistic-Text-To-Speech-Synthesis

A real time system to synthesise emotional and natural human-like sound from unstructured text was built. The whole system has three parts. The first part deals with extracting emotion out of text using text mining and LSTMs. The second part identifies the key features of speech for Speech Emotion Recognition. In the third part these attributes are modified in a neutral speech to give it an emotional and human-like base.