Features of the Implementation of Real-Time Text-to-Speech Systems With Data Variability

This paper discusses the theoretical aspects of real-time text-to-speech synthesis systems and the implementation of such systems based on deep learning models. Such systems consist of two essential components: a text-tospeech generator responsible for generating the audio content of speech and creating mel-spectrograms, and a vocoder that synthesizes speech based on voice data. In this work, an implementation of the text-to-speech generator based on the Tacotron-2 model and a vocoder based on the WaveRNN model is provided. Their advantages and disadvantages are examined based on the obtained results, and potential solutions for adapting the system to different languages are also discussed.