Neural Network-Based Models for Speech Synthesis (EN, SP, CA)

These are speech synthesis models (automatic conversion of text to synthetic speech) for English (EN), Spanish (ES) and Catalan/Valencian (CA). For each of these languages, a different model is included. They solve the problem of automatic generation of voice instructions and audiobooks, speech production in robots and chatbots, and automatic dubbing of videos and audios. They are similar models to those used by the current commercial systems of large technology companies, although they have been obtained in the educational field, so they are better adapted to it. From a technical point of view, these models are trained deep neural networks. Speech synthesis software (not included in this record) requires one of these models to perform speech synthesis in the corresponding language.