Overcoming Linguistic Barriers Developing Advanced Urdu Text-to-Speech Systems

Our research invetstigates the technical obstacles and promising potential of Urdu Text-to-Speech (TTS) technol-ogy. To meet the growing demand for voice-based interfaces and accessibility solutions in Urdu-speaking communities, we give a detailed analysis of the Urdu TTS landscape, including method-ology, advancements, limitations, and potential applications. The study explores into the creation of a cutting-edge TTS system for Urdu, employing the advanced features of the SpeechT5 model within the HuggingFace Transformers library. We rigorously fine-tuned this model using the publicly available Mozilla Common Voice dataset, applying cutting-edge machine learning techniques to generate improved speech synthesis results. Our findings highlight the efficacy of the SpeechT5 model for Urdu TTS, while underlining the necessity for continued modifications to address the intricacies of naturalness and pronunciation and voice rate unique to the Urdu language. This study makes a significant contribution to the improvement of Urdu TTS technology, promoting digital inclusion and increasing accessible for Urdu speakers around the world.