Making Machines Talk In Sanskrit: A systematic exploration Of Text-To-Speech Synthesis For Sanskrit Language
Keywords:
Sanskrit Text-to-Speech (STTS), Text-to-Speech (TTS), Natural Language Processing (NLP), Grapheme-to-Phoneme (G2P), Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA), Deep Neural Network (DNN), Population, Intervention, Context and Outcome (PICO), Voice Operating Demonstrator (VODOR), Orator VerbisElectris (OVE), MUltichannel Speaking Automaton (MUSA), Non Standard words (NSW), Unit Selection Synthesis (USS), Concatenative Speech Synthesis (CSS), Statistical Parametric Speech Synthesis (SPSS), Hidden Markov Model (HMM), Generative Adversarial Network (GAN), Gaussian mixture models (GMM).Abstract
Introduction: For more than four thousand years, Sanskrit, one of the most important languages, has served us. Owing to the philological, social, scientific, and pharmacological significances, Sanskrit needs to be preserved. It can only be done, when it will reach to young minds and people will get the resources to learn, read and use this language.
Context: In the last two decades, few researchers have tried to implement Sanskrit Text-to-Speech (STTS) Systems. In this study, we seek to establish the current state-of-art of STTS by comprehensively identifying and analyzing the previous work.
Methods and Procedures: Accordingly, the article represents findings of the Systematic Literature Review (adhered to the model of Kitchenham and Charter) that collected the most relevant literature produced from 2002 to 2024. After the search conducted on 7 databases, the papers were identified as primary studies and were analyzed in detail.
Outcome and Results: The outcome of this systematic literature review (SLR) depicted that there is a lack of Sanskrit Text-to-Speech system (STTS) related work and publications in high-quality journals. It also elucidates the framework for the development and execution of the Sanskrit Text-to-Speech system (STTS).
Conclusion and Implications: This paper examines the possible methodologies, methods, challenges, and limitations of the Sanskrit Text-to-Speech system (STTS) in order to gain a better understanding of the research dynamics in speech synthesis. Additional Keywords and Phrases: TTS (Text-to-Speech), Systematic review, Natural Language Processing (NLP), Speech Synthesis, Grapheme-to-Phoneme (G2P) conversion, Deep Neural Network (DNN) etc.
