Meri Awaaz Hi Pehchan Hai: A Survey on Multilingual Podcast Processing System

Dr. Vijayalaxmi Mekali, Arpita Rathod, Bhagyashree, Gagana Poojari, Gayana V

Department of Computer Science and Engineering

K. S. Institute of Technology, Bengaluru – 560109, India vijayalaxmimekali@ksit.edu.in

Abstract—Podcast consumption has increased rapidly in re-cent years, making podcasts an important source of education, entertainment, and information sharing. In multilingual countries like India, audiences prefer podcast content in regional languages such as Kannada, Hindi, and Telugu. Processing and summariz-ing multilingual podcasts remains challenging because of regional accents, code-switching, dialectal variations, and long audio content. The survey reviews multilingual podcast processing and summarization techniques, including Automatic Speech Recog-nition (ASR), Natural Language Processing (NLP), extractive and abstractive summarization, machine translation, and Text-to-Speech systems. The study also examines recent advancements in transformer-based models for low-resource Indian languages and discusses datasets, evaluation metrics, research challenges, and future improvements in multilingual podcast summarization.Index Terms—Keywords: Podcast Summarization, Multilin-gual Automatic Speech Recognition (ASR),Natural Language Processing (NLP), Indian Languages, Transformer Models.