Unveiling the State-of-the-Art: A Comprehensive Survey on Voice Activity Detection Techniques

Voice activity detection (VAD) is a binary categorization (1 or 0) task strives to recognize the occurrence or absence of a voiced portion in a section of an acoustic speech signal. It is also called speech activity detection, which plays a key role in almost all real-world speech processing applications for enhancing robustness against background noises and preventing redundant computation of unvoiced speech segments. The VAD research has gathered nearly fifty years of growth and development. Throughout this era, a variety of VAD algorithms were proposed to necessitates the rising need for technology. Developing VAD techniques that are able to work well for an unseen environment is still a challenging task. This research paper offers an extensive examination of VAD, including an in-depth analysis of its challenges and the methodologies employed to overcome them. It serves as an invaluable reference for researchers and practitioners engaged in the field of VAD processing, facilitating a comprehensive understanding and its associated techniques.