MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641075790 A) filed by Vardhaman College Of Engineering on June 18, 2026, for System For Real-Time Acoustic Echo Cancellation And Spatial Audio Isolation.
Inventors include K. Narsimha Reddy; Sunkaraboina Sreenu; A. Vijaya Lakshmi; Dr. V. Purna Chandra Reddy; Ch. Sreenivasa Rao; and K. Ashwini.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: System for Real-Time Acoustic Echo Cancellation and Spatial Audio Isolation is the proposed invention. The invention employs a multi-channel microphone array, loudspeaker reference monitoring, transformer-based acoustic modelling and AI-based signal processing to suppress acoustic echo, reverberation and environmental interference. An adaptive deep learning engine for nonlinear acoustic propagation path estimation and real-time echo prediction and cancellation utilising temporal attention mechanisms. At the same time, a spatial audio isolation module with intelligent beamforming and convolutional recurrent neural networks detects and isolates target speech sources from overlapping voices and background noise. Spectral masking, adaptive gain control, and residual noise suppression techniques are used to enhance the processed audio signals and improve speech intelligibility and listening clarity. The architecture is optimised for low-latency edge deployment via neural network quantisation and parallel processing, enabling implementation in conferencing systems, hearing aids, smart devices, virtual communication platforms, and immersive audio applications. The advantages of the proposed invention are improved accuracy of echo suppression, adaptive environmental awareness, improved spatial sound isolation and stable real time performance in complex acoustic environment.
Disclaimer: Curated by HT Syndication.
