Generalized Sidelobe Canceller

Resource Overview

Implementation of Generalized Sidelobe Canceller (GSC) adaptive beamforming method with time-domain and frequency-domain filtering using LMS adaptive algorithm for speech enhancement. The package includes clean speech samples and noisy speech samples with different SNR ratios for experimental validation.

Detailed Documentation

This project implements the Generalized Sidelobe Canceller (GSC) adaptive beamforming approach, performing simultaneous time-domain and frequency-domain filtering through the LMS adaptive algorithm to achieve speech enhancement. The method effectively suppresses background noise while improving clean speech quality through adaptive weight updates. The implementation typically involves three main components: a fixed beamformer for signal preservation, a blocking matrix for noise reference generation, and an adaptive noise canceller using LMS algorithm for noise suppression. The package includes clean speech samples and noisy speech samples with varying Signal-to-Noise Ratio (SNR) levels, facilitating comprehensive testing and performance evaluation of the noise reduction algorithm under different acoustic conditions.