Summary

This master's thesis presents a practical real-time implementation that combines a delay-sum beamformer and acoustic-echo-canceller to improve far-field speech capture in telephony. The reader will learn how microphone-array beamforming and adaptive echo cancellation interact and how they were implemented and evaluated on real-time DSP hardware.

Key Takeaways

• Implement a delay-sum beamformer to spatially enhance microphone-array signals.
• Design and integrate an adaptive acoustic-echo-canceller to remove loudspeaker echo in hands-free telephony.
• Combine beamforming and echo cancellation while managing interaction effects such as residual echo and double-talk.
• Optimize algorithmic and implementation trade-offs for real-time embedded DSP platforms (latency, complexity, and fixed-point issues).
• Measure performance using objective metrics (e.g., SNR/ERLE) and evaluate robustness in reverberant and noisy environments.

Who Should Read This

Graduate students, DSP engineers, or system designers working on hands-free audio, teleconferencing, or embedded real-time signal processing who need practical guidance on beamforming and echo cancellation integration.

Still RelevantAdvanced

Topics