Protocols for Acquiring Reusable Speech Materials and Presenting Them Reliably Using Structured Test Signals
This paper proposes protocols and tools for acquiring speech materials that are reusable for future investigations, and for presenting speech stimuli more reliably. The protocols and tools are built upon structured test signals and analysis methods, including a new family of Time-Stretched Pulse (TSP) signals, to enable objective evaluation and classification of speech materials.
Improved Direction-of-Arrival Estimation for Multiple Speakers in Dynamic and Reverberant Environments Using a Wearable Microphone Array
This paper proposes an improved direction-of-arrival (DOA) estimation algorithm for multiple speakers in dynamic and reverberant environments using a wearable microphone array. The algorithm incorporates a reliability weight and a quality measure to enhance the robustness and accuracy of DOA estimation under challenging conditions.
A Portable and Scalable Active Noise Mitigation System for Reducing Low-Frequency Construction Machinery Noise
A portable and scalable active noise mitigation system can effectively reduce low-frequency noise emissions from construction machinery through adaptive algorithms and remote management capabilities.
Improving Sound Source Localization Accuracy Through Steered Response Power Techniques: A Comprehensive Review
The Steered Response Power (SRP) method is a widely used technique for sound source localization that can provide satisfactory performance in moderately reverberant and noisy scenarios. This work presents a comprehensive review of over 200 papers on SRP and its variants, with a focus on improving the method's computational efficiency, robustness in adverse environments, and ability to localize multiple sources.
© 2024 by Linnk AI