Efficiency improvements to an auditory-inspired synchrony capture filterbank

Area of Science:

• Auditory Neuroscience
• Signal Processing
• Biomimetic Engineering

Background:

• Prior work established an auditory-inspired adaptive synchrony capture filterbank (SCFB) for acoustic frequency tracking.
• The earlier SCFB architecture utilized a weighted sum of frequency discriminator loop (FDL) and phase-locked loop (PLL) for precise frequency component tracking.

Purpose of the Study:

• To introduce adaptive weights for FDL and PLL within the SCFB architecture.
• To incorporate a place theory-based mechanism for preferential filter activation.
• To enhance biomimicry and computational efficiency in acoustic frequency tracking.

Main Methods:

• Developed an adaptive SCFB architecture with dynamic FDL and PLL weights.
• Integrated a place theory-based mechanism to reduce filter activation.
• Tested the enhanced SCFB using synthetic time-varying signals under various noise conditions.

Main Results:

• Demonstrated improved performance of the adaptive SCFB architecture.
• Showcased enhanced biomimicry through preferential filter activation.
• Validated robust denoising capabilities in challenging noisy environments.

Conclusions:

• The adaptive SCFB architecture with place theory integration offers superior acoustic frequency tracking.
• The enhanced model achieves greater computational efficiency and biomimicry.
• The algorithm exhibits robust performance and denoising capabilities, suitable for real-world applications.