Design and Implementation of a Real-Time Acoustic Feature Visualization System on Raspberry Pi Platform Utilizing Fixed-Point FPGA-Based Mel Spectrum Extraction

This paper presents the design and implementation of a real-time acoustic feature visualization system on the Raspberry Pi platform, incorporating a fixed-point FPGA-based mel spectrum extraction module. The system supports efficient computation of key audio features, particularly the mel spectra, log-mel spectra, and mel-frequency cepstral coefficients (MFCCs), which are computed per frame and widely used in audio classification and acoustic analysis. By offloading mel spectrum extraction to an FPGA and utilizing fixed-point arithmetic, the system achieves low-latency and energy-efficient signal processing, enabling real-time performance on resource-constrained hardware. The Raspberry Pi is responsible for data acquisition and MFCC calculation from the FPGA-generated mel spectra. Additionally, it performs graphical rendering. This configuration forms a low-cost and low-power platform for real-time acoustic applications such as industrial anomaly detection, environmental sound monitoring, and other edge-based acoustic intelligence tasks. Implementation results demonstrate that the system provides accurate and responsive visualization of audio features.