<![CDATA[This study evaluates the effectiveness of a preprocessing pipeline consisting of resampling, normalization, and Finite Impulse Response (FIR) filtering to improve signal consistency for further signal analyses such as feature extraction and classification. Resampling standardizes the sampling rate to 16 kHz, ensuring uniform temporal resolution. Normalization adjusts amplitude across recordings, yielding a mean of 0,0015 and a standard deviation 0,0462. FIR filtering reduces noise, eliminating 77,69% of signal energy above 200 Hz while retaining 29,65% of the main signal. Pipeline evaluation shows a Signal-to-Noise Ratio (SNR) of -1,88 dB, indicating a significant power reduction, but normalization ensures amplitude stability. These results demonstrate that this preprocessing combination effectively reduces noise, although balancing noise reduction and signal preservation remains challenging.]]>
Copyrights © 2025
