Voice Frequency Analysis Based on Gender and Age Using FFT & FIR

Abstrak

This study aims to analyze the dominant frequency characteristics of human voices based on gender and year of birth using the Fast Fourier Transform (FFT) method and Python-based Finite Impulse Response (FIR) filters. The research data consisted of 90 human voice samples, namely 45 male respondents and 45 female respondents with a range of birth years 2002–2016. Each sample was processed through preprocessing, FFT transformation, and screening stages using a low-pass FIR filter with a cutoff frequency of 400 Hz. The results showed that the average dominant frequency of women's voice was higher than that of men, which was 377.630 Hz and 356.877 Hz. In addition, respondents with younger age tended to have a higher dominant frequency. The application of FIR filters also succeeded in improving signal quality by reducing noise, making frequency analysis more accurate.