Análisis comparativo de técnicas de reducción de ruido en señales de audio utilizando Python

Abstract

Abstract: This curriculum integration work (TIC) presents the comparative analysis of three audio signal noise reduction techniques implemented in Python: moving average filtering, the LMS algorithm of adaptive filtering and the Boll algorithm of spectral suppression filtering. The study focuses on two specific types of noise: Additive White Gaussian Noise (AWGN) and Rayleigh fading (multiplicative noise). To evaluate each technique, metrics such as signal- to-noise ratio (SNR) and mean square error (MSE) were used, presenting the results through time domain plots and spectrograms. In addition, it studies the computational complexity and relevant factors for the implementation of each technique in Python. All the work developed provides a solid basis for the understanding and practical application of these techniques in digital audio signal processing. Finally, the structure of this TIC is composed in four chapters, which cover the state of the art, the theoretical foundations, the methodology and implementation, and the comparative analysis of the results obtained.