Research of the impact of noise reduction methods on the quality of audio signal recovery

Анотація

The subject of the study is the analysis  of various filtering algorithms for the  quality of the resulting audio files. The  importance  of audio line filtering has  grown significantly in recent years due  to its key role in a variety of  applications such as speech  reduction and artificial intelligence. Taking into account the growing demand for  solving problems related to speech  recognition, the processing of audio series becomes important for  determining the accuracy and  efficiency of the obtained  solution.
The purpose of the work is to study the impact of noise suppression methods on the quality of restoration of an audio signal, which was alternately noisy with one of five types of noise - white, pink, brown, impulse, Gaussian with different power. To achieve the goal, the following tasks were solved: an analysis of the types of noise was carried out and analysis of noise reduction and filtering methods. A generalized model of noise reduction and filtering was developed, and an experiment was planned depending on the type and power of noise.
Simulation of the experiment was  performed by comparing the  parameters of the signal-to-noise ratio before and after the experiment and  the peak signal-to-noise ratio in the  processed file. The following methods  are used: spectral subtraction, filtering  based on frequency filters and wavelet transformation.
The following results were obtained:  depending on the selected noises and  algorithms, it was possible to achieve  the lowest value of the peak signal-to-noise ratio of 21.52db, and the signal-to-noise ratio increased, which allowed further work with these audio files. The practical significance of this work is the increase in the number of available  audio files for further work.
Conclusions: the analysis of the  obtained results showed that filtering based on frequency filters only  worsened the output signal, that is, not only noise, but also useful information is filtered. In all runs, the SNR deteriorates to - 18dB. which is worse than no filtering. Algorithms of spectral subtraction and wavelet transformation improved SNR parameters and output audio files noisy with the most powerful noises in the range of 20dB, which can be considered acceptable for further processing. The results highlight the importance of using denoising and filtering for complex audio processing tasks, particularly neural  network  training tasks.