APPLICATION OF NEURAL NETWORKS FOR SOUND SIGNAL PROCESSING IN THE EDUCATIONAL PROCESS

This article highlights some aspects of audio signal processing using neural networks in the educational process. The article discusses the steps of data preprocessing, including the conversion of audio signals into a numeric format and their normalization. Particular attention is paid to the use of spectral representations, such as the spectrogram and the chalk spectrogram, and their role in the analysis of sound content. The process of converting frames into a spectral representation, calculating the mel-frequency cepstral coefficients (MFCC) is described in detail. Examples of spectrograms and chalk spectrograms are presented as illustrations. In general, the article provides an overview of the main features of audio signal processing and is a useful resource for researchers and practitioners in the field of sound analysis and audio data processing.