Spectral Analysis of Musical and Random Noise Signals with Audio Classification: A Comparative Study
Abstract
This study presents a comparative analysis of two types of audio signals: random noise and music audio. The random noise was recorded in a crowded public bus on an urban road, while the music audio was captured using the same device in a controlled environment, free from external disturbances. Statistical and spectral analyses, including histogram analysis, power spectrum, Fast Fourier Transform (FFT), and Auto-Correlation Function (ACF), were employed to evaluate the signals. The noise signal showed a bell-shaped histogram with samples concentrated around the mean, an ACF that decreased with lag, and the FFT spectrum, which shows that the energy is distributed uniformly across all frequencies. In contrast, the music signal displayed a sharp histogram, distinct voice and music components, rhythmic repetitions, and specific frequencies in its ACF and FFT spectra. Additionally, Support Vector Machine (SVM) classification was performed on both signals. The noise signal was accurately categorized into three noise levels with an accuracy of approximately 82%, while the music audio was categorized into voice, music, and noise components with 90% accuracy. The superior classification performance for music is attributed to the clear separation between voice and music components. This study demonstrates the potential of SVM in audio signal classification, providing a robust approach for distinguishing between noise and music audio.
Keywords
Full Text:
PDFReferences
A. F. Siegel and M. R. Wagner, “Time series,” in in Practical Business Statistics, Elsevier, 2022, pp. 445–482. doi: 10.1016/b978-0-12-820025-4.00014-2.
M. Müller and A. Klapuri, “Music Signal Processing,” in Academic press library in signal processing, Elsevier, 2014, pp. 713–756. doi: 10.1016/b978-0-12-396501-1.00027-3.
R. Awati, R. Sheldon and J. Burke, “What is signal-to-noise ratio and how is it measured?,” Search Networking, TechTarget, Jun. 3, 2025. Available: https://www.techtarget.com/searchnetworking/definition/signal-to-noise-ratio [Accessed: Jun. 4, 2025]
M. Muller, D. P. W. Ellis, A. Klapuri, and G. Richard, “Signal Processing for Music Analysis,” IEEE Journal of Selected Topics in Signal Processing, vol. 5, no. 6, pp. 1088–1110, Oct. 2011, doi: 10.1109/jstsp.2011.2112333.
J. L. Rojo-Alvarez, C. Figuera-Pozuelo, C. E. Martinez-Cruz, G. Camps-Valls, F. Alonso-Atienza, and M. Martinez-Ramon, “Nonuniform Interpolation of Noisy Signals Using Support Vector Machines,” IEEE Transactions on Signal Processing, vol. 55, no. 8, pp. 4116–4126, Aug. 2007, doi: 10.1109/tsp.2007.896029.
J. R. Reddy, K. K. P. S. Reddy, M. M. Ali, and U. Sesadri, “Music Genre Classification Using Machine Learning,” International Journal for Multidisciplinary Research, vol. 6, no. 1, Jan. 2024, doi: 10.36948/ijfmr.2024.v06i01.8857.
C. Wang and X. Geng, “Simulation of music style classification model based on support vector machine algorithm,” in Sixth International Conference on Intelligent Computing, Communication, and Devices (ICCD 2023), S. Patnaik, Ed., Hong Kong, China: SPIE, Jun. 2023, p. 86. doi: 10.1117/12.2682982.
K. Kumar, “Audio classification using ML methods,” arXiv (Cornell University), May. 2023, doi: 10.48550/arxiv.2305.19304.
M. Xiang, “Music Vocal Spectrum Classification Model based on Improved SVM Algorithm,” 2022 International Conference on Augmented Intelligence and Sustainable Systems (ICAISS), pp. 1244–1247, Nov. 2022, doi: 10.1109/icaiss55157.2022.10010883.
S. Kavitha and J. Manikandan, “Improved Methodology of SVM to Classify Acoustic Signal by Spectral Centroid,” Journal of Trends in Computer Science and Smart Technology, vol. 3, no. 4, pp. 294–304, May 2022, doi: 10.36548/jtcsst.2021.4.005.
L. Pallotta, M. Neri, M. Buongiorno, A. Neri, and G. Giunta, “A Machine Learning-Based Approach for Audio Signals Classification using Chebychev Moments and Mel-Coefficients,” in 2022 7th International Conference on Frontiers of Signal Processing (ICFSP), Paris, France: IEEE, Sep. 2022, pp. 120–124. doi: 10.1109/ICFSP55781.2022.9924832.
K. Zaman, M. Sah, C. Direkoglu, and M. Unoki, “A survey of audio classification using deep learning,” IEEE Access, vol. 11, pp. 106620–106649, Jan. 2023, doi: 10.1109/access.2023.3318015.
S. Akinpelu and S. Viriri, “Speech Emotion Classification: A Survey of the State-of-the-Art,” in Pan-African Artificial Intelligence and Smart Systems, T. M. Ngatched Nkouatchah, I. Woungang, J.-R. Tapamo, and S. Viriri, Eds., Cham: Springer Nature Switzerland, 2023, pp. 379–394. doi: 10.1007/978-3-031-25271-6_24.
“R Studio IDE”, Posit Software, PBC. Available: https://posit.co/downloads/
Copyright (c) 2025 Sandip Bordoloi, Mayuri Sarmah Mazinder, Rashmi Rekha Roy
This work is licensed under a Creative Commons Attribution 4.0 International License.
- Download the AJEEE paper template and submit your paper at the PAPER SUBMISSION page.
ADBU Journal of Electrical and Electronics Engineering (AJEEE) - ISSN: 2582-0257 is an International peer-reviewed Open-Access Online journal in the English language that publishes scientific articles which contribute new novel experimentation and theoretical work in all areas of Electrical and Electronics Engineering and its applications. |
* The views, interpretations and opinions expressed in the articles are those of the author(s) and should not be considered to reflect the opinions of the Editorial Board of this journal- AJEEE.
 | This Journal is published under the Creative Commons Attribution 4.0 International License (CC-BY). |