Harmonic analysis in speech recognition systems

Gaikwad, Amol D.; Kulkarni, Haripriya H.; Patel, Prashant Bansilal; Jape, Vidula Sanjay; Soman, Anagha Rahul; Chudhari, Lalit R.

doi:10.47974/JIM-2508

Open Access Research Article

Harmonic analysis in speech recognition systems

Amol D. Gaikwadamolgaikwad.ag@gmail.comDepartment of Information TechnologyYeshwantrao Chavan College of EngineeringNagpur, Maharashtra, 441110, IndiaView full profile → , ^*Haripriya H. KulkarniCorresponding authorharipriya.kulkarni@dypvp.edu.inDepartment of Electrical EngineeringPimpriDr. D. Y. Patil Institute of TechnologyPune, Maharashtra, 411018, IndiaView full profile → , Prashant Bansilal Patelprashant.patel@dypvp.edu.inDepartment of Instrumentation EngineeringPimpriDr. D.Y. Patil Institute of TechnologyPune, Maharashtra, 411018, IndiaView full profile → , Vidula Sanjay Japejape.swati@moderncoe.edu.inDepartment of Electrical EngineeringProgressive Education Society’sModern College of EngineeringPune, Maharashtra, 411005, IndiaView full profile → , Anagha Rahul Somananagha.soman@zealeducation.comDepartment of Electrical EngineeringNarheZeal College of Engineering and ResearchPune, Maharashtra, 411041, IndiaView full profile → , Lalit R. Chudharilalit.choudhari@dypvp.edu.inDepartment of Instrumentation EngineeringPimpriDr. D. Y. Patil Institute of TechnologyPune, Maharashtra, 411018, IndiaView full profile →

* Corresponding author · click or hover a name for details

pp. 719–727Vol. 29Issue 3March 2026DOI: 10.47974/JIM-2508XML

Received:: 01 Mar 2025
Published Online:: 18 Mar 2026
Article type:: Research Article
Language:: EN
Article no.:: JIM-2508
Pages:: 719–727

Abstract

This paper examines the way in which harmonic analysis can be applied to enhance the speech recognition systems. This is because speech patterns are nearly cyclic and therefore have basic frequencies and harmonics which reveal significant features of sound such as pitch, formants, as well as energy distribution. With the help of harmonic techniques of decomposing these signals, more precise and noise-free features can be extracted. Recognition is much more precise when harmonic analysis is used, as well as when audio models are built using neural networks, with 94.8 % success rate and 31 dB Signal-to-Noise Ratio. The experimental results indicate that harmonic-based models are more effective than the conventional ones like MFCC, LPC, and FFT.

Keywords

Harmonic analysis Speech recognition Neural networks Feature extraction Acoustic modeling Signal processing

Subject Classifications

68T1068T35

References

[1] D. Yin, C. Luo, Z. Xiong, and W. Zeng, “Phasen: A phase-and-harmonics-aware speech enhancement network,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pp. 9458–9465 (2020).
[2] T. Wang, W. Zhu, Y. Gao, S. Zhang, and J. Feng, “Harmonic attention for monaural speech enhancement,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 31, pp. 2424–2436 (2023).
[3] Y.-X. Lu, Y. Ai, and Z.-H. Ling, “MP-SENet: A speech enhancement model with parallel denoising of magnitude and phase spectra,” arXiv preprint, arXiv:2305.13686 (2023).
[4] P. Andreev, A. Alanov, O. Ivanov, and D. Vetrov, “HiFi++: A unified framework for neural vocoding, bandwidth extension and speech enhancement,” arXiv preprint, arXiv:2203.13086 (2022).
[5] S. Abdulatif, R. Cao, and B. Yang, “CMGAN: Conformer-based metric-GAN for monaural speech enhancement,” arXiv preprint, arXiv:2209.11112 (2022).
[6] M. N. Iqbal, L. Kütt, K. Daniel, N. Shabbir, A. Amjad, A. W. Awan, and M. Ali, “Inaccuracies and uncertainties for harmonic estimation in distribution networks,” Sustainability, vol. 16, pp. 6523 (2024).
[7] K. Daniel, L. Kütt, M. N. Iqbal, N. Shabbir, A. Ur Rehman, M. Shafiq, and H. Hamam, “Current harmonic aggregation cases for contemporary loads,” Energies, vol. 15, pp. 437 (2022).
[8] A. S. Shirkande, S. Salunkhe, R. Maral, and K. Dokhe, “AI-driven intelligent attendance system with advanced face recognition and cloud integration,” International Journal of Advanced Computer Engineering and Communication Technology (IJACECT), vol. 13, no. 2, pp. 5–9 (Mar. 2025).
[9] Z. Zecevic and B. Krstajic, “Dynamic harmonic phasor estimation by adaptive Taylor-based bandpass filter,” IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 6500509 (2021).
[10] L. Chen, W. Zhao, Q. Wang, F. Wang, and S. Huang, “Dynamic harmonic synchrophasor estimator based on sinc interpolation functions,” IEEE Transactions on Instrumentation and Measurement, vol. 68, pp. 3054–3065 (2019).
[11] A. Kumar C., C. K. Narayanappa, S. Bhargavi, S. Harish, R. Krishna, K. R. Varalakshmi, and S. Ashwini, “Nonlinear dynamics and chaos theory in engineering with applications in control systems and signal analysis,” Journal of Interdisciplinary Mathematics, vol. 28, no. 6, pp. 2099–2108 (2025), doi: 10.47974/JIM-2351.
[12] D. Motwani, V. Chitre, V. Bhosale, M. M. Nashipudmath, S. Shinde, and A. Nerurkar, “Implementing secure elliptic curve cryptography for blockchain-based financial transactions,” Journal of Discrete Mathematical Sciences and Cryptography, vol. 29, no. 2-A, pp. 617–627 (2026), doi: 10.47974/JDMSC-2504.
[13] S. Bhattacharya, L. Indise, N. Pandey, B. M. Nanche, S. Ramakrishna, and G. Sethi, “Analyzing the performance of wireless sensor networks using polynomial cryptography methods,” Journal of Discrete Mathematical Sciences and Cryptography, vol. 29, no. 2-A, pp. 563–570 (2026), doi: 10.47974/JDMSC-2497.

Views: 155Downloads: 135Citations: 0