Anadili Türkçe olan konuşmacılar tarafından söylenen NATO fonetik alfabesinin akustik analizi: El ve gırtlak mikrofonlarının karşılaştırması

Öz

Bu pilot çalışma, anadili Türkçe olan konuşmacıların NATO fonetik alfabesini telaffuz ederken el tipi ve boğaz mikrofonlarıyla kaydedilen konuşma sinyallerinin karşılaştırmalı akustik analizini sunmaktadır. Toplam 2.080 ses örneği toplanmış ve ses etkinliği tespiti (VAD), gürültü azaltma ve sessizlik kırpma adımlarını içeren bir ön işleme sürecinden geçirilmiştir. Süre, ortalama karekök (RMS) enerjisi, sıfır geçiş oranı (ZCR) ve saniye başına sıfır geçiş yoğunluğu gibi akustik ölçümler çıkarılmıştır. Non-parametrik testler, mikrofon türleri arasında anlamlı farklar olduğunu ortaya koymuştur: El tipi mikrofon daha yüksek enerji ve zamansal kararlılık sağlarken, boğaz mikrofonu daha düşük enerji ancak daha yüksek spektral karmaşıklık göstermiştir. Bazı konuşmacılarda ses üretiminde gecikme ve artikülasyon asimetrileri gözlemlenmiştir. Bulgular, mikrofon türünün fonetik yapı ve sinyal kalitesini etkilediğini ve gürültülü ortamlarda otomatik konuşma tanıma (ASR) sistemleri için önemli çıkarımlar sunduğunu göstermektedir. Bu çalışma, gerçek dünya koşullarına uygun çok kanallı ASR sistemleri için temel oluşturmaktadır.

Anahtar Kelimeler

• NATO fonetik alfabesi
• Gırtlakmikrofonu
• El tipi mikrofon
• Konuşma sinyali analizi
• ASRsağlamlığı.

Acoustic analysis of the NATO phonetic alphabet spoken by native Turkish speakers: A comparison of handheld and throat microphones

Öz

This pilot study presents a comparative acoustic analysis of speech signals recorded with handheld and throat microphones during the pronunciation of the NATO phonetic alphabet by native Turkish speakers. A total of 2,080 voice samples were collected and preprocessed using voice activity detection (VAD), noise reduction, and silence trimming. Acoustic metrics; duration, root mean square (RMS) energy, zero-crossing rate (ZCR), and zero-crossing density, were extracted. Non-parametric tests revealed significant differences between microphone types: the handheld device yielded higher energy and stability, while the throat microphone showed lower energy but higher spectral complexity. Results also indicated phonatory delay and articulation asymmetries in some speakers. The findings suggest that microphone type affects phonetic structure and signal quality, with implications for automatic speech recognition (ASR) systems in noisy environments. Notably, the throat microphone increased zero-crossing density due to internal vibration sensitivity. This study lays the groundwork for multichannel ASR systems designed for real-world acoustic variability.

Anahtar Kelimeler

• NATO phonetic alphabet
• Throat microphone
• Handheld microphone
• Speech signal analysis
• ASRrobustness

Kaynakça

1. I.C.A.O., Manual on the ICAO phonetic alphabet. ICAO Publishing, 2007.
2.     M.D. Keller, J.M. Ziriax, W. Barns, B. Sheffield, D. Brungart, T. Thomas, B. Jaeger, and K. Yankaskas, Performance in noise: impact of reduced speech intelligibility on sailor performance in a Navy command and control environment. Hearing Research, 349, 55–66, 2017. https://doi.org/10.1016/j.heares.201 6.10.007.
3.     P. Boersma and D. Weenink, Praat: doing phonetics by computer. University of Amsterdam, 2023.
4.     M.A.T. Turan, Enhancement of throat microphone recordings using Gaussian mixture model probabilistic estimator. arXiv:1804.05937, 2018. https://doi.org/ 10.48550/arXiv.1804.05937
5.     B.E. Acker-Mills, A.J. Houtsma, and W.A. Ahroon, Speech intelligibility in noise using throat and acoustic microphones. Aviation, Space and Environmental Medicine, 77 (1), 26–31, 2006.
6.     E. Erzin, Improving throat microphone speech recognition by joint analysis of throat and acoustic microphone recordings. IEEE Transactions on Audio, Speech and Language Processing, 17 (7), 1316–1324, 2009.
7.     M.E. Arafat, I. Misra, and M.E. Hamid, A comparative study for throat microphone speech enhancement with different approaches. International Journal of Science and Research Archive, 13 (1), 850–859, 2024. https:// doi.org/10.30574/ijsra.2024.13.1.1631
8.     X. Huang, A. Acero, and H.W. Hon, Spoken language processing: a guide to theory, algorithm, and system development. Pearson Prentice Hall, Upper Saddle River, 2001.

9.     Y. Kobayashi, K. Watanabe, and M. Akagi, Analysis and synthesis of speech captured by a contact microphone using neural vocoders. IEEE/ACM Transactions on Audio, Speech and Language Processing, 28, 2619–2632, 2020.
10.   T. Nguyen and S. Kim, An investigation of throat microphone signals for speech enhancement in noisy environments. Sensors, 21 (3), 987, 2021.
11.   R. Microphones, RØDE AI-1 Audio Interface – Specifications. 2022. Accessed 21 April 2025. https:// www.rode.com/interfaces/ai-1
12.   S. Watanabe, T. Hori, S. Karita, T. Hayashi, J. Nishitoba, Y. Unno, N.E. Yalta Soplin, J. Heymann, M. Wiesner, N. Chen, A. Renduchintala, and T. Ochiai, ESPnet: end-to-end speech processing toolkit. 2018.
13.   S. Karita, N. Chen, T. Hayashi, T. Hori, H. Inaguma, Z. Jiang, M. Someki, N.E. Yalta Soplin, R. Yamamoto, X. Wang, S. Watanabe, T. Yoshimura, and W. Zhang, A comparative study on Transformer vs RNN in speech applications. arXiv:1909.06317, 2019. https://doi.org /10.48550/arXiv.1909.06317
14.   E. Salesky, M. Wiesner, J. Bremerman, R. Cattoni, M. Negri, M. Turchi, D.W. Oard, and M. Post, The Multilingual TEDx Corpus for speech recognition and translation. arXiv:2102.01757, 2021. https://doi.org/ 10.48550/arXiv.2102.01757
15.   N. Wilkinson and T. Niesler, A hybrid CNN-BiLSTM voice activity detector. arXiv:2103.03529, 2021. https://doi.org/10.48550/arXiv.2103.03529
16.   J. Thomas, noisereduce: a Python package for real-time noise removal from speech signals. 2022. Accessed 2 May 2025. https://pypi.org/project/noisereduce/
17.   G. Ioannides and V. Rallis, Real-time speech enhancement using spectral subtraction with minimum statistics and spectral floor. arXiv:2302.10313, 2023. https://doi.org/10.48550/arXiv.2302.10313
18.   B. McFee, C. Raffel, D. Liang, D.P.W. Ellis, M. McVicar, E. Battenberg, and O. Nieto, librosa: audio and music signal analysis in Python. Proceedings of the 14th Python in Science Conference, 18–25, 2015. https://doi.org/10.25080/Majora-7b98e3ed-003
19.   K.A. Abdalmalak and A. Gallardo-Antolín, Enhancement of a text-independent speaker verification system by using feature combination and parallel-structure classifiers. arXiv:2401.15018, 2024. https://doi.org/10.48550/arXiv.2401.15018
20.   S. Verma, R. Banerjee, and A. Singh, A comprehensive review of preprocessing strategies in speech recognition systems. ACM Transactions on Speech and Language Processing, 19 (1), 1–26, 2022.
21.   G. Degottex, P. Lanchantin, M.J. Gales, and S. King, A survey on acoustic representations for voice analysis and synthesis. IEEE/ACM Transactions on Audio, Speech and Language Processing, 29, 116–137, 2021.
22.   J.M. Valin and J. Skoglund, A real-time wideband neural vocoder at 1.6 kb/s using LPCNet. Proceedings of Interspeech, 3406–3410, 2019. https://doi.org/10. 21437/Interspeech.2019-1255
23.   A. Saeed, D. Grangier, and N. Zeghidour, Contrastive learning of general-purpose audio representations. arXiv:2010.10915, 2021. https://doi.org/10.48550/arX iv.2010.10915
24.   W. Wang, J. Yi, M. Wu, and X. Lei, Improving ASR robustness via uncertainty modeling and consistency training. Proceedings of Interspeech, 2022.
25.   Laerd Statistics, Mann–Whitney U Test using SPSS Statistics. 2021. Accessed 2 May 2025. https://stat istics.laerd.com/spss-tutorials/mann-whitne y-u-test-using-spss-statistics.php
26.   Y. Zhao, C. Ni, C.C. Leung, S. Joty, E.S. Chng, and B. Ma, A unified speaker adaptation approach for ASR. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 9398–9410, 2021. https://doi.org/10.18653/v1/2021.emnlp-main.737
27.   Y. Zhang, C. Wu, and Y. Wang, Acoustic environment and recording conditions for voice applications: a comparative study. IEEE Access, 8, 213187–213200, 2020.
28.   J.X. Ou, X. Ming, and A.C.L. Yu, Individual variability in subcortical neural encoding shapes phonetic cue weighting. Scientific Reports, 13 (1), 9991, 2023. https://doi.org/10.1038/s41598-023-37212-y
29.   D. Prabhu, P. Jyothi, S. Ganapathy, and V. Unni, Accented speech recognition with accent-specific codebooks. arXiv:2310.15970, 2023. https://doi.org/ 10.48550/arXiv.2310.15970
30.   M. McAuliffe, M. Socolof, S. Mihuc, M. Wagner, and M. Sonderegger, Montreal Forced Aligner: trainable text–speech alignment using Kaldi. Proceedings of Interspeech 2017, 498–502, 2017. https://doi.org/10.21 437/Interspeech.2017-1386
31.   H. Kallio, S. Antti, and J. Šimko, Fluency-related temporal features and syllable prominence as prosodic proficiency predictors for learners of English with different language backgrounds. Language and Speech, 65 (3), 571–597, 2022. https://doi.org/10.1177/0023 8309211040175
32.   O. Türk and H. Açıkgoz, Phonological influence of Turkish on English pronunciation: a comparative spectrographic study. Journal of Phonetics and Speech Sciences, 14, 157–170, 2022.
33.   K.J. Han, R. Prieto, K. Wu, and T. Ma, State-of-the-art speech recognition using multi-stream self-attention with dilated 1D convolutions. arXiv:1910.00716, 2019. https://doi.org/10.48550/arXiv.1910.00716
34.   K. Tomanek, V. Zayats, D. Padfield, K. Vaillancourt, and F. Biadsy, Residual adapters for parameter-efficient ASR adaptation to atypical and accented speech. arXiv:2109.06952, 2021. https://doi.org/10. 48550/arXiv.2109.06952
35.   A. Radford, J.W. Kim, T. Xu, G. Brockman, C. McLeavey, and I. Sutskever, Robust speech recognition via large-scale weak supervision. arXiv:2212.04356, 2023. https://doi.org/10.48550/arX iv.2212.04356