JBT Journal
Advanced

Raqamli Transformatsiya va Sun’iy Intellekt

NUTQ SIGNALIDAN SHOVQIN VA XALAQITLARDAN TOZALASH MUAMMOLARI

Dusanov Xurshid Toshpulotovich1, Mullaboyev Og‘abek Axmadjon o‘g‘li1

1 Muxammad al-Xorazmiy nomidagi Toshkent axborot texnologiyalari universiteti

Zenodo DOI: 10.5281/zenodo.20556353

Keywords

oq shovqin impulsli shovqin xalaqitni kamaytirish aks-sado signal sifati O‘zgaruvchan shovqin Nutq signali MOS binar niqob Doimiy shovqin

Abstract

Ushbu maqolada nutq signallaridan shovqin va boshqa buzuvchi tovushlarni ajratib olish muammolari, ularning turlari hamda zamonaviy tozalash usullari tahlil qilingan. Audio signallarda uchraydigan statsionar va nostatsionar shovqinlar, aks-sado (reverb), impulsli shovqinlar, fon shovqini hamda elektr interferensiyasining xususiyatlari va ularni kamaytirish algoritmlari o‘rganilgan. Shuningdek, klassik signal qayta ishlash usullari (Wiener filtri, spectral subtraction, notch filtr) va zamonaviy sun’iy intellektga asoslangan modellar (DeepFilterNet2, RNNoise, neyron tarmoqlar)ning samaradorligi taqqoslangan. Tadqiqotda real vaqt tizimlari uchun mos algoritmlarni tanlash muammolari, shovqin va nutq spektrlarining ustma-ust tushishi hamda artefaktlar yuzaga kelish sabablari ham ko‘rib chiqilgan. Natijalar audio signallarni tozalashda gibrid va adaptiv yondashuvlar yuqori samaradorlikka ega ekanligini ko‘rsatadi.

Views: 104 Published: 2026-06-05 Pages: 53-56 Volume: 4 Issue: 3

References

  1. Jurafsky D., Martin J. Speech and Language Processing. — Pearson, 2021.
  2. Valin J.-M. “A Hybrid DSP/Deep Learning Approach to Real-Time Full-Band Speech Enhancement.” // Proceedings of MMSP, 2018.
  3. N.S. Mamatov, X.T. Dusanov Nutq signallari belgilar toplamini shakllantirishning mfcc usuli Современные инновационные исследования актуальные проблемы и развитие тенденции: решения и перспективы 179-182 bet.
  4. N.S.Mamatov, O.J.Babomuradov, X.T. Dusanov Automatic speaker recognition through voice using deep neural networks. Journal of Computer Science Engineering and Information Technology Research (JCSEITR) ISSN (Print): 2250-2416; ISSN (Online): Applied Vol. 14, Issue 1; Jun 30 2024, 13-20
  5. Schröter H., Westhausen N. DeepFilterNet2: Towards Real-Time Speech Enhancement for Full-Band Audio. — 2022.
  6. Vaseghi S. Advanced Signal Processing and Digital Noise Reduction. — Wiley, 2008.