NUTQ SIGNALINI XALAQITLARDAN TOZALASH USULLARI
Keywords:
Nutq signali, shovqinlarni tozalash, Soft Thresholding, Adaptive Smoothing, SNR, qisqa muddatli energiya (STE), pre-emphasis filtr.Abstract
This article considers the issue of cleaning speech signals from various noises.
Improving the quality and ensuring the accuracy of speech signals is relevant for modern voice
communication systems. The study analyzes the performance of the Soft Thresholding (SR) and
Adaptive Smoothing Soft Thresholding (ASST) algorithms. A comparison of the algorithms based
on SNR (signal-to-noise ratio) indicators is presented, and it is determined that the ASST
algorithm shows superior results. Also, information is provided about the importance of methods
such as Short-Term Energy (STE) and pre-emphasis filters.
References
Огнев И.В., Парамонов П.А. Предварительная обработка речевого сигнала для построения базы произношений одиночных слов // Информационные средства и технологии: труды Международной научно-технической конференции (20–22 октября 2012 г.): в 3 т. – М.: МЭИ, 2012. – Т. 1. – С. 53–58.
Sang-Sik A., Yoon-Chang L. An Improved Statistical Model-Based VAD Algorithm with an Adaptive Threshold // Journal of the Chinese Institute of Engineers. – Taipei, 2006. – Vol. 29, No. 5. – P. 783–789.
Зилинберг А.Ю., Корнеев Ю.А. Разработка и исследование временных и спектральных алгоритмов VAD (Voice Activity Detection) // Российская школа-конференция «Мобильные системы передачи данных». – Зеленоград: МИЭТ, 2006. – С. 58–70.
Mamatov N.S., Babomurodov O.J., Dusanov X.T. Automatic Speaker Recognition Through Voice Using Deep Neural Networks // Journal of Computer Science Engineering and Information Technology Research (JCSEITR). – 2024. – Vol. 14, Issue 1. – P. 13–20.
Dusanov X.T. The Issue of Recognizing a Person Based on His Voice // Computer Science and Engineering Technology: Collection of Materials of the International Scientific and Technical Conference. – Jizzakh: Jizzakh Branch of UzMU, 2023. – P. 140–143.