Low-Rate Speech Coding Using Radial Neural Networks
Ekaterina A. Lopukhova, Grigory S. Voronkov, Igor V. Kuznetsov · 2021 International Conference on Information Technology and Nanotechnology (ITNT) · 2021
This article considers methods of low-speed speech coding using neural networks for communication and voice storage applications. It analyzes the use of radial neural networks to solve the problems of low-speed signal coding in multichannel systems of voice information transmission by reducing its dynamic range and increasing the energy efficiency of systems without deteriorating the quality of communication. It describes the construction of a codec model using radial neural networks as a compressor and decompressor for multichannel systems in MatLab and evaluates the codec delay for signal processing depending on the number of channels in the communication system.