During the recording and transmiddion process of audio signals there are situation when the output signals is distorted from norse.The archive recordings which have very important for the art and history are also noisy distorted. The distortion, cracks and picks in the sound signals have very different nature and ifluence on the sound signals ...

This paper reviews different models of time frequency changes and frequency distribution of sound signals. The models, are applied to the representation of room acoustics quality; the restoration (denoising) of measured room impulse responses; artificial reverberation and distance effects in virtual auditory displays by audio signal processing systems.

In the paper are reviewed the disadvantages of a multi-layer neural network and possible ways of optimization of its structure with the purpose of eliminating the disadvantages. A distributed neural network is introduced, describing its structure and showing the advantages and disadvantages of such network compared to a standard multi-layer ...

In this paper is described the possibility of using distributed neural network in phonem recognition. An experimental system for vowels recogniton is described with the purpose of showing the advantages of the distributed network against the standard multilayer neural network. Some preprocessing techniques of the spectrum are examined, which ...

The object-oriented approach for the implementation of an experimental system for extraction of acoustic speech parameters is discussed in the article. The article stresses the building of hierarchy of classes which serve the software system, as superstructure of standard defined classes of Object Windows library (OWL) and ClassLib in the ...