الفهرس 
Speech CodingOnly 14 pages are availabe for public view
 |  | from | 32 |  |  |
| | | from | 32 | | |
Abstract
Development of new Speech coding techniques and bit rate reduction is an important goal of research in mobile communication. Bit rate reduction while maintaining the same speech quality is a challenging task for achievement.
This coder is based on exploiting the characteristics of different parts of speech. In this thesis, three main points are studied.
The first point discusses the Introduction of Wavelet Transform, which is really of great help for mobile communications speech coders since an obvious reduction in the average bit rate was achieved while maintaining the required speech quality. Also decreasing the processing time was achieved, since wavelet transform was used in some signal processing parts computations like the energy calculations and the rate of zero crossings; these two quantities are used for making the decision of speech frame type (voiced, unvoiced, transition or silence).
The second point is introducing the variable bit rate encoder, since each type of frame is given a certain bit rate, resulting in four different bit rates.
The third point is using a new technique in the reconstruction of the speech frame in the decoder, since reconstruction can be done by some of the wavelet coefficients and not all of them.
The applied new technique revealed a better speech coder with lower bit rate with acceptable MOS. In addition that the wavelet transform is much simpler in implementation and faster in processing.