The designing of transmultiplexer systems relies on determining filters for the transmitter and receiver sides of multicarrier communication system. The perfect reconstruction conditions lead to the bilinear equations for FIR filter coefficients. Generally there is no way of finding all possible solutions. This paper describes methods of finding a large family of solutions. Particular attention is devoted to obtaining algorithms useful in fixed-point arithmetic needed to design the integer filters. As a result, the systems perform perfect reconstruction of signals. Additionally, a simple method is presented to transform any transmultiplexer into an unlimited number of different transmultiplexers. Finally, two examples of integer filters that meet perfect reconstruction conditions are shown. The first illustrates a FIR filter which does not require multiplications. The frequency properties of filters and signals are discussed for the second example.
A phoneme segmentation method based on the analysis of discrete wavelet transform spectra is described. The localization of phoneme boundaries is particularly useful in speech recognition. It enables one to use more accurate acoustic models since the length of phonemes provide more information for parametrization. Our method relies on the values of power envelopes and their first derivatives for six frequency subbands. Specific scenarios that are typical for phoneme boundaries are searched for. Discrete times with such events are noted and graded using a distribution-like event function, which represent the change of the energy distribution in the frequency domain. The exact definition of this method is described in the paper. The final decision on localization of boundaries is taken by analysis of the event function. Boundaries are, therefore, extracted using information from all subbands. The method was developed on a small set of Polish hand segmented words and tested on another large corpus containing 16 425 utterances. A recall and precision measure specifically designed to measure the quality of speech segmentation was adapted by using fuzzy sets. From this, results with F-score equal to 72.49% were obtained.