In this study, new implementation techniques of a real-time MPEG-2 audio encoding system are presented. The system is developed using general-purpose DSPs. It consists of one master unit and five slave units, and its structure is basically based on our early work [1]. Two fast algorithms are developed and applied to the most compute-intensive routines of the encoding process. These algorithms play a key role in improving the entire system performance. The implemented system is designed to encode audio signal into MPEG-2 layer II bitstream with full configurations up to 5.1 channels and 640 kbps, and intended to support state-of-the-art quality. The generated bitstream can be stored in a hard disk on a PC or sent to an integration system to be multiplexed with the corresponding video-bitstream.

A speaker adaptation algorithm using formant frequencies is proposed. The formants extracted from the cepstral means in the reference codebook are iteratively shifted toward the formants of a test speaker. The number of cepstral means selected at each iteration decreases as the iteration increases. The decision of the number of selected cepstral means and a formant based distance measure are formulated. The proposed algorithm was implemented in two schemes and evaluated by speaker-independent, male speaker dependent, and female speaker-dependent recognition experiments. A combined scheme with the Bayesian adaptation obtained 9.7% enhancement for the average recognition accuracy in speaker-independent experiments and 52.6% in speaker-dependent recognition experiments.

A duration modeling scheme and a speaking rate compensation technique are presented for the HMM based connected digit recognizer. The proposed duration modeling technique uses a cumulative duration probability. The cumulative duration probability also can be used to obtain the duration bounds for the bounded duration modeling. One of the advantages of proposed technique is that the cumulative duration probability can be applied directly to the Viterbi decoding procedure without additional postprocessing. Therefore, it rules the state and word transition at each frame. To alleviate the problems due to fast or slow speech, a modification to the bounded duration modeling which accounts for speaking rate is described. The experimental results on Korean connected digit recognition show the effectiveness of the proposed duration modeling scheme and the speaking rate compensation technique.

In this paper, we describe a voice transformation meth-od which c hanges source speaker's acoustic features to those of a target speaker.The method developed here, acoustic features are divided into two parts, linear and nonlinear parts.Linear parts are characterized by LPC cepstrum coecients which are obtained from LP analysis.As for nonlinear part, which represent the excitation signal, is modelled by the long-delay nonlinear predictor using a neural net.Conversion rules for excitation signal are generated by the average pitch ratio and the mapping codebook, and those for LPC cepstrum coecients are based on the orthogonal vetctor space conversion.In addition, the spectral envelope compensation is proposed to correct spectral distortion in the transformed speech.A listening test shows that the proposed method makes it possible to convert speaker's individuality while maintaining high quality.

An adaptive multichannel digital filter structure that can be used in multichannel noise-cancelling applications is described. The proposed structure is a hybrid of the lattice and the escalator (Gram-Schmidt) structures, and the coefficients are updated using the least-mean-square algorithm. In the lattice-escalator hybrid structure, complete decorrelation of multiple reference signals is achieved by successively generating forward prediction error vectors via the lattice structure and decorrelating the elements of the error vectors using escalator linear combiners at each lattice stage. Simulation results indicate that the convergence speed of the hybrid structure is as fast as the escalator realization.>