Browsing by Author "Nein, HW"

Now showing 1 - 2 of 2

Incorporating error shaping technique into LSF vector quantization
(2001-02-01)
Nein, HW
;
Lin, CT
;
電控工程研究所
;
Institute of Electrical and Control Engineering
This paper presents an error shaping technique for line spectrum frequency (LSF) vector quantization, The error shaping technique based on the weighted logarithm spectral distortion (WLSD) measure can be used for shaping the spectral distortion distribution of quantization error into any different curve depending on what kind of weighting function is used. However, the high computational complexity of the WLSD measure deters this error shaping technique from practical use. To solve this problem, we approximate the WLSD measure by the quadratically weighted measure or the weighted mean squared error (WMSE) measure and propose an optimal error shaping technique of LSF vector quantization. In this proposed error shaping technique, the optimal WMSE weights (i.e., the optimal weights of LSF parameters) are determined based on the theoretical analysis of the WLSD measure. Three experiments are performed to check the performance of the proposed error shaping technique. One experiment is set up by incorporating human perception into the LSF quantization and another is set up by emphasizing the human-sensitivity frequency band in lower frequency bandwidth 0-3 kHz. In the third experiment, we apply the proposed error shaping technique to the LSF quantization of a CELP coder to test how it affects the overall speech quality in an actual speech coding algorithm.
Speaker adaptation of fuzzy-perceptron-based speech recognition
(1999-02-01)
Lin, CT
;
Nein, HW
;
Lin, WF
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, we propose a speech recognition algorithm which utilizes hidden Markov models (HMM) and Viterbi algorithm for segmenting the input speech sequence, such that the variable-dimensional speech signal is converted into a fixed-dimensional speech signal, called TN vector. We then use the fuzzy perceptron to generate hyperplanes which separate patterns of each class from the others. The proposed speech recognition algorithm is easy for speaker adaptation when the idea of "supporting pattern" is used. The supporting patterns are those patterns closest to the hyperplane. When a recognition error occurs, we include all the TN vectors of the input speech sequence with respect to the segmentations of all HMM models as the supporting patterns. The supporting patterns are then used by the fuzzy perceptron to tune the hyperplane that can cause correct recognition, and also tune the hyperplane that resulted in wrong recognition. Since only two hyperplanes need to be tuned for a recognition error, the proposed adaptation scheme is time-economic and suitable for on-line adaptation. Although the adaptation scheme cannot ensure to correct the wrong recognition right after adaptation, the hyperplanes are tuned in the direction for correct recognition iteratively and the speed of adaptation can be adjusted by a "belief" parameter set by the user. Several examples are used to show the performance of the proposed speech recognition algorithm and the speaker adaptation scheme.