A systematic approach to the modelling of multicellular power transistors is presented. It is based on characterisation and identification (modelling) of the elementary cell of the transistor, and modelling of input a...
详细信息
A systematic approach to the modelling of multicellular power transistors is presented. It is based on characterisation and identification (modelling) of the elementary cell of the transistor, and modelling of input and output interconnections using the physical mask of the device. This model enables the small-signal performance of heterojunction bipolar transistors (HBT) to be predicted precisely. The obtained multicell model has been validated by comparison of its predictions with experimental measurements on various HBTs having different numbers of elementary cells.
In this paper extensions to the analysis-by-synthesis (AbS) loop used in code excited linearpredictive (CELP) speech codecs are considered. Methods for updating the short-term synthesis filter once the excitation par...
详细信息
In this paper extensions to the analysis-by-synthesis (AbS) loop used in code excited linearpredictive (CELP) speech codecs are considered. Methods for updating the short-term synthesis filter once the excitation parameters have been determined are examined. We show that significant improvements can be achieved by updating the synthesis filter, similar to those obtained using the well known methods of interpolation and bandwidth expansion. However our proposed method of update avoids the increase in the delay of a codec that is usually associated with interpolation. Furthermore the traditional sequential method of determining the adaptive and fixed codebook parameters is examined and compared to an exhaustive search of both codebooks. Three sub-optimum techniques are proposed for improving the performance of the codebook search while maintaining a reasonable level of complexity. The most complex of these increases the codec complexity by only about 40% but provides 80% of the maximum possible 1.1 dB segmental SNR improvement associated with an exhaustive codebook search.
This paper describes an isolated-word, speaker independent, speech recognition system based on neural networks. The system was developed for a 32 words vocabulary including the digits, movement commands and arithmetic...
详细信息
This paper describes an isolated-word, speaker independent, speech recognition system based on neural networks. The system was developed for a 32 words vocabulary including the digits, movement commands and arithmetic operation commands. The recognition rate achieved for a directly recorded speech, band-limited to 3.4 kHz, was greater than 96%.
An efficient coding scheme for linear predictive coding (LPC) residuals is proposed based on harmonic and noise representation. New features of the scheme include classified vector quantization of the spectral envelop...
详细信息
An efficient coding scheme for linear predictive coding (LPC) residuals is proposed based on harmonic and noise representation. New features of the scheme include classified vector quantization of the spectral envelope of LPC residuals with a weighted distortion measure. The improvement in performance obtained by classifying codebooks based on a voiced/unvoiced (V/UV) decision is shown. Sequences of the short-term RMS power of the time domain waveforms are also vector quantized and transmitted for unvoiced signals. A fast synthesis algorithm for voiced signals using an FFT is also presented, which reduces the high complexity of the direct sinusoidal synthesis method with interpolated magnitudes and phases. Informal listening tests indicate that, in combination with a known LSP quantization technique, this residual coding scheme provides good communication quality at a total bit rate of less than 2.0 kbps.
Variable rate quantization of the linear predictive coding (LPC) parameters based on phonetic classification of the speech frame results in substantial performance gain. Speech frames are classified as unvoiced or voi...
详细信息
Variable rate quantization of the linear predictive coding (LPC) parameters based on phonetic classification of the speech frame results in substantial performance gain. Speech frames are classified as unvoiced or voiced and are separately quantized with VQ codebooks designed for each class. Performance results, including listening tests, show that for transparent quality roughly 9 bits is sufficient for unvoiced frames and 24 bits for voiced frames. Test results of LPC quantization are described for a variable rate phonetically segmented CELP coder and for the synthesis of speech from the prediction residual.
Packet reservation multiple access (PRMA) assisted adaptive modulation using 1, 2 and 4 bit/symbol transmissions is proposed as an alternative to dynamic channel allocation (DCA) in order to maximise the number of use...
详细信息
Packet reservation multiple access (PRMA) assisted adaptive modulation using 1, 2 and 4 bit/symbol transmissions is proposed as an alternative to dynamic channel allocation (DCA) in order to maximise the number of users supported in a traffic cell. The cell is divided in three concentric rings and in the central high signal-to-noise ratio (SNR) region 16-level star quadrature amplitude modulation (16-StQAM) is used, in the first ring differential quaternary phase shift keying (DQPSK) is invoked, while in the outer ring differential phase shift keying (DPSK) is utilised. A channel SNR of about 7, 10 and 20 dB, respectively, was required in order to maintain a bit error ratio (BER) of about 1%, which can then be rendered error-free by the binary BCH error correction codes used. A 4.7 kbps algebraic code excited linearpredictive (ACELP) speech codec is favoured, which is protected by a quad-class source-sensitivity matched BCH coding scheme, yielding a total bit rate of 8.4 kbps. A GSM-like voice activity detector (VAD) controls the PRMA-assisted adaptive system, which ensures a capacity improvement of a factor of 1.78 over PRMA-aided binary schemes.
This paper argues that many HMM model inaccuracies are a direct consequence of the fact that the HMM is a one dimensional stochastic model applied to a two dimensional process. Thus we argue that a 2D stochastic proce...
详细信息
This paper argues that many HMM model inaccuracies are a direct consequence of the fact that the HMM is a one dimensional stochastic model applied to a two dimensional process. Thus we argue that a 2D stochastic process, known as a Markov random field (MRF) should perform better. We describe a training method for MRFs and analyze its convergence behavior.
Efficient block coding methods for LPC information play an essential role in very-low-rate speech coding systems. The subject of this contribution is a new suboptimal matrix quantization scheme for LPC parameters, cal...
详细信息
Efficient block coding methods for LPC information play an essential role in very-low-rate speech coding systems. The subject of this contribution is a new suboptimal matrix quantization scheme for LPC parameters, called matrix product quantization (MPQ), which operates at bit rates between 300 and 700 b/s. MPQ encodes sequences of LPC parameter vectors using a product formulation of two matrices which describe the average parameter vector and the temporal contour. In fixed-rate coding systems for mobile communication, MPQ achieves a very high coding efficiency at a low coding delay. Compared to the multi-frame coding method (MFC) of Kemp et al. (1991), which causes a delay of 8 frames, the MPQ scheme operates more efficiently even at a coding delay of only 3 frames. Applying MPQ to a variable-rate segment vocoder, a bit rate reduction of 50% compared to memoryless VQ is obtained at a frame period of 20 ms.
A dedicated LPC speech synthesizer is presented in this paper. It mainly acts as a programmable 12-pole lattice filter with a 2-pole low-pass filter, a 12 bit D/A converter and a built-in interface to external CPU. Th...
详细信息
ISBN:
(纸本)0780330625
A dedicated LPC speech synthesizer is presented in this paper. It mainly acts as a programmable 12-pole lattice filter with a 2-pole low-pass filter, a 12 bit D/A converter and a built-in interface to external CPU. The gate level logic simulation and the results of c programs shows that the synthesizer can synthesize high-quality speech at low data rate. It can be used in LPC analyze-synthesize systems and developments of LPC speech synthesizer IC.
暂无评论