This paper describes a BIST circuit for testing SoC integrated multi-channel serializer/deserializer (SerDes) macros. A newly developed packet-based PRBS generator enables the BIST to perform at-speed testing of asynchronous data transfers. In addition, a new technique for chained alignment checks between adjacent channels helps achieve a channel-count-independent architecture for verification of multi-channel alignment between SerDes macros. Fabricated in a 0.13-µm CMOS process and operating at > 500 MHz, the BIST has successfully verified all SerDes functions in at-speed testing of 5-Gbps20-ch SerDes macros.

In this paper, we propose a distributed speaker recognition method using a nonparametric speaker model and Earth Mover's Distance (EMD). In distributed speaker recognition, the quantized feature vectors are sent to a server. The Gaussian mixture model (GMM), the traditional method used for speaker recognition, is trained using the maximum likelihood approach. However, it is difficult to fit continuous density functions to quantized data. To overcome this problem, the proposed method represents each speaker model with a speaker-dependent VQ code histogram designed by registered feature vectors and directly calculates the distance between the histograms of speaker models and testing quantized feature vectors. To measure the distance between each speaker model and testing data, we use EMD which can calculate the distance between histograms with different bins. We conducted text-independent speaker identification experiments using the proposed method. Compared to results using the traditional GMM, the proposed method yielded relative error reductions of 32% for quantized data.

We presented a new text-independent/text-prompted speaker recognition method by combining speaker-specific Gaussian Mixture Model (GMM) with syllable-based HMM adapted by MLLR or MAP. The robustness of this speaker recognition method for speaking style's change was evaluated in this paper. The speaker identification experiment using NTT database which consists of sentences data uttered at three speed modes (normal, fast and slow) by 35 Japanese speakers (22 males and 13 females) on five sessions over ten months was conducted. Each speaker uttered only 5 training utterances (about 20 seconds in total). A combination method reduced the identification error rate by about 50%. We obtained the accuracy of 98.8% for text-independent speaker identification for three speaking style modes (normal, fast, slow) by using a short test utterance (about 4 seconds). Especially, we obtained the accuracy of 99.4% for normal speaking mode. This result was superior to conventional methods for the same database. We show that the attractive result was brought from the compensational effect between speaker specific GMM and speaker adapted syllable based HMM.

A discriminative modeling is applied to optimize the structure of a Partly-Hidden Markov Model (PHMM). PHMM was proposed in our previous work to deal with the complicated temporal changes of acoustic features. It can represent observation dependent behaviors in both observations and state transitions. In the formulation of the previous PHMM, we used a common structure for all models. However, it is expected that the optimal structure which gives the best performance differs from category to category. In this paper, we designed a new structure optimization method in which the dependence of the states and the observations of PHMM are optimally defined according to each model using the weighted likelihood-ratio maximization (WLRM) criterion. The WLRM criterion gives high discriminability between the correct category and the incorrect categories. Therefore it gives model structures with good discriminative performance. We define the model structure combination which satisfy the WLRM criterion for any possible structure combinations as the optimal structures. A genetic algorithm is also applied to the adequate approximation of a full search. With results of continuous lecture talk speech recognition, the effectiveness of the proposed structure optimization is shown: it reduced the word errors compared to HMM and PHMM with a common structure for all models.

We address issues for improving hands-free speech enhancement and speech recognition performance in different car environments using a single distant microphone. This paper describes a new single-channel in-car speech enhancement method that estimates the log spectra of speech at a close-talking microphone based on the nonlinear regression of the log spectra of noisy signal captured by a distant microphone and the estimated noise. The proposed method provides significant overall quality improvements in our subjective evaluation on the regression-enhanced speech, and performed best in most objective measures. Based on our isolated word recognition experiments conducted under 15 real car environments, the proposed adaptive nonlinear regression approach shows an advantage in average relative word error rate (WER) reductions of 50.8% and 13.1%, respectively, compared to original noisy speech and ETSI advanced front-end (ETSI ES 202 050).

In this paper, we provide a practical formulation of the problem of identifying all error occurrences and all failed scan cells in at-speed scan based BIST environment. We propose a method that can be used to identify every error when the circuit test frequency is higher than the tester frequency. Our approach requires very little extra hardware for diagnosis and the test application time required to identify errors is a linear function of the frequency ratio between the CUT and the tester.

In this paper, we describe a parallel decoding-based ASR system developed of ATR that is robust to noise type, SNR and speaking style. It is difficult to recognize speech affected by various factors, especially when an ASR system contains only a single acoustic model. One solution is to employ multiple acoustic models, one model for each different condition. Even though the robustness of each acoustic model is limited, the whole ASR system can handle various conditions appropriately. In our system, there are two recognition sub-systems which use different features such as MFCC and Differential MFCC (DMFCC). Each sub-system has several acoustic models depending on SNR, speaker gender and speaking style, and during recognition each acoustic model is adapted by fast noise adaptation. From each sub-system, one hypothesis is selected based on posterior probability. The final recognition result is obtained by combining the best hypotheses from the two sub-systems. On the AURORA-2J task used widely for the evaluation of noise robustness, our system achieved higher recognition performance than a system which contains only a single model. Also, our system was tested using normal and hyper-articulated speech contaminated by several background noises, and exhibited high robustness to noise and speaking styles.

Since their inception almost fifty years ago, hidden Markov models (HMMs) have have become the predominant methodology for automatic speech recognition (ASR) systems--today, most state-of-the-art speech systems are HMM-based. There have been a number of ways to explain HMMs and to list their capabilities, each of these ways having both advantages and disadvantages. In an effort to better understand what HMMs can do, this tutorial article analyzes HMMs by exploring a definition of HMMs in terms of random variables and conditional independence assumptions. We prefer this definition as it allows us to reason more throughly about the capabilities of HMMs. In particular, it is possible to deduce that there are, in theory at least, no limitations to the class of probability distributions representable by HMMs. This paper concludes that, in search of a model to supersede the HMM (say for ASR), rather than trying to correct for HMM limitations in the general case, new models should be found based on their potential for better parsimony, computational requirements, and noise insensitivity.

A comparison of performances is made of three text-independent speaker identification methods based on dual Penalized Logistic Regression Machine (dPLRM), Support Vector Machine (SVM) and Gaussian Mixture Model (GMM) with experiments by 10 male speakers. The methods are compared for the speech data which were collected over the period of 13 months in 6 utterance-sessions of which the earlier 3 sessions were for obtaining training data of 12 seconds' utterances. Comparisons are made with the Mel-frequency cepstrum (MFC) data versus the log-power spectrum data and also with training data in a single session versus in plural ones. It is shown that dPLRM with the log-power spectrum data is competitive with SVM and GMM methods with MFC data, when trained for the combined data collected in the earlier three sessions. dPLRM outperforms GMM method especially as the amount of training data becomes smaller. Some of these findings have been already reported in [1]-[3].

This paper proposes a technique for synthesizing speech with a desired speaking style and/or emotional expression, based on model adaptation in an HMM-based speech synthesis framework. Speaking styles and emotional expressions are characterized by many segmental and suprasegmental features in both spectral and prosodic features. Therefore, it is essential to take account of these features in the model adaptation. The proposed technique called style adaptation, deals with this issue. Firstly, the maximum likelihood linear regression (MLLR) algorithm, based on a framework of hidden semi-Markov model (HSMM) is presented to provide a mathematically rigorous and robust adaptation of state duration and to adapt both the spectral and prosodic features. Then, a novel tying method for the regression matrices of the MLLR algorithm is also presented to allow the incorporation of both the segmental and suprasegmental speech features into the style adaptation. The proposed tying method uses regression class trees with contextual information. From the results of several subjective tests, we show that these techniques can perform style adaptation while maintaining naturalness of the synthetic speech.

This paper describes a method of generating F0 contours from natural F0 segmental shapes for speech synthesis. The extracted shapes of the F0 units are basically held invariant by eliminating any averaging operations in the analysis phase and by minimizing modification operations in the synthesis phase. The use of natural F0 shapes has great potential to cover a wide variety of speaking styles with the same framework, including not only read-aloud speech, but also dialogues and emotional speech. A linear-regression statistical model is used to "manipulate" the stored raw F0 shapes to build them up into a sentential F0 contour. Through experimental evaluations, the proposed model is shown to provide stable and robust F0 contour prediction for various speakers. By using this model, linguistically derived information about a sentence can be directly mapped, in a purely data-driven manner, to acoustic F0 values of the sentential intonation contour for a given target speaker.

Conventional confidence measures for assessing the reliability of ASR (automatic speech recognition) output are typically derived from "low-level" information which is obtained during speech recognition decoding. In contrast to these approaches, we propose a novel utterance verification framework which incorporates "high-level" knowledge sources. Specifically, we investigate two application-independent measures: in-domain confidence, the degree of match between the input utterance and the application domain of the back-end system, and discourse coherence, the consistency between consecutive utterances in a dialogue session. A joint confidence score is generated by combining these two measures with an orthodox measure based on GPP (generalized posterior probability). The proposed framework was evaluated on an utterance verification task for spontaneous dialogue performed via a (English/Japanese) speech-to-speech translation system. Incorporating the two proposed measures significantly improved utterance verification accuracy compared to using GPP alone, realizing reductions in CER (confidence error-rate) of 11.4% and 8.1% for the English and Japanese sides, respectively. When negligible ASR errors (that do not affect translation) were ignored, further improvement was achieved for the English side, realizing a reduction in CER of up to 14.6% compared to the GPP case.

This study shows the effectiveness of using gamma distribution in the speech power domain as a more general prior distribution for the model-based speech enhancement approaches. This model is a super-set of the conventional Gaussian model of the complex spectrum and provides more accurate prior modeling when the optimal parameters are estimated. We develop a method to adapt the modeled distribution parameters from each actual noisy speech in a frame-by-frame manner. Next, we derive and investigate the minimum mean square error (MMSE) and maximum a posterior probability (MAP) estimations in different domains of speech spectral magnitude, generalized power and its logarithm, using the proposed gamma modeling. Finally, a comparative evaluation of the MAP and MMSE filters is conducted. As the MMSE estimations tend to more complicated using more general prior distributions, the MAP estimations are given in closed-form extractions and therefore are suitable in the implementation. The adaptive estimation of the modeled distribution parameters provides more accurate prior modeling and this is the principal merit of the proposed method and the reason for the better performance. From the experiments, the MAP estimation is recommended due to its high efficiency and low complexity. Among the MAP based systems, the estimation in log-magnitude domain is shown to be the best for the speech recognition as the estimation in power domain is superior for the noise reduction.

This paper describes a hands-free speech recognition technique based on acoustic model adaptation to reverberant speech. In hands-free speech recognition, the recognition accuracy is degraded by reverberation, since each segment of speech is affected by the reflection energy of the preceding segment. To compensate for the reflection signal we introduce a frame-by-frame adaptation method adding the reflection signal to the means of the acoustic model. The reflection signal is approximated by a first-order linear prediction from the observation signal at the preceding frame, and the linear prediction coefficient is estimated with a maximum likelihood method by using the EM algorithm, which maximizes the likelihood of the adaptation data. Its effectiveness is confirmed by word recognition experiments on reverberant speech.

We present a novel trigger-based language model adaptation method oriented to the transcription of meetings. In meetings, the topic is focused and consistent throughout the whole session, therefore keywords can be correlated over long distances. The trigger-based language model is designed to capture such long-distance dependencies, but it is typically constructed from a large corpus, which is usually too general to derive task-dependent trigger pairs. In the proposed method, we make use of the initial speech recognition results to extract task-dependent trigger pairs and to estimate their statistics. Moreover, we introduce a back-off scheme that also exploits the statistics estimated from a large corpus. The proposed model reduced the test-set perplexity considerably more than the typical trigger-based language model constructed from a large corpus, and achieved a remarkable perplexity reduction of 44% over the baseline when combined with an adapted trigram language model. In addition, a reduction in word error rate was obtained when using the proposed language model to rescore word graphs.

As computer hardware components are achieving greater speeds, network link bandwidths are becoming wider. A number of enhancements to TCP have been developed in order to fully exploit these improvements in network infrastructures, including TCP window scale option, SACK option, and HighSpeed TCP (HSTCP) modifications. However, even with these enhancements, TCP cannot provide satisfactory performance in high-speed long-delay networks. As a means addressing this problem, gentle HighSpeed TCP (gHSTCP) has been proposed in [1]. However, its effectiveness has only been demonstrated in simulation experiments. In the present paper, a refined gHSTCP algorithm is proposed for application to real networks. The performance of the refined gHSTCP algorithm is then assessed experimentally. The refined gHSTCP algorithm is based on the original algorithm, which uses two modes (Reno mode and HSTCP mode) in the congestion avoidance phase and switches modes based on RTT increasing trends. The refined gHSTCP algorithm compares two RTT thresholds and judges which mode will be used. The performance of gHSTCP is compared with TCP Reno/HSTCP and parallel TCP mechanisms. The experimental results demonstrate that gHSTCP can provide a better tradeoff in terms of utilization and fairness against co-existing traditional TCP Reno connections, whereas HSTCP and parallel TCP suffer from the trade-off problem.

This paper proposes a new compensation method of acoustic scores in the Viterbi search for robust speech recognition. This method introduces noise models to represent a wide variety of noises and realizes robust decoding together with conventional techniques of subtraction and adaptation. This method uses likelihoods of noise models in two ways. One is to calculate a confidence factor for each input frame by comparing likelihoods of speech models and noise models. Then the weight of the acoustic score for a noisy frame is reduced according to the value of the confidence factor for compensation. The other is to use the likelihood of noise model as an alternative that of a silence model when given noisy input. Since a lower confidence factor compresses acoustic scores, the decoder rather relies on language scores and keeps more hypotheses within a fixed search depth for a noisy frame. An experiment using commentary transcriptions of a broadcast sports program (MLB: Major League Baseball) showed that the proposed method obtained a 6.7% relative word error reduction. The method also reduced the relative error rate of key words by 17.9%, and this is expected lead to an improvement metadata extraction accuracy.

The paper provides an overview of the circuit techniques for CMOS high-speed I/Os, focusing on the design issues in sub-100 nm standard CMOS. First, we describe the evolution of CMOS high-speed I/O since it appeared in mid 90's. In our view, the surge in the I/O bandwidth we experienced from the mid 90's to the present was driven by the continuous improvement of the CMOS IC performance. As a result, CMOS high-speed I/O has covered the data rate ranging from 2.5 Gb/s to 10 Gb/s, and now is heading for 40 Gb/s and beyond. To meet the speed requirements, an optimum choice of the transceiver architecture and its building blocks are crucial. We pick the most critical building blocks such as the decision circuit and the multiplexors and give detailed explanation of their designs. We describe the low-voltage operation of the high-speed I/O in view of reducing the power consumption. An example of a 90-nm CMOS 2.5 Gb/s transceiver operating off a 0.8 V power supply will be described. Operability at 0.8 V ensures that the circuits will not become obsolescent, even below the 60 nm process node.

There has been significant interest in developing new forms of acoustic model, in particular models which allow additional dependencies to be represented than those contained within a standard hidden Markov model (HMM). This paper discusses one such class of models, augmented statistical models. Here, a local exponential approximation is made about some point on a base model. This allows additional dependencies within the data to be modelled than are represented in the base distribution. Augmented models based on Gaussian mixture models (GMMs) and HMMs are briefly described. These augmented models are then related to generative kernels, one approach used for allowing support vector machines (SVMs) to be applied to variable length data. The training of augmented statistical models within an SVM, generative kernel, framework is then discussed. This may be viewed as using maximum margin training to estimate statistical models. Augmented Gaussian mixture models are then evaluated using rescoring on a large vocabulary speech recognition task.

This paper addresses a speech recognition problem in non-stationary noise environments: the estimation of noise sequences. To solve this problem, we present a particle filter-based sequential noise estimation method for front-end processing of speech recognition in noise. In the proposed method, a noise sequence is estimated in three stages: a sequential importance sampling step, a residual resampling step, and finally a Markov chain Monte Carlo step with Metropolis-Hastings sampling. The estimated noise sequence is used in the MMSE-based clean speech estimation. We also introduce Polyak averaging and feedback into a state transition process for particle filtering. In the evaluation results, we observed that the proposed method improves speech recognition accuracy in the results of non-stationary noise environments a noise compensation method with stationary noise assumptions.