This paper describes a new multi-channel method of noisy speech recognition, which estimates the log spectrum of speech at a close-talking microphone based on the multiple regression of the log spectra (MRLS) of noisy signals captured by distributed microphones. The advantages of the proposed method are as follows: 1) The method does not require a sensitive geometric layout, calibration of the sensors nor additional pre-processing for tracking the speech source; 2) System works in very small computation amounts; and 3) Regression weights can be statistically optimized over the given training data. Once the optimal regression weights are obtained by regression learning, they can be utilized to generate the estimated log spectrum in the recognition phase, where the speech of close-talking is no longer required. The performance of the proposed method is illustrated by speech recognition of real in-car dialogue data. In comparison to the nearest distant microphone and multi-microphone adaptive beamformer, the proposed approach obtains relative word error rate (WER) reductions of 9.8% and 3.6%, respectively.

This paper introduces an evaluation framework for Japanese noisy speech recognition named AURORA-2J. Speech recognition systems must still be improved to be robust to noisy environments, but this improvement requires development of the standard evaluation corpus and assessment technologies. Recently, the Aurora 2, 3 and 4 corpora and their evaluation scenarios have had significant impact on noisy speech recognition research. The AURORA-2J is a Japanese connected digits corpus and its evaluation scripts are designed in the same way as Aurora 2 with the help of European Telecommunications Standards Institute (ETSI) AURORA group. This paper describes the data collection, baseline scripts, and its baseline performance. We also propose a new performance analysis method that considers differences in recognition performance among speakers. This method is based on the word accuracy per speaker, revealing the degree of the individual difference of the recognition performance. We also propose categorization of modifications, applied to the original HTK baseline system, which helps in comparing the systems and in recognizing technologies that improve the performance best within the same category.

Behavioral synchronization between speech and finger tapping provides a novel approach to improving speech recognition accuracy. We combine a sequence of finger tapping timings recorded alongside an utterance using two distinct methods: in the first method, HMM state transition probabilities at the word boundaries are controlled by the timing of the finger tapping; in the second, the probability (relative frequency) of the finger tapping is used as a 'feature' and combined with MFCC in a HMM recognition system. We evaluate these methods through connected digit recognition under different noise conditions (AURORA-2J). Leveraging the synchrony between speech and finger tapping provides a 46% relative improvement in connected digit recognition experiments.

This paper describes a spatial spectral subtraction method by using the complementary beamforming microphone array to enhance noisy speech signals for speech recognition. The complementary beamforming is based on two types of beamformers designed to obtain complementary directivity patterns with respect to each other. In this paper, it is shown that the nonlinear subtraction processing with complementary beamforming can result in a kind of the spectral subtraction without the need for speech pause detection. In addition, the optimization algorithm for the directivity pattern is also described. To evaluate the effectiveness, speech enhancement experiments and speech recognition experiments are performed based on computer simulations under both stationary and nonstationary noise conditions. In comparison with the optimized conventional delay-and-sum (DS) array, it is shown that: (1) the proposed array improves the signal-to-noise ratio (SNR) of degraded speech by about 2 dB and performs more than 20% better in word recognition rates under the conditions that the white Gaussian noise with the input SNR of -5 or -10 dB is used, (2) the proposed array performs more than 5% better in word recognition rates under the nonstationary noise conditions. Also, it is shown that these improvements of the proposed array are same as or superior to those of the conventional spectral subtraction method cascaded with the DS array.

CIAIR, Nagoya University, has been compiling an in-car speech database since 1999. This paper discusses the basic information contained in this database and an analysis on the effects of driving status based on the database. We have developed a system called the Data Collection Vehicle (DCV), which supports synchronous recording of multi-channel audio data from 12 microphones which can be placed throughout the vehicle, multi-channel video recording from three cameras, and the collection of vehicle-related data. In the compilation process, each subject had conversations with three types of dialog system: a human, a "Wizard of Oz" system, and a spoken dialog system. Vehicle information such as speed, engine RPM, accelerator/brake-pedal pressure, and steering-wheel motion were also recorded. In this paper, we report on the effect that driving status has on phenomena specific to spoken language

Audio-Visual Speech Recognition (AVSR) is one of techniques to enhance robustness of speech recognizer in noisy or real environments. On the other hand, Deep Neural Networks (DNNs) have recently attracted a lot of attentions of researchers in the speech recognition field, because we can drastically improve recognition performance by using DNNs. There are two ways to employ DNN techniques for speech recognition: a hybrid approach and a tandem approach; in the hybrid approach an emission probability on each Hidden Markov Model (HMM) state is computed using a DNN, while in the tandem approach a DNN is composed into a feature extraction scheme. In this paper, we investigate and compare several DNN-based AVSR methods to mainly clarify how we should incorporate audio and visual modalities using DNNs. We carried out recognition experiments using a corpus CENSREC-1-AV, and we discuss the results to find out the best DNN-based AVSR modeling. Then it turns out that a tandem-based method using audio Deep Bottle-Neck Features (DBNFs) and visual ones with multi-stream HMMs is the most suitable, followed by a hybrid approach and another tandem scheme using audio-visual DBNFs.

We present a multichannel speech enhancement method based on MAP speech spectral magnitude estimation using a generalized gamma model of speech prior distribution, where the model parameters are adapted from actual noisy speech in a frame-by-frame manner. The utilization of a more general prior distribution with its online adaptive estimation is shown to be effective for speech spectral estimation in noisy environments. Furthermore, the multi-channel information in terms of cross-channel statistics are shown to be useful to better adapt the prior distribution parameters to the actual observation, resulting in better performance of speech enhancement algorithm. We tested the proposed algorithm in an in-car speech database and obtained significant improvements of the speech recognition performance, particularly under non-stationary noise conditions such as music, air-conditioner and open window.

A sound field reproduction method is proposed that uses blind source separation and a head-related transfer function. In the proposed system, multichannel acoustic signals captured at distant microphones are decomposed to a set of location/signal pairs of virtual sound sources based on frequency-domain independent component analysis. After estimating the locations and the signals of the virtual sources by convolving the controlled acoustic transfer functions with each signal, the spatial sound is constructed at the selected point. In experiments, a sound field made by six sound sources is captured using 48 distant microphones and decomposed into sets of virtual sound sources. Since subjective evaluation shows no significant difference between natural and reconstructed sound when six virtual sources and are used, the effectiveness of the decomposing algorithm as well as the virtual source representation are confirmed.

This paper describes subband-crosscorrelation analysis (SBXCOR) using two input channel signals. SBXCOR is an extended signal processing technique of subband-autocorrelation analysis (SBCOR) that extracts periodicities associated with the inverse of center frequencies present in speech signals. In addition, to extract more periodicity information associated with the inverse of center frequencies, the multi-delay weighting (MDW) processing is applied to SBXCOR. In experiments, the noise robustness of SBXCOR is evaluated using a DTW word recognizer under (1) a simulated acoustic condition with white noise and (2) a real acoustic condition in a sound proof room with human speech-like noise. As the results, under the simulated acoustic condition, it is shown that SBXCOR is more robust than the conventional one-channel SBCOR, but less robust than SBCOR extracted from the two-channel-summed signal. Furthermore, by applying MDW processing, the performance of SBXCOR improved about 2% at SNR 0 dB. The resultant performance of SBXCOR with MDW processing was much better than those of smoothed group delay spectrum (SGDS) and mel-filterbank cepstral coefficient (MFCC) below SNR 10 dB. The results under the real acoustic condition were almost the same as the simulated acoustic condition.

This paper describes a new method for estimating the direction of arrival (DOA) using a nonlinear microphone array system based on complementary beamforming. Complementary beamforming is based on two types of beamformers designed to obtain complementary directivity patterns with respect to each other. In this system, since the resultant directivity pattern is proportional to the product of these directivity patterns, the proposed method can be used to estimate DOAs of 2(K-1) sound sources with K-element microphone array. First, DOA-estimation experiments are performed using both computer simulation and actual devices in real acoustic environments. The results clarify that DOA estimation for two sound sources can be accomplished by the proposed method with two microphones. Also, by comparing the resolutions of DOA estimation by the proposed method and by the conventional minimum variance method, we can show that the performance of the proposed method is superior to that of the minimum variance method under all reverberant conditions.

In this paper, we address issues in improving hands-free speech recognition performance in different car environments using multiple spatially distributed microphones. In the previous work, we proposed the multiple linear regression of the log spectra (MRLS) for estimating the log spectra of speech at a close-talking microphone. In this paper, the concept is extended to nonlinear regressions. Regressions in the cepstrum domain are also investigated. An effective algorithm is developed to adapt the regression weights automatically to different noise environments. Compared to the nearest distant microphone and adaptive beamformer (Generalized Sidelobe Canceller), the proposed adaptive nonlinear regression approach shows an advantage in the average relative word error rate (WER) reductions of 58.5% and 10.3%, respectively, for isolated word recognition under 15 real car environments.

A single-dimensional interface which enables users to obtain diverse localizations of audio sources is proposed. In many conventional interfaces for arranging audio sources, there are multiple arrangement parameters, some of which allow users to control positions of audio sources. However, it is difficult for users who are unfamiliar with these systems to optimize the arrangement parameters since the number of possible settings is huge. We propose a simple, single-dimensional interface for adjusting arrangement parameters, allowing users to sample several diverse audio source arrangements and easily find their preferred auditory localizations. To select subsets of arrangement parameters from all of the possible choices, auditory-localization space vectors (ASVs) are defined to represent the auditory localization of each arrangement parameter. By selecting subsets of ASVs which are approximately orthogonal, we can choose arrangement parameters which will produce diverse auditory localizations. Experimental evaluations were conducted using music composed of three audio sources. Subjective evaluations confirmed that novice users can obtain diverse localizations using the proposed interface.

The separation and localization of sound source signals are important techniques for many applications, such as highly realistic communication and speech recognition systems. These systems are expected to work without such prior information as the number of sound sources and the environmental conditions. In this paper, we developed a dodecahedral microphone array and proposed a novel separation method with our developed device. This method refers to human sound localization cues and uses acoustical characteristics obtained by the shape of the dodecahedral microphone array. Moreover, this method includes an estimation method of the number of sound sources that can operate without prior information. The sound source separation performances were evaluated under simulated and actual reverberant conditions, and the results were compared with the conventional method. The experimental results showed that our separation performance outperformed the conventional method.

Blind source separation is a technique that can separate sound sources without such information as source location, the number of sources, and the utterance content. Multi-channel source separation using many microphones separates signals with high accuracy, even if there are many sources. However, these methods have extremely high computational complexity, which must be reduced. In this paper, we propose a computational complexity reduction method for blind source separation based on frequency domain independent component analysis (FDICA) and examine temporal data that are effective for source separation. A frame with many sound sources is effective for FDICA source separation. We assume that a frame with a low kurtosis has many sound sources and preferentially select such frames. In our proposed method, we used the log power spectrum and the kurtosis of the magnitude distribution of the observed data as selection criteria and conducted source separation experiments using speech signals from twelve speakers. We evaluated the separation performances by the signal-to-interference ratio (SIR) improvement score. From our results, the SIR improvement score was 24.3dB when all the frames were used, and 23.3dB when the 300 frames selected by our criteria were used. These results clarified that our proposed selection criteria based on kurtosis and magnitude is effective. Furthermore, we significantly reduced the computational complexity because it is proportional to the number of selected frames.

To improve speech recognition performance, acoustic feature transformation based on discriminant analysis has been widely used. For the same purpose, discriminative training of HMMs has also been used. In this letter we investigate the effectiveness of these two techniques and their combination. We also investigate the robustness of matched and mismatched noise conditions between training and evaluation environments.

Acoustic feature transformation is widely used to reduce dimensionality and improve speech recognition performance. In this letter we focus on dimensionality reduction methods that minimize the average classification error. Unfortunately, minimization of the average classification error may cause considerable overlaps between distributions of some classes. To mitigate risks of considerable overlaps, we propose a dimensionality reduction method that minimizes the maximum classification error. We also propose two interpolated methods that can describe the average and maximum classification errors. Experimental results show that these proposed methods improve speech recognition performance.

We carried out a one year field trial of a voice-activated automatic telephone exchange service at KDD Laboratories which has about 200 branch phones. This system has DSP-based continuous speech recognition hardware which can process incoming calls in real time using a vocabulary of 300 words. The recognition accuracy was found to be 92.5% for speech read from a written text under laboratory conditions independent of the speaker. In this paper, we describe the performance of the system obtained as a result of the field trial. Apart from recognition accuracy, there was about 20% error due to out-of-vocabulary input and incorrect detection of speech endpoints which had not been allowed for in the laboratory experiments. Also, we found that the recognition accuracy for actual speech was about 18% lower than for speech read from text even if there were no out-of-vocabulary words. In this paper, we examine error variations for individual data in order to try and pinpoint the cause of incorrect recognition. It was found from experiments on the collected data that the pause model used, filled pause grammar and differences of channel frequency response seriously affected recognition accuracy. With the help of simple techniques to overcome these problems, we finally obtained a recognition accuracy of 88.7% for real data.

In this paper, we discuss the construction of a large in-car spoken dialogue corpus and the result of its analysis. We have developed a system specially built into a Data Collection Vehicle (DCV) which supports the synchronous recording of multichannel audio data from 16 microphones that can be placed in flexible positions, multichannel video data from 3 cameras, and vehicle related data. Multimedia data has been collected for three sessions of spoken dialogue with different modes of navigation, during approximately a 60 minute drive by each of 800 subjects. We have characterized the collected dialogues across the three sessions. Some characteristics such as sentence complexity and SNR are found to differ significantly among the sessions. Linear regression analysis results also clarify the relative importance of various corpus characteristics.

In this study, toward the development of smartphone-based monitoring system for life logging, we collect over 1,400 hours of data by recording including both the outdoor and indoor daily activities of 19 subjects, under practical conditions with a smartphone and a small camera. We then construct a huge human activity database which consists of an environmental sound signal, triaxial acceleration signals and manually annotated activity tags. Using our constructed database, we evaluate the activity recognition performance of deep neural networks (DNNs), which have achieved great performance in various fields, and apply DNN-based adaptation techniques to improve the performance with only a small amount of subject-specific training data. We experimentally demonstrate that; 1) the use of multi-modal signal, including environmental sound and triaxial acceleration signals with a DNN is effective for the improvement of activity recognition performance, 2) the DNN can discriminate specified activities from a mixture of ambiguous activities, and 3) DNN-based adaptation methods are effective even if only a small amount of subject-specific training data is available.

This paper proposes a semi-supervised source separation method for stereophonic music signals containing multiple recorded or processed signals, where synthesized music is focused on the stereophonic music. As the synthesized music signals are often generated as linear combinations of many individual source signals and their respective mixing gains, phase or phase difference information between inter-channel signals, which represent spatial characteristics of recording environments, cannot be utilized as acoustic clues for source separation. Non-negative Tensor Factorization (NTF) is an effective technique which can be used to resolve this problem by decomposing amplitude spectrograms of stereo channel music signals into basis vectors and activations of individual music source signals, along with their corresponding mixing gains. However, it is difficult to achieve sufficient separation performance using this method alone, as the acoustic clues available for separation are limited. To address this issue, this paper proposes a Cepstral Distance Regularization (CDR) method for NTF-based stereo channel separation, which involves making the cepstrum of the separated source signals follow Gaussian Mixture Models (GMMs) of the corresponding the music source signal. These GMMs are trained in advance using available samples. Experimental evaluations separating three and four sound sources are conducted to investigate the effectiveness of the proposed method in both supervised and semi-supervised separation frameworks, and performance is also compared with that of a conventional NTF method. Experimental results demonstrate that the proposed method yields significant improvements within both separation frameworks, and that cepstral distance regularization provides better separation parameters.