Extended VTS for noise-robust speech recognition

Model compensation is a standard way of improving the robustness of speech recognition systems to noise. A number of popular schemes are based on vector Taylor series (VTS) compensation, which uses a linear approximation to represent the influence of noise on the clean speech. To compensate the dynamic parameters, the continuous time approximation is often used. This approximation uses a point estimate of the gradient, which fails to take into account that dynamic coefficients are a function of a number of consecutive static coefficients. In this paper, the accuracy of dynamic parameter compensation is improved by representing the dynamic features as a linear transformation of a window of static features. A modified version of VTS compensation is applied to the distribution of the window of static features and, importantly, their correlations. These compensated distributions are then transformed to distributions over standard static and dynamic features. With this improved approximation, it is also possible to obtain full-covariance corrupted speech distributions. This addresses the correlation changes that occur in noise. The proposed scheme outperformed the standard VTS scheme by 10% to 20% relative on a range of tasks. © 2006 IEEE.

Fracturing of sand in compensation grouting

The phenomenon of fracturing in sand as a result of compensation grouting was studied. Processes of fracture initiation and propagation were explained and a parametric study was conducted in order to investigate the factors that cause sand fracturing to occur. Experimental results indicate that fracture initiation requires the existence of a local inhomogeneity around the injection position. Grout mixture in terms of water-cement ratio and fines content had major roles in sand fracturing, whereas injection rate had a minor influence under the tested conditions. © 2009 Taylor & Francis Group.

Factor analysis based VTS and JUD noise estimation and compensation

Model based compensation schemes are a powerful approach for noise robust speech recognition. Recently there have been a number of investigations into adaptive training, and estimating the noise models used for model adaptation. This paper examines the use of EM-based schemes for both canonical models and noise estimation, including discriminative adaptive training. One issue that arises when estimating the noise model is a mismatch between the noise estimation approximation and final model compensation scheme. This paper proposes FA-style compensation where this mismatch is eliminated, though at the expense of a sensitivity to the initial noise estimates. EM-based discriminative adaptive training is evaluated on in-car and Aurora4 tasks. FA-style compensation is then evaluated in an incremental mode on the in-car task. © 2011 IEEE.

Speaker and noise factorisation on the AURORA4 task

For many realistic scenarios, there are multiple factors that affect the clean speech signal. In this work approaches to handling two such factors, speaker and background noise differences, simultaneously are described. A new adaptation scheme is proposed. Here the acoustic models are first adapted to the target speaker via an MLLR transform. This is followed by adaptation to the target noise environment via model-based vector Taylor series (VTS) compensation. These speaker and noise transforms are jointly estimated, using maximum likelihood. Experiments on the AURORA4 task demonstrate that this adaptation scheme provides improved performance over VTS-based noise adaptation. In addition, this framework enables the speech and noise to be factorised, allowing the speaker transform estimated in one noise condition to be successfully used in a different noise condition. © 2011 IEEE.

Rapid joint speaker and noise compensation for robust speech recognition

For speech recognition, mismatches between training and testing for speaker and noise are normally handled separately. The work presented in this paper aims at jointly applying speaker adaptation and model-based noise compensation by embedding speaker adaptation as part of the noise mismatch function. The proposed method gives a faster and more optimum adaptation compared to compensating for these two factors separately. It is also more consistent with respect to the basic assumptions of speaker and noise adaptation. Experimental results show significant and consistent gains from the proposed method. © 2011 IEEE.

Model-based approaches to handling additive noise in reverberant environments

Model-based approaches to handle additive and convolutional noise have been extensively investigated and used. However, the application of these schemes to handling reverberant noise has received less attention. This paper examines the extension of two standard additive/convolutional noise approaches to handling reverberant noise. The first is an extension of vector Taylor series (VTS) compensation, reverberant VTS, where a mismatch function including reverberant noise is used. The second scheme modifies constrained MLLR to allow a wide-span of frames to be taken into account and projected into the required dimensionality. To allow additive noise to be handled, both these schemes are combined with standard VTS. The approaches are evaluated and compared on two tasks, MC-WSJ-AV, and a reverberant simulated version of AURORA-4. © 2011 IEEE.

An allelic series reveals essential roles for FY in plant development in addition to flowering-time control

The autonomous pathway functions to promote flowering in Arabidopsis by limiting the accumulation of the floral repressor FLOWERING LOCUS C (FLC). Within this pathway FCA is a plant-specific, nuclear RNA-binding protein, which interacts with FY, a highly conserved eukaryotic polyadenylation factor. FCA and FY function to control polyadenylation site choice during processing of the FCA transcript. Null mutations in the yeast FY homologue Pfs2p are lethal. This raises the question as to whether these essential RNA processing functions are conserved in plants. Characterisation of an allelic series of fy mutations reveals that null alleles are embryo lethal. Furthermore, silencing of FY, but not FCA, is deleterious to growth in Nicotiana. The late-flowering fy alleles are hypomorphic and indicate a requirement for both intact FY WD repeats and the C-terminal domain in repression of FLC. The FY C-terminal domain binds FCA and in vitro assays demonstrate a requirement for both C-terminal FY-PPLPP repeats during this interaction. The expression domain of FY supports its roles in essential and flowering-time functions. Hence, FY may mediate both regulated and constitutive RNA 3'-end processing.

Force ripple compensation in a tubular linear generator for marine renewable generation

Tubular permanent magnet linear generators are a promising generator technology for use in marine renewables. One aspect of their design relates to the conditions necessary for achieving a smooth thrust response from the generator, free from cogging and periodic variations due to spatial harmonics of the flux cutting the generator coils. This paper presents an experimental and finite element study of the sources of thrust ripple in a prototype linear generator for marine generation. A simple self-commutated control scheme is shown, which uses linear Hall-effect sensors and look-up-table based feed-forward compensation to derive the excitation currents required to drive the machine with constant force. Details of the controller's FPGA based implementation are given, including its strategy for detecting sensor failure. © 2011 IEEE.

IGBT series connection under active voltage control

In this paper an Active Voltage Control (AVC) technique is presented, for series connection of insulated-gate-bipolar-transistors (IGBT) and control of diode recovery. The AVC technique can control the switching trajectory of an IGBT according to a pre-set reference signal. In series connections, every series connected IGBT follows the reference and so that the dynamic voltage sharing is achieved. For the static voltage balancing, the AVC technique can clamp the highest collector-to-emitter voltage to a pre-set clamping voltage level. By selecting the value of the clamping voltage, the difference among series connected IGBTs can be controlled in an accepted range. Another key advantage for AVC is that by changing the reference signal at turn-on, the diode recovery can be optimized. © 2011 EPE Association - European Power Electr.

Magnetic sensor with compensation for lead and coil resistance

A measurement system for magnetic fields or electric currents uses a single-core fluxgate, magneto-inductive or magneto-impedance device driven from a radio frequency excitation source. Flux nulling feedback circuitry is provided to maintain the core of the sensor at substantially zero net flux and improve the linearity and dynamic response of the sensor system. A high pass filter is provided for reducing the dc effects of the ohmic resistance of the coil and lead wires on the effectiveness of the flux nulling feedback.

Improving reverberant VTS for hands-free robust speech recognition

Model-based approaches to handling additive background noise and channel distortion, such as Vector Taylor Series (VTS), have been intensively studied and extended in a number of ways. In previous work, VTS has been extended to handle both reverberant and background noise, yielding the Reverberant VTS (RVTS) scheme. In this work, rather than assuming the observation vector is generated by the reverberation of a sequence of background noise corrupted speech vectors, as in RVTS, the observation vector is modelled as a superposition of the background noise and the reverberation of clean speech. This yields a new compensation scheme RVTS Joint (RVTSJ), which allows an easy formulation for joint estimation of both additive and reverberation noise parameters. These two compensation schemes were evaluated and compared on a simulated reverberant noise corrupted AURORA4 task. Both yielded large gains over VTS baseline system, with RVTSJ outperforming the previous RVTS scheme. © 2011 IEEE.

Diffraction based phase compensation method for phase-only liquid crystal on silicon devices in operation.

A method to measure the optical response across the surface of a phase-only liquid crystal on silicon device using binary phase gratings is described together with a procedure to compensate its spatial optical phase variation. As a result, the residual power between zero and the minima of the first diffraction order for a binary grating can be reduced by more than 10 dB, from -15.98 dB to -26.29 dB. This phase compensation method is also shown to be useful in nonbinary cases. A reduction in the worst crosstalk by 5.32 dB can be achieved when quantized blazed gratings are used.