This paper addresses the theory, design, and applications of surface acoustic wave (SAW) Fourier-transform processors. These analog processors are shown to perform several sophisticated real-time signal-processing functions at wide bandwidth (tens megahertz) making them attractive for use in radar, sonar, communication equipments. Theoretical results show how specific arrangements physically realizable SAW chirp filters permit Fourier transformation both baseband IF input signals. The...

This paper presents three experiments designed to empirically evaluate humanoid synthetic agents in electronic retail applications. First, human-like were evaluated a single e-retail application, home furnishings service. The second experiment explored application dependency effects by evaluating the same different personalized CD third effectiveness of range cartoon-like agents. Participants eavesdropped on spoken dialogues between "customer" and each agents, which played role...

Artists’ adoption and adaption of film have long been underwritten by benevolent, if often underpowered, pipelines funding. With the devolution British cultural policy in twentieth century, however, public subsidy for this marginal practice began to develop unevenly, brokering discrepancies that left artists Scotland economically disadvantaged a measure decades. Holding artistic production cannot be untethered from socio-economic context, article advances compromise contingency as immutable...

A parallel genetic algorithm is applied to assign the codevector indices for noisy channels so as minimise distortion caused by bit errors. The property of multiple global optima and average memoryless binary symmetric channel any error are also introduced. Experimental results confirm this approach.

HARP is a two-year project tnded by the European Community s TIDE programme (Technology Initiative for Disabled and Elderly people), with aim of developing speech rehabilitation system heating-impaired people.The system, based on an [BM-PC compatible microcomputer,

Addresses the problem of speech recognition with signals corrupted by additive noise at moderate signal-to-noise ratio (SNR). A model for is presented and used to compute uncertainty about hidden clean signal so as weight estimation provided spectral subtraction. Weighted dynamic time warping (DTW) Viterbi (HMM) algorithms are tested, results show that weighting information along can substantially increase performance subtraction, an easily implemented technique, even a poor without using...

A semicontinuous hidden Markov model (HMM), which can be considered as a special form of continuous-mixture HMM with the continuous output probability density functions sharing in mixture Gaussian codebook, is proposed. The function represented by combination discrete probabilities and codebook. amount training data required, well computational complexity HMM, reduced comparison to HMM. Parameters codebook mutually optimized achieve an optimal model/codebook combination, leads unified...

A weighted Viterbi algorithm (HMM) is proposed and applied in combination with spectral subtraction cepstral mean normalization to cancel both additive convolutional noise speech recognition. The approach compared used state duration modelling. results presented show that a proper weight on the information provided by static parameters can substantially reduce error rate, weighting procedure improves better robustness of than introduction temporal constraints low computational load. Finally,...

It is shown how distributed arithmetic techniques can be applied in parallel-data computations to achieve highly regular and efficient VLSI structures on silicon. Two individual processor chips are described as examples of the technique. The described, which intended primarily for computation FFT butterfly, each contain functional equivalence two parallel pipelined multipliers. first chip an 8-bit prototype device has been designed fabricated a standard 5-/spl mu/m silicon-gate n-channel MOS...

This paper reports an experiment to investigate users' preferences amongst three modes of data entry in automated home shopping service: DTMF input on the telephone keypad, and isolated word (IW) connected (CW) speech input. Preferences were measured both by means attitude questionnaires giving participants explicit choice among versions service once they had experienced them all. Users' attitudes with a given mode found vary according their cognitive skills (verbal spatial abilities)...

Star_pak (Signal Transform Analysis for Recognition PAcKage) is a software suite which provides all front end signal processing requirements the Edinburgh speech recognition system based on feature extraction techniques.The design of package "frame", Le. short-time segment order 25.6ms, muo-is processed by an array techniques provide rich acoustic description signal.The star_.pak strategy can be viewed as applying firstly set kernel transformations, such Fourier transformations and...

A bound for a Minkowski metric based on Lp distortion measure is proposed and evaluated as means to reduce the computation in vector quantisation. This provides better criterion than absolute error inequality (AEI) elimination rule Euclidean measure. For of order n, this contributes from L1 Ln metric. can also be extended quadratic which applied hidden Markov model with Gaussian mixture probability density function.

Dysphonic voices were used to compare electroglottographic (EGG) and acoustic measures of fundamental frequency (F/sub 0/) jitter using a wavematching an event based technique. Continuous speech was considered in the first part study, where effects pre filtering signals linearly smoothing F/sub 0/ contours analysed. The second investigation compared from sustained vowels (/i/ /a/, /u/), resulting poor agreement for /i/ /u/. In /a/ vowels, however, relatively small mean normalised absolute...

This paper addresses the problem of temporal constraints in Viterbi algorithm speaker-dependent and independent tasks. The results here presented suggest that a task introduction can lead to high improvement with additive or convolutional noise, statistical modeling state durations is not relevant if max min duration restrictions are imposed, truncated probability densities give better than metric previously proposed. Finally, word position dependent compared connected speech recognition...

This paper considers the nature of speech mechanism, and effects on spectrum a high pressure helium-air environment. A comparison is made between characteristics distortions in mixture certain well-known normal air which give rise to similar effects.The criteria for good intelligibility are related performance various helium unscrambling techniques have been used. These classified here into two main categories: those essentially using signal processing frequency domain time domain....

Although wikis are common in both the workplace and Higher Education, little research has studied wiki user experience. Recent literature highlights that users may be anxious about editing content; yet most of this anxiety not been measured quantitatively. computer metrics exist to measure towards technology, they lack specificity relevance context. This paper reports two studies used validity reliability inventory-editing (WAI-E), an inventory developed (Study 1) explore factor structure...

The problem of speech pulse detection with additive noise at a signal-to-noise ratio (SNR) as low 0 and –6 dB is addressed. assumed to be reasonably stationary correlated. Three techniques have been examined: the autoregressive analysis noise; spectral density comparison; non-stationarity measure.

A prototype real time cepstrum analyzer incorporating surface acoustic wave (SAW), Fourier transform processors is reported. This system offers sophisticated wideband signal processing for radar, sonar, and communications applications. Practical results demonstrate its capabilities when analyzing bandwidths in excess of 10 MHz a few microseconds with simulated pulsed RF waveforms the presence multipath echoes. Pulse duration, repetition interval, binary code length are resolved potential to...