generated using an efficient dynamic method, and test data were time registered wi th the templates. A set of ten mel-frequency cepstrum coefficients computed every 6 " 4 ms resulted in the best performance, namely 96.. 5 % and 9500 % recognition with each of two speakers.. The superior performance

of operating at ten times the current design luminosity, i.e. at about 2×10 33 cm −2 s −1. The aim is to collect a data sample of ∼100 fb −1 as well as to increase the trigger efficiency for hadronic decays by at least factor of two. The present paper descibes briefly the physics motivation for an LHCb upgrade

This paper demonstrates a speaker identification system based on recurrent neural networks trained with the Real-time Recurrent Learning algorithm (RTRL). A series of speaker identification experiments based on isolated digits has been conducted. The database contains four utterances of ten digits

Recently, the quantum molecular sci-ence world was in uproar [1,2]. The preliminary list of approximately 25 speakers for the International Congress of Quantum Chemistry (ICQC) was published online, with no women speakers listed. One reaction to this list was to set up a petition to ‘‘condemn

developed by recording two different texts read consecutively by a set of ten speakers (male and female voices). The method involves extraction of the speech signal fundamental frequency (Fo) from the recorded speech, determination of the Fo intensity waveform within the utterances of given duration time

This study investigated how perceptual constraints affected the contrastiveness and intelligibility of spoken vowels from speakers with hearing loss. Ten speakers with normal hearing, 10 with severe hearing loss and 10 with profound hearing loss produced the vowels /a/, /i/, hi and IvJ each in a

This paper evaluates continuous density hidden Markov models (CDHMM), dynamic time warping (DTW) and distortion-based vector quantisation (VQ) for speaker recognition, across incremental amounts of training data. In comparing VQ and CDHMMs for text-independent (TI) speaker recognition, it is shown

technique is implemented in C. The data used in the tests are taken from the Danish EUROM. 1 database. All results are averaged over ten speakers. Three issues have been addressed: 1) the effect of varying the amount of adaptation material, 2) the effect of using different regression class divisions and 3

In cross-cultural interaction in medical settings, there is great potential for mtscommunication This potential is increased when the language proficiency of one of the interactants is low In the data presented here of such interaction, however, the instances of miscommumcation are quickly resolved and communication proceeds Three sources of communicative success are suggested. Most important is the ability of both parties to draw contextually triggered inferences which provide a basis for (1) making perceptual contributions and (2) using communication strategies These inferenctng abilities and communication strategies are guided by the expectation of and search for relevance (Sperber and Wilson 1995) which allows interaction to proceed A third source of success is the professional knowledge and skill of the interactants which not only enable interaction to proceed but to do so toward professional goals The implications of these findings contribute to our understanding of models of intentional communication.

This paper describes an efficient method for unsupervised speaker adaptation. This method is based on (1) selecting a subset of speakers who are acoustically close to a test speaker, and (2) calculating adapted model parameters according to the previously stored sufficient HMM statistics