Perceptual linear predictive (PLP) analysis of speech

A new technique for the analysis of speech, the perceptual linear predictive (PLP) technique, is presented and examined. This technique uses three concepts from the psychophysics of hearing to derive an estimate of the auditory spectrum: (1) the critical-band spectral resolution, (2) the equal-loudness curve, and (3) the intensity-loudness power law. The auditory spectrum is then approximated by an autoregressive all-pole model. A 5th-order all-pole model is effective in suppressing speaker-dependent details of the auditory spectrum. In comparison with conventional linear predictive (LP) analysis, PLP analysis is more consistent with human hearing. The effective second formant F2' and the 3.5-Bark spectral-peak integration theories of vowel perception are well accounted for. PLP analysis is computationally efficient and yields a low-dimensional representation of speech. These properties are found to be useful in speaker-independent automatic-speech recognition.     

Dynamical replica analysis of disordered Ising spin systems on finitely connected random graphs

We study the dynamics of macroscopic observables such as the magnetization and the energy per degree of freedom in Ising spin models on random graphs of finite connectivity, with random bonds and/or heterogeneous degree distributions. To do so, we generalize existing versions of dynamical replica theory and cavity field techniques to systems with strongly disordered and locally treelike interactions. We illustrate our results via application to, e.g., +/-J spin glasses on random graphs and of the overlap in finite connectivity Sourlas codes. All results are tested against Monte Carlo simulations.     

Acoustic analysis of induced vocal stressby means of cognitive workload tasks

The purpose of this study was to determine the acoustic effects on voice of three tasks of cognitive workload and their possible relationship to stress. Acoustic analysis was used to measure stress and workload in four experimental tasks and two experiments. In the first experiment, subjects performed cognitive workload tasks under a stressful condition, performing the tasks as rapidly as possible without errors and with the knowledge that any errors committed would reduce their grade in a course. The second condition was to perform the same tasks but without the condition of stress related to the final grade. Four testing conditions were included. One was a baseline measure in which subjects spelled the Spanish alphabet. The second was the reading of a tongue twister, the third was the reading of a tongue twister with delayed auditory feedback, and the fourth was spelling the Spanish alphabet in reverse order. In each condition the subjects prolonged the vowel /a/ for, approximately 5 sec. All subjects performed a test to determine their overall level of anxiety. The results suggest that in conditions of experimentally induced stress there is an increase in the fundamental frequency (F₀) relative to baseline, an increase in jitter and shimmer, an increase in the high-frequency harmonic energy, and a decrease in spectral noise.     

Analysis of vocal tract parameters in parkinsonian speech

A method for the analysis of vocal tract parameters is developed, aimed to perform quantitative analysis of rigidity from speech signals of Parkinsonian patients. The cross-sectional area function of the vocal tract is calculated using pitch synchronous autoregressive moving average (ARMA) analysis. The changes in Parkinsonian subjects of the cross-sectional area during the utterance of sustained sounds are attributed to both Parkinsonian tremor and rigidity. In order to isolate the effects of the rigidity on the vocal tract from those of the tremor, an adaptive tremor cancellation (ATC) algorithm is developed, based on the correlation of tremor signals extracted from different locations of the speech production system.     

Analysis of vocal pulses in a speech signal

An algorithm for estimating the vocal pulse positions and durations in an actual speech signal is described. Testing of the algorithm shows that it outperforms the best of the competitor algorithms in accuracy on the average by a factor of two. The algorithm is less sensitive to spectrum distortions in telephone channels, to various types of noise, and to instability in duration and amplitude of pulses produced by the voice source. The accuracy of the pulse position estimate is sufficient for a synchronous speech signal analysis, while the speed of signal processing makes the algorithm suitable for real-time operation.     

Level statistics in disordered linear networks

We consider a linear electrical impedance network, and study its frequency of nearest level statistics, P(s). We find that this probability density evolves from a sum of gaussians at small disorder, to a linear combination of Poisson and gaussian orthogonal ensemble for large values of the disorder.     

Estimation of molecular orientations in disordered samples by a proton-NMR-based method

We introduce a procedure on the basis of proton nuclear magnetic resonance (NMR) for investigation of the orientation state of disordered samples like amorphous or nematic polymers. Advantageous features of this method are the following: (i) disorder of the sample is not a problem (other than in the case of X-ray); (ii) the method works faster than multidimensional NMR techniques; (iii) this procedure can be implemented also at more simple and inexpensive NMR spectrometers; and (iv) for the data evaluation it will be not necessary to know the molecular geometry. The latter is possible by introducing the expressions “relative orientation distribution” and “relative orientation degree” which characterize the difference of the orientation of the current sample in comparison to a reference sample. Contrary to the absolute orientation degrees the relative ones are easily available from wide-line proton NMR spectra. The method is demonstrated by applying it to monitor the qualitatively different behavior of the director fields of two liquid-crystalline polymer samples with different molecular weights which are exposed to a suddenly switched magnetic field. A temporary asymmetry of the orientation distribution could be detected and quantified.     

Homogeneity examination of a disordered magnetic alloy by means of CEMS

The depth sensitivity of the Mössbauer effect conversion electron spectroscopy (CEMS) offers a possibility to check the homogeneity of an alloy as a function of the sample thickness. We present experimental results on (Fe_0.65Ni_0.35)_1-xMn_x (x=0.026) prepared by diffusion of Mn into a Fe_0.65Ni_0.35 foil. On an inhomogeneous sample we determined the diffusion coefficient of Mn in Fe_0.65Ni_0.35 at 1300 K to be about 10^?11 cm²/s.     

Spastic/spasmodic vs. tremulous vocal quality: motor speech profile analysis

Strained, strangled, and tremulous vocal qualities that are typically seen in adductor spasmodic dysphonia (ADSD), voice tremor (Tremor), and the spastic dysarthria of amyotrophic lateral sclerosis (ALS) may sound similar and be difficult to differentiate. The purpose of this study was to determine if these vocal qualities of neurologic origin could be differentiated on the basis of acoustic and motor speech parameters. Three groups of subjects (ADSD, ALS, and Tremor) were analyzed by the Motor Speech Profile System (Kay Elemetrics, Lincoln Park, NJ) for fundamental frequency (Fo), standard deviation of Fo, diadochokinetic rate (ddk), standard deviation of ddk, mean intensity and standard deviation of ddk, frequency and amplitude variability in connected speech, and speaking rate in connected speech. Profiles of the three groups are presented with the significant features that differentiated one from the other.     

Selection of dominant multi-exciton transitions in disordered linear J-aggregates

We show that the third-order optical response of disordered linear J-aggregates can be calculated by considering only a limited number of transitions between (multi-) exciton states. We calculate the pump-probe absorption spectrum resulting from the truncated set of transitions and show that, apart from the blue wing of the induced absorption peak, it agrees well with the exact spectrum.     

On the synchronization of chains of nonlinear pendula connected by linear springs

In this work the theoretical model of multidimensional physical systems, representable as chains of nonlinearly coupled chaotic pendula subjected to harmonic excitations, is formulated and its nonlinear dynamics and synchronization characteristics are studied by means of a numerical approach. Some considerations on the role of the main system parameters are drawn. Dynamic perturbations, due for example to background interactions or to intrinsic pathological imperfections of the chain, are also taken into account. Their effect is analyzed with reference to two distinct situations: uniform application to all the pendula and localized application to the extremities of the chain.     

Study of local magnetic non-homogeneities in disordered magnetics by means of Mössbauer spectroscopy

The analysis of the results of experimental studies of hyperfine interactions in disordered magnetics Fe_1?y Pd_y and FePd₂Au presented in the given work allowed to establish the interrelation between their atomic and magnetic structures. The magnetic moments of iron atoms in these alloys were defined and the character of their dependences on local atomic environment was clarified. The prevailing effect of local atomic environment on magnetic states realizing in FePd₂Au alloy was approved.     

Probing quantum-mechanical level repulsion in disordered systems by means of time-resolved selectively excited resonance fluorescence

We argue that the time-resolved spectrum of selectively-excited resonance fluorescence at low temperature provides a tool for probing the quantum-mechanical level repulsion in the Lifshits tail of the electronic density of states in a wide variety of disordered materials. The technique, based on detecting the fast growth of a fluorescence peak that is redshifted relative to the excitation frequency, is demonstrated explicitly by simulations on linear Frenkel exciton chains.     

Radical detection in harsh environments by means of laser-induced fluorescence using a single bidirectional optical fiber

A new experimental method is described enabling detection of hydroxyl radicals (OH) by laser-induced fluorescence in high-temperature gas-phase reactions. This is accomplished by means of a bidirectional optical fiber probe, which is of interest for applications where optical access is limited. An optical setup that allows simultaneous excitation and detection of fluorescence using one and the same fiber has been developed. Complications resulting from coupling as well as laser-induced scattering are addressed, and different fibers are compared with regard to core material composition and geometric collection efficiency. On this basis, a suitable fiber is identified, and OH detection and profile measurements are demonstrated in a premixed laminar flame as reference experiment.     

Decay kinetics of metastable states in a disordered linear system

Kolmogorov’s method for describing statistical crystallization is generalized to take into account disorder in the structure and is applied to the calculation of the decay kinetics of metastable states in linear (in the sense of spatial dimensionality) systems. The theory describes the influence of defects on the dynamics of polymers and domain boundaries between phases on substrates, on crystal growth involving the movement of steps over a surface, on the plastic flow of materials due to the motion of dislocations in a Peierls-Nabarro relief, and on the kinetics of other physical systems. Fiz. Tverd. Tela (St. Petersburg) 41, 1988–1993 (November 1999)     

An approach to normalization of coarticulation effects for vowels in connected speech

A method is proposed to reduce the ambiguity of vowels in connected speech by normalizing the coarticulation effects. The method is applied to vowels in phonetic environments where great ambiguity would be likely to occur, taking as their features the first and second formant trajectories. The separability between vowel clusters is found to be greatly improved for the vowel samples. In addition, distribution of the vowels on a feature plane characterized by this method seems to reflect their perceptual nature when presented to listeners without isolation from their phonetic environments. The results suggest that the method proposed here is useful for automatic speech recognition and help infer some possible mechanisms underlying dynamic aspects of human speech recognition.     

Estimating speech spectra for copy synthesis by linear prediction and by hand

Linear prediction is a widely available technique for analyzing acoustic properties of speech, although this method is known to be error-prone. New tests assessed the adequacy of linear prediction estimates by using this method to derive synthesis parameters and testing the intelligibility of the synthetic speech that results. Matched sets of sine-wave sentences were created, one set using uncorrected linear prediction estimates of natural sentences, the other using estimates made by hand. Phoneme restrictions imposed on linguistic properties allowed comparisons between continuous and intermittent voicing, oral or nasal and fricative manner, and unrestricted phonemic variation. Intelligibility tests revealed uniformly good performance with sentences created by hand-estimation and a minimal decrease in intelligibility with estimation by linear prediction due to manner variation with continuous voicing. Poorer performance was observed when linear prediction estimates were used to produce synthetic versions of phonemically unrestricted sentences, but no similar decline was observed with synthetic sentences produced by hand estimation. The results show a substantial intelligibility cost of reliance on uncorrected linear prediction estimates when phonemic variation approaches natural incidence.     

Comparison of direct and indirect calculations of laryngeal airway resistance in connected speech

Direct measures of subglottal pressure obtained through a tracheal puncture were used to calculate laryngeal airway resistance. Six subjects completed tasks including syllable trains and more natural speech samples produced at three loudness levels. Direct calculations of natural speech resistance values were compared with indirect estimates obtained during syllable train production. The degree of correspondence between direct and indirect calculations varied by subject. Overall, the smallest relative errors among calculations occurred for syllable trains, with higher relative errors for the monologue and sentence. For loudness conditions, the smallest and largest relative errors occurred for soft and loud productions, respectively. The clinical utility of indirect estimation is questioned and suggestions for improving its validity are provided.     

Estimation of crystal sample structure by means of fast electron radiation

The dependence of the characteristics of X-ray radiation occurring under fast electron transmission through a crystal on a sample structure has been analyzed. A method to estimate the size of the crystal microblocks by means of the yield ratio of the parametric X-ray radiation and diffracted bremsstrahlung and transition radiation for different observation angles has been suggested. The possibility of estimating the dimension of the microblocks using the ratio of the intensities of these emission mechanisms for the same observation angle and several reflection orders has been discussed.