Encoding of speech in convolutional layers and the brain stem based on language experience

Comparing artificial neural networks (ANNs) with outputs of brain imaging techniques has recently seen substantial advances in (computer) vision and text-based language models. Here, we propose a framework to compare biological and artificial neural computations of spoken language representations and propose several new challenges to this paradigm. Using a technique proposed by Begus and Zhou (2021b), we can analyze encoding of any acoustic property in intermediate convolutional layers of an artificial neural network. This allows us to test similarities in speech encoding between the brain and artificial neural networks in a way that is more interpretable than the majority of existing proposals that focus on correlations and supervised models. We introduce fully unsupervised deep generative models (the Generative Adversarial Network architecture) trained on raw speech to the brain-and-ANN-comparison paradigm, which enable testing of both the production and perception principles in human speech. We present a framework that parallels electrophysiological experiments measuring complex Auditory Brainstem Response (cABR) in human brain with intermediate layers in deep convolutional networks. We compared peak latency in cABR relative to the stimulus in the brain stem experiment, and in intermediate convolutional layers relative to the input/output in deep convolutional networks. We also examined and compared the effect of prior language exposure on the peak latency in cABR, and in intermediate convolutional layers of a phonetic property. Specifically, the phonetic property (i.e., VOT =10 ms) is perceived differently by English vs. Spanish speakers as voiced (e.g. [ba]) vs voiceless (e.g. [pa]). Critically, the cABR peak latency to the VOT phonetic property is different between English and Spanish speakers, and peak latency in intermediate convolutional layers is different between English-trained and Spanish-trained computational models. Substantial similarities in peak latency encoding between the human brain and intermediate convolutional networks emerge based on results from eight trained networks (including a replication experiment). The proposed technique can be used to compare encoding between the human brain and intermediate convolutional layers for any acoustic property.

The acoustic property and impact behaviour of 3D printed structures filled with shear thickening fluids

In this study, the microstructures of the silica and styrene/acrylate particles and rheological behaviour of the three STFs were measured. The acoustic property and impact behaviour of 3D printed structures filled with STFs were investigated. The results showed that sound transmission loss (STL) of the structures filled with 46.5 vol% silica-based and 58.8 vol% styrene/acrylate-based STFs have been significantly improved, while their sound absorption coefficient (SAC) reduced greatly. The internal damage mechanism and energy absorption of honeycomb structures filled with different volume fraction STFs under low-velocity impact (LVI) loading were analysed, finding that the volume fractions and nanoparticles hardness of STFs has a significant influence on the impact resistance of the 3D printed honeycomb structures.

Pitch Range and Voice Quality in Dimasa Focus Intonation

This paper presents an analysis of Dimasa focus intonation. The acoustic analysis shows that narrow focus sentences undergo a jump in the pitch range irrespective of the underlying tonal value of the morpheme it attaches to. In addition to f0 expansion, the prosodic property of focus in Dimasa was found to have different (tense) phonation in morphologically marked narrow focus sentences when compared to the broad focus context. Thus, the tense phonation property of sentences bearing morphological focus is not only an acoustic property of a higher pitch range but may also be an acoustic cue of discourse-level intonation.

Advanced design of Close-Proximity (CPX) trailer enclosure acoustics on tyre/road noise measurement

The PolyU Mark II Twin-wheeled CPX trailer was developed for the measurement of tyre/road noise in Hong Kong urban environment according to a standard methodology (ISO/CD 11819-2) - the Close-Proximity (CPX) method. Numerical simulations of the acoustics of PolyU Mark II CPX enclosure were conducted and a good agreement between numerical and experimental results was obtained. In order to extend the capacity of the Mark II CPX trailer and enhance the acoustic performance within the enclosure for future tyre/road noise studies, the validated numerical simulations were carried on to design the next generation of the PolyU CPX system. Through analyzing the acoustic performance within the enclosures of different dimensions and the distributions of sound pressure level (SPL) inside the anechoic chamber, the geometry of the PolyU Mark III CPX enclosure was finally determined. With newly designed enhanced interior wall absorption, the new PolyU Mark III CPX enclosure design was delivered into numerical simulations for acoustic analysis. Fewer room modes and high uniformity of SPL distributions were observed within the new enclosure design. The PolyU Mark III CPX enclosure was fabricated based on the corresponding dimensions and the specific absorption layers. Great consistency was achieved between the numerical and measured results of the Mark III CPX enclosure. In addition, the PolyU Mark III CPX enclosure shows an improved acoustic property with a lower background noise level during road tests than Mark II CPX enclosure. The outcome of this study firmly establishes the feasibility of designing advanced CPX enclosure with numerical simulations with results that can be realized in realistic CPX measurement.

Convolution Neural Network Based Visual Speech Recognition System for Syllable Identification

Introduction: This paper introduces novel and reliable approach for speech impaired people to assist them to communicate effectively in real time. A deep learning technique named as convolution neural network is used as its classifier. With the help of this algorithm, words are recognized from an input which is a visual speech, disregards with its audible or acoustic property. Methods: This network extracts the features from mouth stances and different images respectively. With the help of a source, non-audible mouth stances are taken as an input and then segregated as subsets to get desired output. The Complete Datum is then arranged to recognize the word as an affricate. Results: Convolution neural network is one of the most effective algorithms that extracts features, performs classification and provides the desired output from the input images for speech recognition system. Conclusion: Recognizing the syllables at real time from visual mouth stances input is the main objective of the proposed method. When tested, datum accuracy and quantity of training sets is giving satisfactory output. A small set of datum is taken as first step of learning. In future, large set of datum can be considered for analyzing the data. Discussion: On the basis of type of Datum, network proposed in this paper is tested to obtain its precision level. A network is maintained to identify the syllables but it fails when syllables are of same set. Requirement of Higher end graphics pro-cessing units is there to bring down the time consumption and increases the efficiency of network.

Propagation characteristics of ultrasonic P-wave velocity in artificial jointed coal briquettes

Abundant engineering practice shows that coal and rock dynamic disasters often occur in fault zones; hence, the detection of faults in coal mines is important. As an essential branch of seismic tomographic application, research shows that ultrasonic P-wave velocity can detect mine faults and provide technical support for dynamic disaster prewarning systems. Traditional P-wave testing mostly takes rock or raw coal as the research object, which has some shortcomings in controllability, homogeneity and comparability. This paper compares the difference in ultrasonic P-wave propagation velocity in jointless and jointed briquettes through laboratory research, focuses on the effect of macro-joints on P-wave velocity and makes a preliminary theoretical analysis. The results show that: (i) the three-dimensional P-wave velocity throughout a jointless briquette specimen is similar, which reflects the high homogeneity of this medium and avoids the influence of the random distribution of primary bedding, joints and structural planes; (ii) the P-wave velocity in jointless and jointed briquettes is positively correlated with density and forming pressure, and is negatively correlated with the angle between the ultrasonic wave and the joint surface in a sample and (iii) when the P-wave encounters the macroscopic joint surface, it may reflect and refract, changing the propagation direction and inducing wave mode conversion. This study provides the necessary technical support and a theoretical guide to optimise acoustic property analysis of coal and rock as well as a field application for seismic tomography technology.

A Study on Acoustic Property of Composites from Waste Tyres

The basic aim in using the waste is to reduce the level of pollution that harms the environment. However a proper approach to manage waste or reduce the level of waste helps in conservation of resources for the succeeding generation and contributes to a healthy, clean and pollution free environment. In recent years, control on two major ecological issues such as waste disposal and noise pollution can help in achieving a green and healthier environment. For disposal, waste tires are normally burnt. Extinguishing fire from tire is very difficult. The fire develops a thick, dark smoke which contains CO2, SO2, cyanide, and products of butadiene & styrene. This study focuses on the fabrication of composites from epoxy-waste tire crumb by hand lay-up method and the investigation of its acoustical property by Impedance tube method using LABVIEW. In the present work three different materials such as PVC, Stainless Steel and Glass has been used to prepare impedance tube as per ASTM standards. All the sound testing has carried out at a frequency range of 500 Hz to 5000 Hz. This study clearly shows a promising waste management for tyres by developing a cost effective composite and thus contributing to a sustainable waste granule (waste tyre) reinforced polymer composite industry.