Search

In noisy environments, hearing protection is used to occlude the ear canal’s external meatus to limit air conduction. However, acoustic waves travel through several pathways such as bones, cartilages, and soft tissues, which impact protections’ effectiveness. Physical properties of cartilage, as well as soft tissues can generate airborne sound in the occluded ear. To assess the part of the cartilage which contribute to the limitation of hearing protection, Bekesy audiometries with electrodynamic transducers are performed. Four zones (ear tragus, scaphoid fossa, ear lobe, behind the concha) are excited by a pulsed signal sweeping from 125 Hz to 8 kHz. The frequency response tends to denote that the ear tragus has the best transducer-cartilage coupling efficiency. The occlusion outcome is also evaluated at the mastoid and ear tragus, showing cartilage ability to transmit sound, especially at low frequency. For higher frequencies, the contribution of cartilage decreases, and the involvement of direct air conduction from transducer acoustic radiation increases, prominently when the ear canal is not occluded. Nevertheless, considering the transducer-skin coupling effects evaluated with airborne radiation measure and the contribution to the hearing perception of the earlobe, composed mainly of skin and adipose tissue, this paper raises the question of the conduction of the sound in soft tissues, and in a first step, in the understanding of the pathway of the sound when the ear canal is occluded.

Fabric noise generated by fabric-to-fabric friction is considered as one of the auditory disturbances that can have an impact on the quality of some textile products. For this reason, an instrument has been developed to analyse this phenomenon. The instrument is designed to simulate the relative movement of a human arm when walking. In order to understand the nature of the relative motion of a human arm, films of the upper half of the human body were taken. These films help to define the parameters required for movement simulation. These parameters are movement trajectory, movement velocity, arm pressure applied on the lateral part of the trunk and the friction area. After creating the instrument, a set of soundtracks related to the noise generated by fabric-to-fabric friction was recorded. The recordings were treated with a specific software to extract the sound parameters and the acoustic imprints of fabric were obtained.

A reproducible method to measure the intelligibility of communication systems is required to assess their efficiency. The current study seeks to develop a French version of the Modified Rhyme Test (MRT) [(House, Williams, Heker, and Kryter (1965). J. Acoust. Soc. Am. 37, 158-66], an intelligibility test composed of 50 six-word lists, originally developed for military applications and now widely used. An evaluation of the authors' French MRT was carried out, reproducing the original experimental conditions used by House and colleagues. Very similar results were found between the original MRT and the French MRT, validating the latter for further use and dissemination.

The objective of this study is to analyze the characteristics of frictional sounds and to determine correlations between sound parameters and mechanical properties of fabrics classically used for protective garments. The set of samples includes six fabrics whose surface was previously worn out at different degrees using a Martindale abrasion tester. The frictional sounds were obtained by a device reproducing the friction engendered under arm during walking and were recorded and analyzed using Fast Fourier Transform (FFT) analysis. In addition, the mechanical properties of samples were measured using the Kawabata Evaluation System. The FFT spectra of fabrics showed different shapes and different noise levels according to the degree of wear. Several correlations between sound parameters and mechanical properties were highlighted. It was shown that the garment’s wear increases the sound level of rustling sounds.

International audience; Tactical Communication and Protective Systems (TCAPS) are hearing protection devices that sufficiently protect the listener' s ears from hazardous sounds and preserve speech intelligibility. However, previous studies demonstrated that TCAPS still deteriorate the listener' s situational awareness, in particular, the ability to locate sound sources. On the horizontal plane, this is mainly explained by the degradation of the acoustical cues normally preventing the listener from making front-back confusions. As part of TCAPS development and assessment, a method predicting the TCAPS-induced degradation of the sound localization capability based on electroacoustic measurements would be more suitable than time-consuming behavioral experiments. In this context, the present paper investigates two methods based on Head-Related Transfer Functions (HRTFs): a template-matching model and a three-layer neural network. They are optimized to fit human sound source identification performance in open ear condition. The methods are applied to HRTFs measured with six TCAPS, providing identification probabilities. They are compared with the results of a behavioral experiment, conducted with the same protectors, and which ranks the TCAPS by type. The neural network predicts realistic performances with earplugs, but overestimates errors with earmuffs. The template-matching model predicts human performance well, except for two particular TCAPS.

Recent advances in the development of hearing protection devices open new fields of applications on the battlefield. While the TCAPS (Tactical Communication And Protective Systems) protect against acute acoustic traumas, they maintain information relative to the acoustic environment of the soldiers. Today, these systems are efficient for both hearing protection and communication. We propose to use the microphones equipping the TCAPS headsets in order to detect and localize shooters on the battlefield. The microphone underneath the hearing protection is used in order to detect the shock and muzzle waves generated by supersonic shots. A meshed network between the TCAPS deployed on the field allows transmitting asynchronous information relative to the detected waves to data fusion nodes that allow estimating the shooter’s position. Solutions are proposed in order to compensate the effects of the presence of the head between the microphones underneath the hearing protection. Results concerning the estimation of the time difference of arrival of a transient wave in free field and in the presence of an artificial head are presented. The data fusion process is tested thanks to simulations in various deployment configurations.

International audience; This paper deals with voice transformation techniques in the case of natural-to-shouted voice transformation. A database is first recorded and analyzed in order to find relevant cues relative to the speaker's vocal effort, especially concerning the fundamental frequency. Then, a new experiment aiming at transforming natural voices into shouted voices is presented. It is based upon an observed repetitive f0-pattern. The transformation also takes into account intensity variations, spectral tilt variations and duration variations. A perceptual test was carried out in order to evaluate the quality and the perceived sensation of the transformations. The results show that about 80% of the proposed transformed voices are perceived as shouted voices.

Snipers have emerged as a major threat to troops in recent conflicts. To reduce this menace, the objective of the French- German Research Institute of Saint Louis (ISL) research project “IMOTEP” is to improve the detection of snipers on the battlefield. Our basic approach is to combine several sources of information for a fast and appropriate reaction when an unusual signal (e.g. a flash or a shot) is detected. The project includes several technologies developed at ISL: acoustical detection, fusion of distributed sensor network data, active imaging and 3D audio communication. The protection of camps, convoys or dismounted soldiers rests on a distributed acoustical sensor network that detects and localizes sniper attacks. An early estimation of the threat position is transmitted through a network to an active imaging system in order to confirm and refine this position by 3D imaging. The refined position is then sent to the control center which generates an alert message that displays the threat position using two formats: a tactical map and a 3D audio signal. In addition, the camp is protected by an ad-hoc sensor network used for intruder detection.

This paper presents a new sonar target classification approach based on the use of time-frequency filters. Their design is carried out from the free field response of a reference target, and more precisely from the analysis of echo formation mechanisms in the time-frequency plane. The study of the relevance and the robustness of this approach in approximately real sonar conditions is conducted from experimental measurements in a tank. A data base is set up that contains a large set of target responses in the free field, near different interfaces and in waveguide situations. First, the efficiency of the method for the recognition of a nickel molybdenum spherical shell, corresponding to a class of man made targets whose size is much smaller than the sonar beam (finite size) is shown (100% of recognition). Second, a classification procedure between different targets of finite size is conducted: more than 85% of good classification is obtained (except for the marble solid target). Finally, in the presence of ...

Hearing-protection system that provide level-dependent sound attenuation can protect the ear against potentially damaging sounds (such as loud impulsive noises), while at the same time allowing the perception of moderate-level signals (such as speech). Such systems come in two forms: passive (nonlinear-attenuation earplugs) and active (talk-through system). This study sought to quantify the effect of these systems on spatial hearing. To this aim, sound-localization performance was measured in twenty subjects, with and without ear protectors on. Five protectors (two passives and three actives) were tested. The results showed significant increases in the proportions of errors during the use of one of the systems tested. To clarify the origin of this effect, "protected head-related transfer functions" (PHRTFs), i.e., HRTFs obtained with the ear-protectors on, were measured in the horizontal plane for each of the systems tested. The comparisons of these measures between PHRTFs with HRTFs were found to be in agreement with the subjective tests.

International audience; In this study we focus on the relationship between the talker-to listener distance (TLD) and the dynamics of speech intensity and fundamental frequency. A new experiment for the extraction of pertinent information from prosodic parameter variations for different TLDs is described. The proposed new recording protocol allows simulating a real TLD by using acoustically insulated rooms. We propose to measure the dynamics of ID and intensity by using the variation between the minima and the maxima of their contours and the initial/final slopes of f0. The results show a strong correlation between these parameters and the TLD.

This paper presents a method for target recognition and classification in a shallow water environment. It is based on time-frequency filtering matched to a free field reference target response. The decision strategy lies on the comparison of the reference and the filter output signal. The method is applied to an experimental data base containing target acoustic responses measured in a tank for typical configurations (free field, semi-infinite space and waveguide). First, the recognition of a spherical shell is carried out. The obtained rate of recognition and confusion are more than encouraging. Then, a classification procedure is conducted and a degradation of the mean performances is to be noted in the more general case. However, the classification of 3D targets independently of their attitude gives quite satisfactory results.

In 2003 the Directive 2003/10/EC has been published by the European Union. It defines the maximum noise exposure levels for workers and the related necessities for hearing protection and for conservation programs. The limits for continuous noise are based on A weighted exposure levels as defined by ISO 1999:1990. For impulse noise the Directive limits the maximum peak pressure level to 140 dB including hearing protection. This single value limit which does not take into account any frequency or duration information is not adequate to evaluate the hazard of an impulse noise and lacks any scientific validation. Although it may not present many problems in industry, for the armed forces it has become a major problem, as it limits training and bans the use of some weapon. Therefore Germany has not implemented this Directive for its forces but uses the Damage Risc Critera (DRC) designed by Pfander. In other countries which have implemented the Directive “As-is”, other criteria based on A-weighted energy (LAex,8h) proposed by Dancer or the AHAAH proposed by Price are taken into consideration for being used. The presentation will discuss scientific basis of the different DRCs for impulse noise in use or being candidate in Europe.

When speaking, not only air conducted noise is generated, but also vibrations can be recorded at different places on the head using accelerometers. The bone conduction microphones are less sensitive to noise than regular acoustical microphones, they are usable in harsh environments and they are compatible with head equipment such as NBC protection devices. This paper reports the first results of a study designed to evaluate the differences in perception between speech recorded via a acoustic microphone and speech recorded using bone conduction. These differences may cause of a bad intelligibility of bone conduction communication, even if presented in an undisturbed environment. We study the recognition of ten French phonetic vowels recorded by a bone conduction microphone. A listening test is designed to show the confusions of phonetic vowels when listening to speech picked up by an air or bone conduction microphone. The tests show confusion between the vowels [i] [y] [u]. All of these have the frequency ...

Commercially available active hearing protectors are usually implemented as ear muffs with an analog driven feedback loop. The active attenuation bandwidth of these devices is limited to about 800 Hz. Moreover, it is not possible to adapt the analog system for specific noise events or to the user’s morphology. Active earplugs allow one to extend the bandwidth of the active protection to frequencies higher than 2 kHz. A digital feedback system has been implemented for active earplugs, in order to allow an adaptation to different noises and users. As the acoustic delays are too short to obtain a causal system when using FIR filters, the filters have been implemented as IIR type filters. The paper presents numerical simulations of different possible algorithms, used for the cascaded IIR filter of the ANR system. It shows how numerical errors are propagated in cascaded Bi‐Quad implementations, and how an optimum signal‐to‐noise ratio may be obtained. The numerical predictions and experimental results will be presented and compared. As these filters will be used in a closed‐loop feedback system, a numerical simulation of the behavior of the closed‐loop system, including the electroacoustic transfer function of the ear plug, will be presented.

Many professionals, e.g. mining and construction workers, ground crews or soldiers are exposed to impulsive or constant high level noise. In order to prevent hearing loss, they depend on hearing protections devices (HPDs). On the contrary, HPDs interfere with situational awareness and sound source localization. This contradiction makes users pondering between hearing loss and situational awareness. Often last mentioned dominates over first mentioned. This work aims to bring in line hearing protection and situational awareness. A virtual acoustic environment (VAE) with 16 circularly, horizontally arranged loudspeakers is set up. Localization performance with commercially available HPDs, including active and passive earplugs and earmuffs, is assessed in the VAE with 40 subjects. Earplugs with small geometries show better results than large-sized earmuffs. These results coincide with the study on modifications of the Head Related Transfer Function (HRTF) caused by HPDs. Earplugs preserve many individual spectral cues, while earmuffs cancel out most of these cues. We compare methods of combining a simulated, generic HRTF with the simulated, individual Pinna Related Transfer Function. An analytic model of HRTFs, controlled by the azimuth angle, is developed. Respecting the limitations of embedded systems, regarding energy supply and computational power, 14 digital filters are defined. A headphone based listening test is conducted to rate these filters regarding subjective front-back discrimination performance, resulting in better performance with low order filters than with high order filters. We present 4 designs of advanced HPDs which are aimed to improve the sound localization performance. Prototypes are manufactured and evaluated in a subjective listening test with 36 participants, showing that it is possible to improve sound localization of a commercially available active HPD., De nombreux professionnels sont exposés à des bruits impulsifs ou constants de très fort niveau. Pour se prémunir d'une perte auditive ils portent des protecteurs auditifs. Il s'ensuite la réduction de la localisation des sources sonores cependant, souvent les utilisateurs jugent la localisation plus importante que d'être protéger. L'objectif de cette étude est de concilier la protection auditive en gardant la connaissance situationnelle. Un environnement acoustique virtuel (EAV) de 16 haut-parleurs disposés circulairement et horizontalement est mis en place. La performance de localisation avec des protections acoustiques de types actives et passives, bouchons et casques, disponibles sur le marché, est évaluée avec 40 sujets. Ce test a montré que les bouchons sont à préférées aux casques. Ces résultats sont en corrélation avec l'étude sur les modifications des Head Related Transfer Functions (HRTFs) introduites par les protecteurs auditifs. Les bouchons conservent plus d'indices spectraux individuelles que les casques, qui enlèvent de nombreux d'indices individuelles. On compare 2 méthodes pour combiner une HRTF générique et pré-simulée avec la fonction de transfert individuelle relative à la coquer simulée. Un modèle analytique des HRTFs, contrôlé par l'angle d'azimut, est développé. Avec des contraintes imposées sur des systèmes embarqués en termes de ressources d'énergie et de puissance de calcul, 14 filtres sont définis. Un test d'écoute a permis d'évaluer ces filtres, concernant la discrimination des sons placés devant et derrière l'auditeur. Les filtres à l'ordre bas montrent des meilleurs résultats que ceux à un grand ordre. On propose 4 approches pour des protecteurs auditifs avancés de type casque, qui ont comme but d'améliorer la localisation des sons. Ces prototypes sont assemblés et évalués dans un test d'écoute avec 36 participants, qui montre qu'il est possible d'améliorer la performance de localisation d'un protecteur auditif du commerce.

While moving the soldier will generate fabric-to-fabric frictions that will generate itself friction sounds. It is important to study them because these noises can be annoying or even disabling during the soldier’s progress. Previous studies have shown that sound properties of fabrics depend on various parameters such as the weave pattern or the surface of the fabric. Wear of fabrics is one of the main factors affecting the surface of fabrics therefore the aim of this study is to analyze the wear’s influence on fabric friction sound by two different approaches. The friction sounds of four technical fabrics were generated and recorded using an experimental device to simulate the movement of the arm when a person is walking. In order to assess the wear’s influence, the samples were worn out by two different methods, by multiple washes and by the Martindale process. For each of these samples, acoustic parameters, mechanical and sensory properties were measured. Acoustic parameters such as sound level were obtained thanks to a Fast Fourier transform, the Kawabata Evaluation System (KES) was used to measure the properties of compression and surface of fabrics. Finally, a trained panel carried out the assessment of sensory properties. The results showed a different evolution of the total sound level depending on how the fabric was worn. A study of the correlations especially by principal component analysis between all these parameters was realized. Thus, a strong relationship between the sound level, the compressive resilience and the friction coefficient of fabrics was demonstrated. This multidisciplinary work has shown the wear’s influence on the noise level of friction. Indeed, this level is different according to the state of fabric’s surface therefore according to its degree of wear., Le combattant, en se déplaçant, va générer un frottement tissu contre tissu qui générera lui-même des bruits de friction. Ces bruits pouvant être gênants, voire fortement handicapants, lors de la progression du combattant, il est important de les étudier. Des études antérieures ont montré que les propriétés sonores des tissus dépendent de nombreux paramètres tels que l’armure ou l’état de surface du tissu. L’usure étant l’un des principaux facteurs entraînant des modifications de cet état de surface, l’objectif de cette étude est d’analyser l’influence de l’usure sur le bruit de frottement des étoffes par deux approches différentes.Lors de ces travaux, les bruits de friction de quatre étoffes techniques ont été générés et enregistrés grâce à un dispositif expérimental permettant de simuler le mouvement du bras lors de la marche. Afin de pouvoir évaluer l’influence de l’usure, les échantillons ont été usés par deux méthodes différentes, par de multiples lavages et par le procédé Martindale. Pour chacun de ces échantillons, les paramètres acoustiques, les propriétés mécaniques et sensorielles ont été mesurés. Les paramètres acoustiques tels que le niveau sonore ont été obtenus à partir de la transformée de Fourier des spectres sonores, le système d'évaluation Kawabata (KES) a été utilisé pour mesurer les propriétés de compression et de surface du tissu. Enfin, l'évaluation des propriétés sensorielles a été effectué par un panel entraîné.Les résultats ont montré une évolution différente du niveau sonore selon la façon dont l’étoffe a été usée. Une étude des corrélations par analyse en composantes principales entre l’ensemble de ces paramètres a été réalisée, une forte relation entre le niveau sonore, la résilience en compression et le coefficient de frottement des étoffes a ainsi été démontrée. Ces travaux pluridisciplinaires ont permis de mettre en évidence l’impact des conditions d’usage des textiles sur le niveau sonore des bruits de friction, en effet, celui-ci varie en fonction de l’état de surface de l’étoffe et donc en fonction de son degré d’usure.

This thesis focuses on speaker voice transformation in the aim to indicate the distance of it: a spokento-whispered voice transformation to indicate a close distance and a spoken-to-shouted voicetransformation for a rather far distance. We perform at first, in-depth analysis to determine mostrelevant features in whispered voices and especially in shouted voices (much harder). The maincontribution of this part is to show the relevance of prosodic parameters in the perception of vocaleffort in a shouted voice. Then, we propose some descriptors to better characterize the prosodiccontours. For the actual transformation, we propose several new transformation rules whichimportantly control the quality of transformed voice. The results showed a very good quality oftransformed whispered voices and transformed shouted voices for relatively simple linguisticstructures (CVC, CVCV, etc.)., Cette thèse porte sur la transformation de la voix d’un locuteur dans l’objectif d’indiquer la distance de celui-ci : une transformation en voix chuchotée pour indiquer une distance proche et une transformation en voix criée pour une distance plutôt éloignée. Nous effectuons dans un premier temps des analyses approfondies pour déterminer les paramètres les plus pertinentes dans une voix chuchotée et surtout dans une voix criée (beaucoup plus difficile). La contribution principale de cette partie est de montrer la pertinence des paramètres prosodiques dans la perception de l’effort vocal dans une voix criée. Nous proposons ensuite des descripteurs permettant de mieux caractériser les contours prosodiques. Pour la transformation proprement dite, nous proposons plusieurs nouvelles règles de transformation qui contrôlent de manière primordiale la qualité des voix transformées. Les résultats ont montré une très bonne qualité des voix chuchotées transformées ainsi que pour des voix criées pour des structures linguistiques relativement simples (CVC, CVCV, etc.).