Littérature scientifique sur le sujet « Physics-based sound synthesi »

Two seemingly incompatible worlds of quantum physics and acoustics have their meeting point in experiments with the Bose-Einstein Condensate. From the very beginning, the Quantum Music project was based on the idea of converting the acoustic phenomena of quantum physics that appear in experiments into the sound domain accessible to the human ear. The first part of this paper describes the experimental conditions in which these acoustic phenomena occur. The second part of the paper describes the process of sound synthesis which was used to generate final sounds. Sound synthesis was based on the use of two types of basic data: theoretical formulas and the results of experiments with the Bose-Einstein condensate. The process of sound synthesis based on theoretical equations was conducted following the principles of additive synthesis, realized using the Java Script and Max MSP software. The synthesis of sounds based on the results of experiments was done using the MatLab software. The third part or the article deals with the acoustic analysis of the generated sounds, indicating some of the acoustic phenomena that have emerged. Also, we discuss the possible ways of using such sounds in the process of composing and performing contemporary music.

The proposed sound synthesis tool converts a physics-based frequency-domain model of wind turbine trailing edge noise to a time-domain signal while accounting for the appropriate time shift due to the propagation between the moving blades and the fixed observer. A window function that implements cross-fading between consecutive signal grains is proposed and a method to objectively estimate the influence of the synthesis parameters is described. As the synthesis tool is independent of the aerodynamic noise model, it can be readily adapted to auralize other noise sources such as turbulent inflow noise or stall noise.

Turbulence, internal waves, surface roughness, and other environmental variations in the atmosphere and ocean randomly scatter sound. Realistic representation of these variations is important for numerical wave propagation calculations. In principle, there are two primary approaches to create these representations: (1) physics-based, dynamical models for the atmosphere or ocean and (2) kinematic synthesis of random fields with prescribed statistical properties that do not necessarily capture the medium dynamics. The most common kinematic approach involves synthesizing fields from randomly phased Fourier modes. For statistically inhomogeneous media, the Fourier modes generalize to empirical orthogonal functions. An alternative kinematic approach, called filtered Poisson processes, distribute spatially localized functions with randomized positions and orientations. Quasi-wavelets (QWs) are a filtered Poisson process intended for turbulence and other self-similar media. While both the Fourier and QW approaches can be formulated to reproduce specified second moments of the random field, if the representations are constructed too sparsely, higher-order moments such as the kurtosis will be unrealistic. The kinematic approaches also underlie phase screen methods, which can be quite useful when the Markov approximation is valid.

AbstractPeople working on stellar populations can look forward to an exciting decade ahead. Investigations of stellar populations lie at the heart of the science cases being used to justify the development of upcoming telescopes and emerging instrumentation technologies. Examples abound, but I will focus on three case studies: (1) Wide field astronomy with upcoming ground-based and space-based survey facilities; (2) Adaptive optics, which has the potential to revolutionize our understanding of stellar populations in both nearby and distant galaxies; (3) The James Webb Space Telescope, which may well extend the reach of stellar population work to encompass the full range of the star-forming history of the Universe. However, most of these developments will require extensive advance preparation in order to be used effectively. The time to start that preparation is now (if not yesterday). Three areas which need urgent development are highlighted in these proceedings: (1) We need a wide-field high-resolution spectroscopic capability to augment wide-area imaging surveys; (2) We need a set of AO-friendly extragalactic deep fields in order to exploit upcoming AO-fed instrumentation; and (3) Existing tools for population synthesis modeling need to be extended in order to incorporate the effects of dust. Because the physics of dust creation and destruction is so complicated and uncertain, the latter capability sounds almost impossibly hard to develop, but in this talk I will argue that some simple approaches already exist that allow dust to be injected rather naturally into population synthesis models. I will show a concrete example where incorporation of dust into spectral synthesis models allows one to detect and characterize rate of formation of circumstellar disks at high redshifts.

Technological progress supports the development of digital game, so that, through people's lifestyle that relies more and more on wireless devices, games become mobile, easy to access, likable and popular. Accordingly, the purpose of this study is to find the relation between games and its players in regards to the communication process and interactive pattern that exist in Cut The Rope and Angry Birds as mobile physics game.This research is an explorative field analysis that uses Creswell case study divided into literature study, case identification, primary data collection and interpretation as the synthesis of the whole. The result of this research is focused on the visual perception and interaction between players and two researched objects.The result shows that visual element, game play, puzzle-like character and sound effect are the most influential aspects of the games since they are considered suitable to the condition and the ability of the casual game players. Besides that, it is known that visual perception is a representation of the function of a common-theme-object with its interpretation that fits the context in a game. Therefore, visual interaction between the players and the researched games are based on four crucial factors: theme, game play, physics movement and simplified presentations of the visual elements. Through the perceptive process, similar kind of games would create a bonding, called challenge based immersion experience, in which this involvement demands the players to understand rules of the games. Such interaction is not influenced by playing time duration and usually does not cause an addiction to the games. Hence, it is clear that the concept of a mobile game has to be understandable through the games' highly comprehensible way, rules, and purpose.This research is mostly focused on the study of the visual aspect of a mobile game, so others still have the chance to do deeply study of the other three crucial aspects of the mobile game: game play, puzzle-like character, and sound effect.

In questo lavoro sono state affrontate alcune questioni inserite nel tema più generale della rappresentazione di scene e ambienti virtuali in contesti d’interazione uomo-macchina, nei quali la modalità acustica costituisca parte integrante o prevalente dell’informazione complessiva trasmessa dalla macchina all’utilizzatore attraverso un’interfaccia personale multimodale oppure monomodale acustica. Più precisamente è stato preso in esame il problema di come presentare il messaggio audio, in modo tale che lo stesso messaggio fornisca all’utilizzatore un’informazione quanto più precisa e utilizzabile relativamente al contesto rappresentato. Il fine di tutto ciò è riuscire a integrare all’interno di uno scenario virtuale almeno parte dell’informazione acustica che lo stesso utilizzatore, in un contesto stavolta reale, normalmente utilizza per trarre esperienza dal mondo circostante nel suo complesso. Ciò è importante soprattutto quando il focus dell’attenzione, che tipicamente impegna il canale visivo quasi completamente, è volto a un compito specifico.
This work deals with the simulation of virtual acoustic spaces using physics-based models. The acoustic space is what we perceive about space using our auditory system. The physical nature of the models means that they will present spatial attributes (such as, for example, shape and size) as a salient feature of their structure, in a way that space will be directly represented and manipulated by means of them.

The work in this thesis presents and discusses new techniques for modelling string instrument vibrations for the purpose of physics-based sound synthesis. The work focuses on coupled systems, where phenomena such as sympathetic vibrations should naturally occur. Finite difference (FD) methods are chosen for modelling string vibrations for to their flexibility both in terms of local adjustments and possible extensions to nonlinearities. The resonating body is represented using a modal formulation. Modelling the body in such a way has the advantage of scalability which can improve efficiency. This is shown to be possible without affecting the overall timbre of the sound. The models were formulated using idealised shapes such as the beam and plate, but the modal formulation is general for all linear systems. A technique for interfacing the FD model of the string to the modal formulation of the body is presented. In this way the advantages of both methods are exploited, improving the balance between accurate and efficiency. Initially, this coupling is formulated using only transverse motion but is then extended to include longitudinal motion. In simulations of a harp-like instrument where the strings are coupled to the body at an angle, results obtained with numerical experiments show that including longitudinal vibrations impacts the eigenmodes of the system and prove essential for accurately modelling sympathetic vibrations. Comparisons with previous studies validate these results. By applying the proposed method to model resembling a simplified piano, online changes to parameters such as soundboard density further demonstrate the proposed technique.

AbstractThis chapter addresses the first building block of sonic interactions in virtual environments, i.e., the modeling and synthesis of sound sources. Our main focus is on procedural approaches, which strive to gain recognition in commercial applications and in the overall sound design workflow, firmly grounded in the use of samples and event-based logics. Special emphasis is placed on physics-based sound synthesis methods and their potential for improved interactivity. The chapter starts with a discussion of the categories, functions, and affordances of sounds that we listen to and interact with in real and virtual environments. We then address perceptual and cognitive aspects, with the aim of emphasizing the relevance of sound source modeling with respect to the senses of presence and embodiment of a user in a virtual environment. Next, procedural approaches are presented and compared to sample-based approaches, in terms of models, methods, and computational costs. Finally, we analyze the state of the art in current uses of these approaches for Virtual Reality applications.

Abstract Over the past decade, Machine Learning or, more generally, Artificial Intelligence, has made a stelar entry into the O&G world and is today routinely used from exploration all the way up to retail. At the same time, physics-based tools are still being seen as key for value generation and the O&G community is increasingly looking at combining "traditional" with "new" technologies in what has come to be known as "hybrid" tools. In this paper we explain how we Baker Hughes, C3 and KBC are combining Physics and Machine Learning to create Hybrid Digital Twins that help operators improve their margins, reduce their emissions and, in general, position themselves for sustainable growth. Before they are in operation, plants (and oilfields alike) are typically simulated using physics-based tools. In practice, therefore, physics-based simulation models pre-date any plant data. Once plant data starts being produced, however, it becomes clear they do not match exactly the theoretical models that were used before operation. It also becomes clear that plants are operated in a very narrow range, due to quality, production and HSE requirements. In this paper we document a hybrid approach for a digital twin used to generate optimized targets to a Crude Unit operation: synthetic data has been generated to overcome the limitations of available plant data. These synthetic data have then been used as additional training input for ML models, complementing instrumentation data, while at the same time, instrumentation data is used for calibration of the physics-based model. This hybrid approach has been applied to a Crude Unit, in an optimization use case, with remarkable results. The final goal was to make energy optimization targets to operations through its machine learning algorithms which are based on two years of historical data optimized by a rigorous process simulator. The targets are manually downloaded to the unit multivariable controller by the operators. The hybrid approach has been found to combine the benefits of each type of technique: First Principles provides a sound bases for (limited) extrapolation, while Machine Learning ensures the First Principles models remain tuned to the reality of the process and complements the physics in areas where it is insufficient to model the reality (as is the case of equipment performance degradation/failure prediction). The range of what-if studies (and, correspondingly, optimization options) is thus radically extended. This paper shows incremental benefits in optimization cases illustrated by a Crude Unit use case, resulting from the application of a hybrid digital twin using both physics-based and machine-learning models. The hybrid approach allows optimizers to evaluate scenarios beyond the historical operating envelope and opens the door to incorporating equipment condition considerations.

Mixed (MR) and Virtual Reality (VR) simulations are hampered by requirements for hand controllers or attempts to perseverate in use of two-dimensional computer interface paradigms from the 1980s. From our efforts to produce more naturalistic interactions for combat medic training for the military, we have developed an open-source toolkit that enables direct hand controlled responsive interactions that is sensor independent and can function with depth sensing cameras, webcams or sensory gloves. From this research and review of current literature, we have discerned several best approaches for hand-based human computer interactions which provide intuitive, responsive, useful, and low frustration experiences for VR users. The center of an effective gesture system is a universal hand model that can map to inputs from several different kinds of sensors rather than depending on a specific commercial product. Parts of the hand are effectors in simulation space with a physics-based model. Therefore, translational and rotational forces from the hands will impact physical objects in VR which varies based on the mass of the virtual objects. We incorporate computer code w/ objects, calling them “Smart Objects”, which allows such objects to have movement properties and collision detection for expected manipulation. Examples of smart objects include scissors, a ball, a turning knob, a moving lever, or a human figure with moving limbs. Articulation points contain collision detectors and code to assist in expected hand actions. We include a library of more than 40 Smart Objects in the toolkit. Thus, is it possible to throw a ball, hit that ball with a bat, cut a bandage, turn on a ventilator or to lift and inspect a human arm.We mediate the interaction of the hands with virtual objects. Hands often violate the rules of a virtual world simply by passing through objects. One must interpret user intent. This can be achieved by introducing stickiness of the hands to objects. If the human’s hands overshoot an object, we place the hand onto that object’s surface unless the hand passes the object by a significant distance. We also make hands and fingers contact an object according to the object’s contours and do not allow fingers to sink into the interior of an object. Haptics, or a sense of physical resistance and tactile sensation from contacting physical objects is a supremely difficult technical challenge and is an expensive pursuit. Our approach ignores true haptics, but we have experimented with an alternative approach, called audio tactile synesthesia where we substitute the sensation of touch for that of sound. The idea is to associate parts of each hand with a tone of a specific frequency upon contacting objects. The attack rate of the sound envelope varies with the velocity of contact and hardness of the object being ‘touched’. Such sounds can feel softer or harder depending on the nature of ‘touch’ being experienced. This substitution technique can provide tactile feedback through indirect, yet still naturalistic means. The artificial intelligence (AI) technique to determine discrete hand gestures and motions within the physical space is a special form of AI called Long Short Term Memory (LSTM). LSTM allows much faster and flexible recognition than other machine learning approaches. LSTM is particularly effective with points in motion. Latency of recognition is very low. In addition to LSTM, we employ other synthetic vision & object recognition AI to the discrimination of real-world objects. This allows for methods to conduct virtual simulations. For example, it is possible to pick up a virtual syringe and inject a medication into a virtual patient through hand motions. We track the hand points to contact with the virtual syringe. We also detect when the hand is compressing the syringe plunger. We could also use virtual medications & instruments on human actors or manikins, not just on virtual objects. With object recognition AI, we can place a syringe on a tray in the physical world. The human user can pick up the syringe and use it on a virtual patient. Thus, we are able to blend physical and virtual simulation together seamlessly in a highly intuitive and naturalistic manner.The techniques and technologies explained here represent a baseline capability whereby interacting in mixed and virtual reality can now be much more natural and intuitive than it has ever been. We have now passed a threshold where we can do away with game controllers and magnetic trackers for VR. This advancement will contribute to greater adoption of VR solutions. To foster this, our team has committed to freely sharing these technologies for all purposes and at no cost as an open-source tool. We encourage the scientific, research, educational and medical communities to adopt these resources and determine their effectiveness and utilize these tools and practices to grow the body of useful VR applications.