One of the most significant features of our media-dominated society is the power of images. A good image can swing a decision in favor of one project and spell the end of another equally viable one. “But these images still need to match reality. We use our software to make things look realistic rather than just pretty,” explains Jacques Delacour, the founder and president of Optis, a software vendor in Toulon specializing in optical simulation. “When we create computer-generated images, we work on real renderings of the chosen materials under fixed lighting conditions. The real scene will look like a screen image. That’s the secret of our success.”
These simulations that come very close to reality are the fruit of almost three decades of research, initiated by Delacour. They are based on a simple premise: light is energy rather than mere geometrically propagated rays. To produce a realistic simulation, you need to take into account physics of both light itself, and of the surfaces of the objects it lights up.
Delacour began by developing algorithms to simulate radiation using Monte Carlo methods [Ed. note: calculating a numerical value using probability techniques]. But he soon focused on studying light energy in optical systems, looking at radiation and spectrum, and including how light-energy reflecting surfaces behave. To do this, he used the bidirectionalreflectance distribution function (BRDF), which describes the appearance of materials in terms of their interaction with light at each point on their surface. Other simulation software vendors wishing to simplify the calculations did not use this notion of energy.
These light/material interactions drove Optis engineers to become interested in geometric modeling of objects and hence in CAD. They then created computer-generated images to validate simulation results. This is how SPEOS software came to be launched on the market in 1994.
Prediction by Age Bracket
Studying light energy has also enabled development of a human vision model. Optis went back to Yves Le Grand’s abacuses. According to the amount of light energy in a spectrum, these abacuses predict what, and with what degree of sensitivity, the human eye will see. Optis added a visual perception layer, which gives the image’s spectral radiance. “The eye is sensitive to this magnitude. Depending on the spectrum of an image’s pixel and its energy level, as well as the age of the person looking, there will be glare that changes how objects are perceived,” explains Delacour. “In the car industry, for example, this is used to predict how well people in different age brackets will see and read light information according to its type, color, and illumination level, or to determine how much discomfort is created by a reflection.”
In partnership with several business clusters, Optis has kept this technology at the heart of its R&D programs since 2005. The MARVEST project, run with the Mer Méditerranée innovation and business cluster, is a hyperrealistic, multi-sector ship-steering simulator, which evaluates users’ reactions to glare or changing weather conditions. The second project, Virtu’ART, backed by Pôle Pégase (an aerospace cluster in Provence), Airbus Group and Airbus Helicopter, uses digital mock-ups to manipulate ultra-realistic, real-time visual renderings in an immersive virtual reality system. These projects have brought hyperrealistic, real-time simulation out of the laboratory, leading Optis to develop an industrial platform based on 3D animation technology produced by the SimplySim start-up, which Optis acquired in 2011.
Moving in this direction has boosted the company’s growth: it has expanded from 40 to 120 people – mainly R&D engineers – in five years. Today, the car industry represents over half of its sales. All manufacturers in this sector use Optis software to design their vehicle headlights, signal lights, and dashboards. High-end car manufacturers use it to validate perceived quality, notably for evaluating gap and flush on bodywork that varies depending on the chosen paint color.
From Realistic to Real
In this field, Optis acquired the British software vendor Icona Solutions and its Aesthetica software at the end of 2013. This is specialist software for 3D visualization of the impact of manufacturing tolerances on the perceived quality of assembled end products. For example, the biggest car manufacturers (Nissan, GM Opel, Fiat, Chrysler, Porsche, and Bentley) use it to inspect gap and flush on their vehicles, a determining factor in car buying decisions. Since mastering the physical behavior of light and materials is integral to the Optis offer, they produce real 3D-model images, moving in real time and in complex environments, rather than merely realistic images. The software is also very popular in the electronics sector as it helps improve management of flat-screen backlighting energy and facilitates the transition to LED. In industry, the US Air Force uses the software to validate its pilots’ anti-glare sunglasses by parasitic lasers. The CEA (French Atomic Energy and Alternative Energies Commission) uses it to establish what the optical control system will see of nuclear fusion in the ITER reactor core. “These applications are about to come out of design and engineering offices and go to end users. Thanks to configurators, it will be possible to use them with clients for visual validation of choices they make from among the numerous options they are offered. It won’t matter whether we are dealing with aircraft, cars, or control rooms,” predicts Delacour.
Controlling ITER’s light energyThe ITER program [see page 32] integrates optical simulation developed by Optis. Plasma circulating in the core of the ring-shaped magnetic containment vessel is a mixture of two hydrogen isotopes (deuterium and tritium), heated to temperatures in excess of 150 million degrees. This plasma emits very intense light, comparable to sunlight. It is reflected by the internal wall to prevent any absorption, which would cause fusion of the reactor shell. Optical cameras measure two pieces of information to control this reaction: light emitted by the plasma itself, i.e. the signal, and the plasma’s reflection in the walls, which is called ‘noise’. Using simulation to separate out these two pieces of information is easily done, and gives a better understanding of how the plasma will behave. ??