(written in September 2010)
Computer graphics, a passion.
40 years old, passionate graphic designer for more than fifteen years. From the time of the very first festivals in Imagina, from the time of the first feature films like Toy-Story, from the time of the first short films of Supinfocom’s, in short, you will have understood the digital media that makes you dream has fascinated me for a long time.
My goal then was to be an integral part of this dream world. But as always the game had only just begun. So it is certain there are few places for many elected officials, but hey that’s how it is. And then talent can’t be caught like a virus, so you might as well be patient and do your best to build your talent or in any case, lay the foundations of a building that you hope will ultimately be majestic once finished.
After my Technology & Electronic Baccalaureat in Aix-en-Provence (yes I know the guidance counselors in 1994 did not know too much about computer graphics, well it’s also my fault ^^) to fall back on an art preparation graphics at Ax-Sud – Marseille in 1998 before leaving for the deep end at ESMA – Montpellier, which I must admit has come a long way since the first 3D sections.
So in 2004, I joined APPORTMEDIA – Paris, a company that I liked very much, as my first professional experience, after my end-of-study internship, carried out a few months earlier with the same project manager. Company specializing in Serious-Gaming, in other words, real-time simulation (as in a graphically speaking video game for professional) training. Focused on driving simulators for companies managing large fleets and wanting to do prevention; this axis then turned to drive schools as a niche market. We use the same techniques as in video games (even if the means are not comparable) while keeping in mind that the physical precision of the vehicle, as well as the educational part, must remain the main objectives. The graphics are not yet comparable to great titles called « AAA » like Gran Tourismo 5, but the challenge is not the same either. At the time, competing simulators still looked like old Virtua-Racing games for a long time, that’s to say how we stood out once our first product was finalized.
Now even a serious simulator intended for professionals could and had to be more realistic and less austere than in the past while marking a subtly identifiable distinction with the world of video games. It is not easy to make a student understand that he is not sitting on an arcade station like Sega Rally, but a driving position intended for learning the basics of driving. The border is easily crossed if we do not do our design correctly.
The first product was then created between 2004 and 2005, then consisted of a circuit approximately 500 meters long and 300 meters wide. No more than 2 to 3 vehicles were on the screen, but many risky situations were staged. It was already a good start. This step allowed us to better prepare the graphic evolution, the environmental limits, and the technical choices that we would find in the following products. The challenge was then great because our goal was to drive in open space 5km long and 5km wide. The environmental part then became one of the key parts on which I was going to be able to express my knowledge and acquire know-how in this field.
A team of about 10 people, including 2 coworkers on the graphic design part for 4 years then 2 other years alone in charge of the projects. Suffice to say that it is a job that involves good group cohesion and a good dose of communication and listening in this small team, to bring these products to light. The Import / Export tools and the graphics engine being a pure in-house product (Thank you Damien), the link with the graphics development manager was all the more important. The physical part was also partly my responsibility because I had to optimize the structure on which the car was going to be able to run as well as possible. Definition of the spatial limits of the zone passable by vehicles, as well as the development of systems for identifying the different surfaces for the management of various adhesion.
It was a great experience for me to learn more about an area that we 3D « artists » often kindly overlook. When the creation « pipeline » is done via more robust and well-established tools such as publishers from the world of video games such as UnrealEd from Epic or the Sandbox editor from Crytek. There the processes are well known, the bugs too and everything is done very quickly and smoothly spent a few weeks on the workflow. The graphics constraints are not the same either and the CPU load allocated to physics limited us even more on the graphics part. To give you an idea, the vehicles had a polygon budget of 4 to 6000 triangles surroundings (interior and exterior included), while at the same time the games like Grand Turismo 4 released in 2005 on Playstation 2 included vehicles that were much higher in polycount than ours, our vehicle comprised of the order of 3 to 5,000 triangles (outside), but all for a console released in 2000 and which could display a good dozen vehicles on the screen when we could then only display 3 or 4. Today (2011) a Gran Turismo 5 is close to 100,000 to 200,000 triangles per vehicle (interior and exterior), while we are still in a low range of 6,000 to 12,000 triangles. The means of development and production are not the same either. The shading and lighting methods are not comparable either. Finally, like all young people who are starting, we take inspiration from what surrounds us and especially what works and Gran Turismo was the graphic objective to be achieved in a way, or at least towards which to strive the most. For example, where we dreamed of implementing shadow mapping, we had to fall back on the shadow volume technique, to light the vehicles on road. Here again, we were interested in the start of shading via a normal map, but not only did it require more resources, but also more time to create the car models. Suddenly, we fell back on shading techniques more within our reach with Diffuse Map including a Prelight pass (AO and Direct Lightning via baking). No specular map to differentiate plastics/fabrics, chrome, and other bodywork elements but segmentation of the elements constituting the vehicle into very distinct parts, to apply a shader with or without Cube Mapping reflections. I would speak again on the occasion of these technical choices and of what currently exists in terms of techniques intended for the rendering of vehicles.
STARS AE product on PC in 2006:
Model of the standard vehicle used on the version 2 simulator in 2005. A lot of information related to the shading is baked in the diffuse texture. Diffuse + Reflection via a CubeMap. Model has 7K triangles & 1024² Diffused texture.
Gran Turismo 4 sur PS2 en 2005:
Example of a model from the Gran Tourismo series. The shading is managed differently from what we can do with the technology at Apportmedia. But at the same time, the technology is not comparable. Model with 100 to 200K triangles.
But where the strength of the latest products released in 2008-2009 at Apportmedia lies is that we have managed to have an open environment covering an area of 400 km² (25km / 25 km) there or in 2004, when we started we were barely there. to 250,000 m², then 25 km ² in 2006. To give you an idea, the circuits often closed in current video games like the last GT5 or GRID is around 4 to 8 km ² in area, which is ultimately not much. But their objective is not the same either (at the same time it offers no less than 71 different circuits where we have a single modular infrastructure) but we want the user to circulate for a much longer time also than 6 to 10 mins and over distances of around 20 to 30 km.
Here we had to think, reflect and realize the tools that would suit us best, keeping in mind our medium and long-term objectives, as well as their implementation costs. A rich experience for the whole team.
This part of the work was the most interesting part for me because we had to think in collaboration with the graphic development manager (engine and import/export tools), on the best technical and practical choice to create such an environment in a minimum of time. Knowing that the scalability of this space had to be taken into account for other products to come, these tools had to be as robust and reliable as possible. It was not an easy task but given what was done in the market for game publishers and our needs, we needed to have essential tools such as the ability to update all the positions of the tree on-demand in the scene if the geography of the land were to change. For example how to manage to come placed the ground of the ground, so that there is no interpenetration between the road and this one. At that time I wasn’t using Houdini or any Procedural DCC app. All these things can perhaps make you smile knowing that this kind of thing is often automated from A to Z on today’s or yesterday’s editors like the Sandbox mentioned above, the QuadSoftware Grome, or the BF2 Editor from Battlefield 2 / EA. The snapping of objects or meshes, preset functions … etc, all this had to be managed manually in a way and above all to plan your work plan well so as not to have to go back on a tedious step that could put involved everything that would have already been done. Not simple, but with rigor no insurmountable problems.
It was 6 wonderful years. I who at first imagined myself staying only one or two years to then see the world of special effects cinema, advertising, or video games, got caught up in the game. And then it must also be said that 2004 coincided with my marriage, so hey, making life plans while changing jobs all the time, was quite risky at that time. I ended up taking a liking to it, seeing that the company was developing and that new professional and artistic perspectives presented themselves. But it was without counting on the caprices of life and I was obliged to leave France in February 2010, as well as my job to go and live in South Korea, my wife’s country of origin and live like this, with my wife and 3-year-old daughter. It has now been almost a year and a half that I am there and I must admit that I am acclimating to this culture quite a bit. On the one hand, because few people speak English well enough, at least in my region (because in Seoul I think it must be different), on the other hand, because I have to lead several fronts at the same time, the French education of my daughter far from her native land, the learning of Korean for my part, and the construction of a portfolio worthy of the name to find work here.
My life is pleasant and now turned largely to my daughter. Some work was done last year as a freelance, still for Apportmedia.
Well, otherwise I discovered Modo at the time of version 101 via my work colleague Borhan whom I thank in passing for having made me discover this pretty toy in more depth. My 3D software knowledge goes from Modo through Maya mainly (used in my professional and student career), ZBrush, Terragen, Leveller, and Xnormal also for some very specific tasks. As well as 3Ds Max that I discovered as a self-taught enthusiast and later a student.
Here I hope that this will allow you to know me better this September 2010.
Welcome to you.