Damage Modeling - How It's Done...

  • Thread starter Thread starter cmpinberkeley
  • 2 comments
  • 3,233 views
Messages
180
Messages
GTP_DarthDrift
Hey guys. I love GT5P but I had gotten bored with it a few months back for all the reasons that are well documented in the forum. I decided to try iRacing 👍. Because of the quality of the iRacing physics, graphics, great online racing, much less punting (hits your safety rating), damage, replays, qualifying, tire wear, fuel usage, pit stops, track modeling, comprehensive statistics, variety and coordination of events and enthusiastic community, I have rarely even thought of GT5P since. The purpose of this thread is not to rave about iRacing or knock GT5P (I love them both), but rather to share an iRacing forum thread from the designer of iRacing. The topics are upcoming updates and damage. I'm sharing this for two reasons:

1) iRacing gets updates just like GT5P and it's interesting how much more info gets shared with users. Perhaps PD will consider sharing more to keep interest up...speculation gets tiring after a while.

2) The damage modeling technical concept from iRacing is explained here. Their modeling is excellent, although perhaps too sensitive to low speed contact. It definitely promotes clean driving as even a light punt or contact with a wall will damage your car :nervous: and likely eliminate your chance for a Top 5. It's probably too late for PD to incorporate any of this, but who knows, maybe PD will do it better...

Anyway, just thought I would share this because it sums up why I've stopped playing GT5P. I'll keep checking the forums because I really do like GT5P and the community here, and I'm looking forward to the full version of GT5! Happy Racing!!!!

From Dave Kaemmer - 11 Sept 2008

Hello fellow iRacing Members:

As much as I enjoy working on the iRacing code, physics, tire modeling, the competition structure and new models and features, I also enjoy reading the forum posts as often as I get a chance. I often read the posts debating what should be next or to be blunt what is going on inside the walls of iRacing. After all these years, I guess I should not be surprised that my passion for sim racing is shared by so many people. That is a good thing! It certainly gives me the chance to do what I love.

I really do appreciate the passion for what is coming next at iRacing. We really donÂ’t have any big secrets (although we are working on some pretty cool surprises for next year). We are simply hard at work on the things that both you and we want for the service. Some of the bigger feature-related items that we are currently working on were included in one of the questions in our recent survey. Those items include pitting, flags, race control, replays and open practice. Some of these items should be out soon. I hope they ALL will be out soon, but in this business things never go as planned. It is a very complicated and technically challenging endeavor. I have learned not to make promises if IÂ’m not 100% certain weÂ’ll be able to deliver.

Anyway, we will post the results of that survey once we wrap it up – in case you’re curious. Thanks for taking the time to do it; the results are very helpful to us. We also have some very exciting new cars and tracks that we will be announcing shortly and bringing into the sim. I hope you saw the Lotus 79 announcement this week.

I was going to wait until we added blogging and social-network functionality to iRacing and then start a weekly developerÂ’s blog and let the technical, production and design teams here rotate and tell you the nitty gritty as to what they are working on and discuss some of the more technical aspects of iRacing, since there seems to be a portion of the membership interested in that detail. We should have the blogging capability done and up live on the site in a few weeks or so, but in the meantime, I figured I would send out at least a few quick development thoughts each week. Again, some people do seem interested in the details, so I asked Kevin Combs our 3D Art Manager and a terrific technical artist to talk about our collision system (below), one of the many models within iRacing. Hope you enjoy. Talk to you next week and good racing!

Dave


Our collision system is unique compared to systems used in the past. Rather than using a static collision volume to represent the driverÂ’s car, each car has a collection of spheres created by the artists, which represent the carÂ’s solid surface. Each sphere is animated to match the deformation of the visual model, so if you crunch the nose of your race car out of its normal position, the collision will be moved back to match its new position the next time you hit something. The graphical animation of the sphere also controls the damage assigned to the visual model, which in turn controls the damage texture drawn on the model.

Each sphere is imported from the graphical model into the physics definition for each car where physical characteristics such as the material type (plastic, steel, aluminum, carbon fiber, fiberglass, rubber, etc.), material strength, and its breaking point are defined. When a sphere exceeds its breaking point it is removed from the collision detection and any visual model associated with the sphere is detached from the car. In the future, visual models that are detached will remain in the world with their own collision spheres and mass; creating debris that is hittable and can cause damage to other cars.

Although the spheres themselves determine collisions between the driver and the world, as well as the driverÂ’s collisions with other drivers, the remote cars use a convex hull defined by the spheres for performance reasons; in a full field of cars, if all cars used the collision spheres, there would be thousands of sphere collision calculations. Spheres that surround your wheels and tires are treated specially, since the tire forces use a much more sophisticated force model.
 
I know nobody wants to hear an iRacing "Fanboy" talk about iRacing on the GTP website, but since I'm also a GT5P "Fanboy" and it's relevant to GT5P, I figure I can do it :sly:. Similar to the post above on damage modeling, I quote Steve Myers, the Executive Director of iRacing below on the process for introducing new content in the sim. He goes into some detail on licensing, scanning cars and tracks and physics.

This has given me a whole new appreciation for why it is taking PD quite a bit longer than we all think it should to come out with new content. Again, I am soooo looking forward to the new GT5P updates, but I just wish PD would approach the sharing of information a bit differently. Then again, who am I to question strategy/tactics/marketing if PD makes millions and millions more $$ than other driving software developers?

Posted by Steve Myers, Executive Director iRacing Oct 08

As I alluded to in my last blog post I will touch on the content creation process in this weekÂ’s blog. I am going to be going over this at a fairly high level as I would have to write a novel to explain exactly how it all works. We are also in the last stage of readying what will hopefully be a release candidate build for final testing so things are hopping around here.

I think the best place to start on how we create content is how we license it. We have signed many agreements with a wide range of partners through the years and I can tell you that is a time consuming and often frustrating task. Some agreements happen in a matter of weeks and some take years. Our wonderful Divina Galica (who I need to convince to start a blog to talk about all the incredible experiences she has had in her life) does most of the heavy work on acquiring our partners and I will tell you that she is focusing her energy on international properties right now (cough, cough, Spa). Tony Gardner and I also have a few other projects that we are working on domestically that will have a significant impact on our service if we can make them happen. Nothing surprises me anymore with this whole process and the bottom line is literally nothing happens with content until the bottom line is signed.

Once we have a signed agreement in hand we evaluate what is in the content development pipeline already and where this particular partner may fit in our racing schedule. The issue that takes the most precedence right now is trying to fill out our racing ladder to keep pace with the license level progression of our customers. I know that some of you are frustrated by us announcing the signing of wonderful partners that you still have not seen in our service, but our content takes a significant amount of time to create and we need to fill out the license ladders first. I think itÂ’s probably fair to say that once an agreement is signed it will generally take three to six months before it actually starts its path down the development pipeline.

The development pipeline is constituted by many different steps. The first step is obviously the laser scanning process done by a team which is headed by Dave Moulthrop and includes our scanner operators Reed Rundell and Kevin Iannarelli. Scanning cars is pretty straight forward as they can generally be done in a day. It really is only a matter of finding a car and an owner who does not mind it smelling like baby powder for a few days. Yes that’s right, baby powder. We use the baby powder on the cars to dull the finish as it is too reflective for the scanner. Scanning a track is a much different animal and is often an extreme undertaking. We generally try and book scan trips to do three or four tracks at a time so the guys may be on the road for a month or more straight. They often work at night as this is the only time that the tracks are not rented for racing events or testing and cannot work in the rain. These guys are the unsung heroes of this company and the next time you take a lap around the Jefferson circuit you can thank this terrific team for dealing with a night of cows getting loose and making “deposits” on the track to bring this circuit to your PC.

Once the track scanning is complete the data makes its way to Greg Hill. Greg is our VP of Art and Production and is responsible for managing the car and track content creation process. Three specialized teams contribute towards making an iRacing racetrack. The track geometry itself (the area you can drive on) is created by a member of the production staff who builds the racing surface from the laser scanned data with our proprietary tools. The objects that surround the racetrack, such as the buildings, flag stands, and grandstands are created by one our object art teams. The textures that cover the track, such as the grasses, asphalts, dirt, and concretes are worked on by yet another art team who specialize in these types of materials. At any given time we are usually working on four to five tracks.

Our car creation process is a little different than tracks because there is quite a bit more data collection required and obviously engineering support needed. We still have multiple art teams that will take the scan and sometimes CAD data and begin the process of building and texturing the models. The difficult part of creating the cars is getting access to all the data we need to build the actual physics models. When we are working with a major manufacturer we generally can get everything we can dream of for data. Unfortunately this is often not the case and we have to send Ian Berwick who is our Vehicle Dynamics Engineer to the race shop that houses the car and he goes through an extensive process of weighing and measuring everything he can get his hands on. He also works with the manufacturers or teams to get tire and aero data if they have it but we also gather data ourselves by going to Calspan and the tire manufacturers which has given us a ton of data that we have barely scratched the surface on to begin implementing. One of the last steps we do in creating a car is to take it to a dyno or track to record the engine sound. We use the Aussie Greg Hill (we have two Greg HillÂ’s) at Soundwave Concepts to do the actual processing of the sound data and he really does a wonderful job of giving us terrific samples to plug into our sound engine. Generally, cars take us about three months to create but this is entirely dependent on how long it takes us to collect the data we need. For instance, we have had three data collection trips canceled or abruptly ended for the Daytona Prototype car because the car we planned on using was destroyed in testing or a race. We actually had our equipment on a car once when it was wreckedÂ…

Again, I just described an incredibly detailed and difficult process in 5 paragraphs which is really not doing any of the topics much justice. Maybe down the road I can talk some of the people who actually do some of these tasks into writing their own blog on what they do.

Steve
 

Latest Posts

Back