[Focus] la fisica di FM4

What can you tell us about Forza 4’s new physics?

My favorite subject. There have been several key changes to the physics and handling of FM4. The four primary areas of change to the physics model itself are in the tires, steering suspension, and track bumpiness. Let’s get into the details.

Tires:
In Forza, we’ve always focused on our tire simulation. As a result, we’ve struck up multiple partnerships over the years to get tire data. In the past, we’ve worked with Goodyear, Toyo, and Michelin. We also used reference data from the Society of Automotive Engineers (SAE). In past Forza Motorsport games, we even featured tire carcass flex, tire surface heat, as well as tire gas heat and expansion. However, the model was based on a combination of data from all of our previous partnerships. And we had to do fix-up on the data to make all of the different data formatting play nicely together.

For Forza 4, we took an entirely new approach to our tire simulation. This time, we threw all of the old data away and asked Pirelli to provide us with all-new data for everything. Pirelli did custom tests on a huge variety of tires to cover all of our cases—including tire width and height, compound, inflation pressure, heat, wear, sidewall height, load, angle, etc. We then changed our system to accept the real-world data directly and without any fix-up from us at all. This means that the tires in Forza 4 behave exactly as the Pirelli test tires did, even in complex situations where multiple parameters are changing rapidly.

Just looking at graphs of the Pirelli data versus our Forza 3 data, there are several obvious differences. Unfortunately, I’m not permitted to expound on the data differences. As part of getting this level of access and partnership from Pirelli, we have to keep their data completely under wraps. From a gameplay perspective, these changes are most noticeable as you move load from tire to tire. There is a lot more nuance in the model than ever before.

Steering:
For Forza 4, we changed the steering model in three major ways. First, we spent time watching how quickly real-world drivers (both pros and amateurs) tend to rotate the steering wheel as they turn-in, correct, and countersteer. This led to a change in the steering rate for Forza 4. We haven’t found that this change is really noticeable, but we know it’s more accurate—especially when coupled with our second change.

Second, we reevaluated our max steering angles and the system that creates them per car. We were able to use our researched curb-to-curb turning radius in combination with our new tire data (per compound peak slip angle) and other researched parameters, such as track width and wheelbase to reverse engineer each car’s max steering angle. We already had this data researched for some cars, but not for all of them. So we filled in the research gaps. In Forza 4, lock-to-lock steering is far more accurate for every car in the game.

Finally, we added the “simulation steering” option. This option removes all of the controller aides. There were very few of these aides on the wheel controller, but there were several on the console controller. With these aides removed, you will no longer get help finding the right counter-steering angle. If you over-countersteer (meaning you overcorrect by steering into the angle of the slide), the wheels do as commanded. This usually results in a wicked tank-slapper. Also, initial turn-in is direct and linear. This can feel very twitchy. When you command the game to go lock to lock-to-lock, it assumes you mean it and lets you live with the consequences.

Suspension:
There were a couple of suspension architecture types that we felt could use better research. Specifically, we wanted to rethink our implementation for solid-axle rear ends (again). Once we started looking at the system, we decided to reevaluate our MacPherson, double-wishbone and push-rod implementation as well. We had the overall simulation correct, but we found a better way of researching and simulating the per car differences within these systems. From a high level, this change is most noticeable on the classic and low-end cars. They feel more “boaty.” It’s not necessarily less spring and damping stiffness (though in some cases, it’s that too). It’s how the tire moves in relation to the body and the road. When combined with the new tires and steering, all cars have an entirely new feel in Forza.

Track Bumpiness:
The “feel” of Forza 4 is also different for two additional reasons: track surface bumpiness and dynamic camera animation. In past versions, we’ve optimized our track surface anomalies based purely on researched data and firsthand accounts from professional race drivers. As you know, we take great lengths to capture track as accurately as possible. What we capture, we recreated to the last detail. And, then we bring in pro drivers to further refine each track. When Gunnar Jeannette or Stéphane Sarrazin or other visiting drivers tell us there are bumps or chatters, we add them.

For this version, we added a new pass for all tracks. We evaluated the general age and quality of the track surface for each track and then added bumps around the entire track surface according to our evaluation. Different tracks and different driving surfaces around the track got different levels of bumpiness. However, these are tuned to be realistic bumps and cracks (this means subtle). While they are big enough to affect friction and upset the car, we found that most of the bumps and cracks we added were being absorbed by the suspension and tires before appearing noticeably in the car’s body and driver’s view.

In order to make bumps more perceptible, some games make them unrealistically big. However, this simply isn’t accurate and therefore we opted against it. Rather than taking those liberties, we changed our camera system to boost up the perception of the surface bumpiness. We trigger camera animation based on several new layers in Forza 4. We now have layers for accentuating bumps and cracks in the camera, based on the car physics. We even have new layers for heightening the perception of wind buffeting at high speeds. Combined, the tracks are noticeably more bumpy and engaging.

What about Performance Index changes?

I’m going to get a bit more technical in this explanation. The PI and classification system have been significantly changed in Forza 4. As you know, Forza features a dynamic car classification system we call Performance Index (PI). This system was rewritten for Forza 2 and again for Forza 3. Thus far, each new approach has addressed some big issues and created new ones. For this version, we decided to keep the same general system and refine it based on data rather than rewrite it. This approach was born out of the wisdom we’ve gained over the years through rewriting it—aka “the devil we know.” Forza 3 was incredibly well-explored and exploited by you, the players. Specifically, the community found issues with how the system scored AWD drivetrains and, to a lesser degree, downforce and weight. We took all of the data from our scoreboards and data logs (uploaded to the server, but not visible by players) to help us refine the PI simulation.

Unfortunately, we couldn’t realistically address the Forza 3 PI issues with a title update. In order to address PI in Forza 3, we would have had to put an entirely new database out there in a title update. And, the database is larger than what is allowed for a title update. Therefore, the only way to get the new database out would be with a user-selected or “Trojan Horse” DLC. There are many restrictions on this type of DLC. Also, the likelihood of mismatched databases between players is very high. The final straw was this: In order to make this fix, we would have had to wipe all of the scoreboards and “fix up” everyone’s saved games, as well as delete all of the tuning files, ghosts, and replays in the community (including those on the storefront, in your local save game, and those currently installed on all of the cars in your garage). In the end, we elected to study the issues in Forza 3 with better data-logging so that we could hopefully address them all in Forza 4.

In Forza 4, the PI and classification system has evolved considerably. First, we changed the simulation track that our PI system uses to evaluate each car. This new track was more scientifically derived. Not only did we look at the Forza 3 data results to find a more realistic stackrank target based on how our community’s fastest drivers perform, we also exported and studied the track geometry from all of our Forza 4 tracks to find a better way to design the PI simulation track. Our goal was to make a PI simulation track that not only better represents actual car stackrank target, but also does so by recreating the types of track combinations found throughout the game. This new track is longer and features more types of corners and combinations.

There were also two big changes in the PI simulation itself. Looking at the PI simulation in Forza 3, we found that AWD cars were not accelerating out of the corners as effectively as they would in the real world. This meant that the average speed for their simulated lap was lower—thus they had lower PI. This error was related to the code as well as the simulation track geometry. So we changed the track geometry to have more increasing radius lines to follow. We also changed the code to push the front tire limits more while accelerating (especially for AWD cars). These changes made the AWD cars faster in the simulation. By making them faster in the simulation, they get a higher PI. Obviously, these changes don’t change how fast they are on the track when you’re driving them. This just changes how they are scored by the PI system. Therefore, higher PI simulation results for AWD cars means less headroom for upgrades before hitting the class boundary.

On a related note: I saw questions in the community regarding Forza’s simulation of drivetrain loss in AWD cars. Forza 3 (as well as Forza 4) feature both higher driveline inertia as well as higher drivetrain friction for AWD cars, relative to RWD and FWD cars. This system is highly nuanced. Some AWD drivetrains are actually very efficient and some RWD systems are highly inefficient. However, as an average, AWD systems are less efficient than RWD systems in Forza. Drivetrain efficiency was not the source of the AWD PI issue in Forza 3.

In order to make bumps more perceptible, some games make them unrealistically big. However, this simply isn’t accurate and therefore we opted against it.

grande ludus io lo avevo letto all'acqua di rose e nel tuo resoconto ho compreso molti dettagli in piu...

tuttavia col mio inglese scalcinato questo l'ho capito bene pure da solo...direi che è abbastanza interessante non trovi?
Non per dire ma rispecchia quello che ho sempre pensato di fronte a certi giochi salterini...alcuni pure con un certo blasone...........

Trovo molto interessante l'analisi del problema che c'era con le AWD e il fatto che vada così nello specifico!
Hanno anche pensato di patchare F3 ma si sono resi conto praticamente subito che era una cosa infattibile, in pratica si trattava di resettare il gioco da zero!

L'aiuto sullo sterzo sarà duro da disattivare con i pad, se non impossibile, ma con i volanti renderà la guida molto più realistica.

Una cosa che spero abbiano implementato, vista anche la collaborazione con Pirelli, è il decadimento degli pneumatici, in F3 quasi non si sente, soprattutto nelle classi alte dove il consumo dovrebbe essere elevato. Sarebbe molto bello avere una gomma che si consuma un pò come quella della F1, costringendo chi guida a "sentire" quando è il momento di fermarsi, a meno di non voler perdere un sacco di tempo.

Da possessore del volante gt2 Posso solo dire che leggere queste info in merito non mi fanno pentire dei soldi spesi per la postazione e per il volante e quanto pare renderà ancora di più realistica la guida.

Parlando delle gomme è davvero importante leggere che la turn 10 sta facendo dei gran passi avanti immagino già i vari tornei e serate organizzate con il prosimo capitolo ci.sarà veramente da divertirsi, non vedo l ora di provare le demo se uscira.

Aspettando forza 4

immagina se tutto questo fosse riservato ai possessori del csr xD