PD's physics calculations

  • Thread starter Thread starter Saladine12
  • 21 comments
  • 2,076 views
Messages
69
Messages
FREEMASON007
Messages
SALADINE1
Can someone enlighten me on how PD translates say, the Ferrari F40's real life handling to the virtual world?
I mean what calculations do they do to come up with their idea of what the car in the game will feel like?
Another example, a car that has great turn in as opposed to one that doesn't.
How much of difference in information does that constitute?
What is it measured in?
Questions, questions, all i have is questions. Any answers?
 
Here is a quote from Craig Morrison at the US playstation, Gran Turismo board.

3GVette
To answer the question about handling, I spoke with CG about this a bit - there were around 70 questions that we had to answer so PD could plug the data into the physics engine. Some were pretty basic, weight, tires, height, etc., but when it came to the more complicated like inertia momemts, cg position, roll center locations, etc. all that requires a lot of work to get right. One of our engineers was crunching numbers for most of a day, and some were calculated based on simplified models. If we hadn't done this, we wouldn't have gotten any figure due to the complexity of the calculations.


Keep in mind these are the questions they as private owners/builders, automakers would be able to give much more detailed info at PD's request.
 
Can someone enlighten me on how PD translates say, the Ferrari F40's real life handling to the virtual world?
I mean what calculations do they do to come up with their idea of what the car in the game will feel like?
Another example, a car that has great turn in as opposed to one that doesn't.
How much of difference in information does that constitute?
What is it measured in?
Questions, questions, all i have is questions. Any answers?
From GT4, you can expect the same work but on a higher level for GT5:

gtscan1.jpg


gtscan2.jpg


gtscan4.jpg



And that was from a recent GT5P interview:
Kazunori Yamauchi: The F2007 is currently I think the most precise simulation of an F1 car in existence, and I think it's something that everyone should try out. We've had to implement a lot of new physics into the game, and a lot of mechanisms that we had left out before that we had to add to simulate the car properly, and it's something we're looking forward to having everyone experience.
 
Yeah but i really doubt PD managed to get all of the calculations necessary for every single car, especially in GT4. Some of those older models were probably fudged and "guesstimated' a bit. But maybe i'm wrong.
 
Can someone enlighten me on how PD translates say, the Ferrari F40's real life handling to the virtual world?
I mean what calculations do they do to come up with their idea of what the car in the game will feel like?
Another example, a car that has great turn in as opposed to one that doesn't.
How much of difference in information does that constitute?
What is it measured in?
Questions, questions, all i have is questions. Any answers?

This was the explanation given by Craig Morrison of the 3G Vette (Art Morrison Corvette in Prologue):

"To answer the question about handling, I spoke with CG about this a bit - there were around 70 questions that we had to answer so PD could plug the data into the physics engine. Some were pretty basic, weight, tires, height, etc., but when it came to the more complicated like inertia momemts, cg position, roll center locations, etc. all that requires a lot of work to get right. One of our engineers was crunching numbers for most of a day, and some were calculated based on simplified models. If we hadn't done this, we wouldn't have gotten any figure due to the complexity of the calculations.

I think with the Buick [refers to the Buick Special in GT4], because it was a privately owned car, there were a lot of assumptions made with these numbers and the car definitely handles a lot better [in-game] than the real thing. While we did the chassis, we didn't do the suspension, and it rides pretty rough. Enough so that the car's suspension was re-done, converting it from air suspension to coils and increasing the amount of travel and trying to eliminate the suspension bind."

So each car will be calculated a bit differently than all others based on the amount and quality of input data that is available and how much has to be guessed or fudged.

Craig also admits that the Prologue 3GVette handles better in the game than in real life.

Cheers,

MasterGT
 
Craig also admits that the Prologue 3GVette handles better in the game than in real life.

Cheers,

MasterGT
So, it's safe to say that what i'm driving in the game isn't exactly as described?
If that's the case, then how many other cars are not represented accurately?
 
In my book, that's impossible to predict. Imagine you get to drive a car you know very well from the game. You would expect it to be the same, but there are so many factors involved in the real word: bad weather, dirty tarmac, bad tires, chassis wear, engine wear, ...

So it's pretty unlikely that you'll be able to 100% accurately represent a car in the game. It's highly likely that the cars will drive similarly though.
 
I think they said they are only HALF WAY reaching this SIMULATION goal to 100%
 
I think they said they are only HALF WAY reaching this SIMULATION goal to 100%

I don't think they'll ever get to 100% accuracy. There are just too many outside factors that the Interceptor mentioned that are impossible to duplicate in the game.
 
In my book, that's impossible to predict. Imagine you get to drive a car you know very well from the game. You would expect it to be the same, but there are so many factors involved in the real word: bad weather, dirty tarmac, bad tires, chassis wear, engine wear, ...

So it's pretty unlikely that you'll be able to 100% accurately represent a car in the game. It's highly likely that the cars will drive similarly though.
Indeed.

There still things in the real world that engineers may still know nothing about. All the the ways tyres act on the limit is one example. And then there's the brand of similar tyres that can be used on a car that can change its handling considerably compared to the game.

Plus, there's likely the limitation of CPU power to work out all these figures anyway.
 
The 3G Vette handles better in the game than in real life?!?!?! :eek:

The thing is like a weasel on speed in the game. Wholly criminy man, can't imagine what it's like to really drive. :nervous:

I'll be curious to see if any of the anomalies make into the game, ie the 'wheelie' glitch. That's sooooooo nuts. As is the whole lifting the nose of the car to gain speed. It's fun though....👍
 
I don't think they'll ever get to 100% accuracy. There are just too many outside factors that the Interceptor mentioned that are impossible to duplicate in the game.

bad weather, dirty tarmac, bad tires, chassis wear, engine wear,

Those are all things that can be duplicated in game. Chassis and engine wear are the only ones I see difficult problems with :)

I do agree though. 100 % simulation can never be achieved, but we can get pretty damn close.
 
bad weather, dirty tarmac, bad tires, chassis wear, engine wear,

Those are all things that can be duplicated in game. Chassis and engine wear are the only ones I see difficult problems with :)
Theoretically those things can be included in the game, but that misses the point I was trying to make. What I meant was that if you compare the game to real life, it is extremely unlikely that you'll meet the exact same driving conditions. Therefore, there is no neutral basis to judge how real your experience is.
 
You can get close, but there's so many more varialbes that you account for irl than any computer game does. It's simple a case of how many calculations per second the PS3 can do compared to how many it would have to be able to do to simulate real life conditions wich would be varying every milli-second and every square milimeter of the game environment. Thinks like wind speed calcualted like a 3d grid, so each square of the grid would be different, the same would be true for direction temperature and density. Wind doesn't travel at one constant speed in one constant direction, it's always changing. Then you'd have the temerature of the the air changing every second, the density of the air changing and all of these things would be tied into each other. Already your talking about a combination of billions of calculations and we haven't even touched the road surface or the car yet.
 
Just to make things clear, there are no "calculations" associated with any single car, and there's no special "X-factor" in any of them. Each car is simply a long list of specifications, variables, and mathematical curves (eg. the engine's torque curve), and the game's physics engine uses this data combined with your control inputs to decide what the car should do. Some of the calculations are similar to what you might find in a college physics class, while others are simplified to save on processing power, but while you or I could manage one at a time with a pencil and paper, the CPU goes through dozens of them in the blink of an eye. As it does these calculations over and over, it extracts the important bits -- velocity, heading, traction, etc. -- and moves the car accordingly.

However, that doesn't mean it's easy to recreate realistic handling. If something is oversimplified or incorrect in the physics engine, then all of the cars in the game are flawed. If something goes wrong when entering data for a car, then that one car is flawed.

There's no such thing as a perfect simulator, and 99% of the time the problem is in the physics engine itself. Data entry is relatively simple and easy to double-check. So, for example, when Craig Morrison said the 3G Vette handled better in the game than it does in real life, I doubt he was alluding to a data entry problem. On the other hand, I also doubt he was using N tires, which of course would give him the impression of a very grippy car.
 
Theoretically those things can be included in the game, but that misses the point I was trying to make. What I meant was that if you compare the game to real life, it is extremely unlikely that you'll meet the exact same driving conditions. Therefore, there is no neutral basis to judge how real your experience is.

My post was a reply to 91hondawagon, I was just listing the items from your post he was referring to :) I know what you're saying and I agree.
 
Kazunori Yamauchi: "The F2007 is currently I think the most precise simulation of an F1 car in existence, and I think it's something that everyone should try out. We've had to implement a lot of new physics into the game, and a lot of mechanisms that we had left out before that we had to add to simulate the car properly, and it's something we're looking forward to having everyone experience."

Sorry, but that comes off as being completely out of touch with current sim offerings. Best F1 experience on a console? Yes.

Kaz is a genius, but he's admitted in the past to not really spending time playing or even paying attention to other driving/racing sims.
 
To be serious, there is no such thing as a 'perfect model' of a car because even car manufacturers can reproduce the same car day in, day out with the same performance results. It's a matter of fact that engine outputs differ from car to car that run off the production line, tyres differ, suspension geometry also differs slightly (but not to such a great degree).

Having been in the automotive industry for more than a decade, I've driven everything from Mini Coopers to Ferrari's. I used to work for a Lamborghini dealership and rarely came across two cars that handled exactly the same. I've owned and driven 4 different cars of the same MODEL and none of them were comparable.

Having Gran Turismo replicate ONE version of a car that is reproduced a million times over means that the version they feature in the game is bound to be different according to each individual owner.

Some people put too fine a point on a game.
 
Standard physics is just a tweaked version of GT4's to making driving easier for people. It's more "arcade" like compared to the pro.

Pro physics is the "new" physics engine that PD are using to make the driving more realistic. In comparison to the standard physics, grip is less, there's more oversteer/understeer, you can't brake and turn in so well, etc etc.
 
I wonder if in GT5 there will be a new name for the PRO physics since they will (I BET THEY WILL) update the physics a bit... are they going to call it SIM physics?
 
Back