A Racing Driver's Perspective of Gran Turismo 5: Prologue

  • Thread starter Thread starter Nitrof1
  • 52 comments
  • 4,370 views
but its not rally simple as putting some numbers in...
When you have an established physics engine, that's all there is to do.

Polyphony Digital does it; they don't program individual physics engines for each model. They program one physics engine, then attach a whole bunch of values and curves (representing specifications like curb weight and the torque curve of the engine) to each car. As you drive, the physics engine reads these values and curves to figure out what the car should do.

If the physics engine and specification values/curves are sophisticated and accurate enough, you shouldn't need to do anything "special" to make a car feel like its real-world counterpart. It'll feel like its real-world counterpart simply because all of the numbers are the same.
 
What's imaginary about 140hp, front-engine, RWD, and ~2600lbs.? That's a summary of the XR GT, and tells us that the car should behave like a real-world car that has 140hp, a front-mounted engine, RWD, and weighs ~2600lbs. It's as simple as that.

The game even describes (and represents visually, in real-time) the suspension geometry of each car. There's also the various other specifications provided and the fact that the suspension/etc. settings use real units of measurement.

It's the specifications that make the simulation, not the 3D models and licensed names.

The specifications aren't the same real cars for most the cars though, just loosely based on certain models&makes. So until I drive a car in LFS and real life that have the same spec, I'm not willing to say LFS is the most realistic.
 
If you create a car that has all the same specs as a Mini Cooper S but isn't a Mini Cooper S, it will drive like a Mini Cooper S. Thoes are the principals Wolfe was describing in a simple way.
Yep a 100% cloned car should drive the same but not all <200 hp ~1215 Kg FWD cars are clones.

A R32 GT-R not drive the same as R34, a twenty years old car not drive the same as an actual model with same specs, etc.. right now there's plenty of electronic applied on engine, drivetrain, suspension, gears, etc that give the caracter and personality on an specific model and all this is intented on each car and brand on GT.

LFS has good physics but too vague simulation of a modern real car.
 
The specifications aren't the same real cars for most the cars though, just loosely based on certain models&makes. So until I drive a car in LFS and real life that have the same spec, I'm not willing to say LFS is the most realistic.

That's not really it's point. It uses a certain amount of specifications such as the ones Wolfe stated (140hp, front-engine, RWD, and ~2600lbs.) and simulates how that would work in the real world. It comes pretty close.

If anyone's only wanting to know how LFS simulates a real car, then get the Formula BMW, and drive that.
 
Yep a 100% cloned car should drive the same but not all <200 hp ~1215 Kg FWD cars are clones.

A R32 GT-R not drive the same as R34, a twenty years old car not drive the same as an actual model with same specs, etc.. right now there's plenty of electronic applied on engine, drivetrain, suspension, gears, etc that give the caracter and personality on an specific model and all this is intented on each car and brand on GT.
An R32 and an R34 are not the same spec, that's why they are different. If two cars had all the same specs they would drive the same. Every aspect of a car is a specification, the ECU settings, the size of the fuel tank, the material the chassis is made with, the weight of the engine, the position of the engine. Every single thing is a spec and corresponds to a numerical value that can be inputted and interpreted by a physics engine. LFS basically takes a selection of specifications and says that is how it is irl. Ofcourse it's not perfect, it's well of the real experience, but it's the closest yet.

There's a program where you can alter the cars specs in the game, essentially recreating any real life car if you want. It won't look like the rea car, the model will still be one of LFS's models but the spec can be the same. The pricipal of what Wolfe originally said is still correct. I have to say, licensed cars still have a big appeal especially in the numbers Gran Turismo tpyically offers. It's worth noting that a couple of LFS's cars are real and licensed.
 
That's not really it's point. It uses a certain amount of specifications such as the ones Wolfe stated (140hp, front-engine, RWD, and ~2600lbs.) and simulates how that would work in the real world. It comes pretty close.

If anyone's only wanting to know how LFS simulates a real car, then get the Formula BMW, and drive that.

My point is you can't truly know if it's realistic or not until you drive the car in real life.
 
No but you can get an idea in general of how cars behave. An inaccurate physics engine produces oddities when you try to do things that the makers perhaps didn't expect you too or wern't concerned with. Yes you will need first hand experience of a Nissan R34 GT-R to know that GT5 simulates it accurately, but and this is a big but, you don't need first hand experience to know if it doesn't.

Physics are physics, when things happen in the game that don't in real life, there's no "yeah but have you driven that car irl" argument to come back on, certain flaws are easy to spot. GT4 had big problems with grip transition, big problems. Then you have the issue of cars being rooted to the road on impact, for example performaing the pit manouver was almost impossible, you don't need first hand experience of performing the pit manouver to know that. Then there's power delivery and the effect of flooring a car without touching the steering. I don't need to try it in a TVR Griffith 500 to know that that would leave you facing in the wrong direction especially on a cambered road. Then there's some problems that become evident when you mod a car like the speeds you can get some of the cars to go, there is no chance you'll get a BMW M5 to go 260mph with the power that you can give it in the game, it's simply not possible when you do the math. And while we're at it, the wheelie "trick", just because I've never tried in real life doesn't mean I can't say that with 100% certainty, putting thoes setting on a car on a long straight would result in it doing a wheelie and accelerating to speeds of 2000kph and more. Now you can argue that that's a glitch all you want, but it's simply exploitation of the flaws in the physics engine. Just because you won't be doing that in a serious race doesn't mean that the flaws that exagerates arn't already there.

I'd like to reiterate, I'm not interested in a GT vs argument here, I'm just not intereted in listening to GT is the god of sims and what not. As I said I haven't read too much about GT5:P's engine and I can only assume that if anything GT5's will be even better. But I am enjoying GT5:P and while there are one or two niggles I've found they arn't hampering my enjoyment like the flaws in GT4 did.
 
That's part of my point, with LFS the average guy is forced to go off a general idea of how the car would behave in real life when judging realism. I think it gives LFS more leeway, as with GT5P you have the option of driving the car in real life and judging by what you know the car is like, instead of what you think the car would be like. Also I'm trying to make an argument against comparing the two. Atleast until LFS gets some real roadcars that people can judge based on firsthand knowledge.
 
My point is you can't truly know if it's realistic or not until you drive the car in real life.
That's pretty much why it's called a simulator....The game is simulating those real world calculations.

But you won't be seeing any real cars until licensing is actually achievable. Until then, the only real car is the BMW Formula. And I have no doubt the game simulates it very well, mostly due to reading from actual racecar drivers on RSC.net.
 
If the only racing games you've played are Burnout, NFS, Dirt and GT - then comparatively GT absolutely whoops the rest of them in the physics department.

I wonder how many different racing games this guy has actually played?
 
If financial resource allocation is any indication of--or maps on the increase in--realism of chassis dynamics, then Sony's flagship title would be unrivaled by even LFS and rFactor.
 
That's part of my point, with LFS the average guy is forced to go off a general idea of how the car would behave in real life when judging realism. I think it gives LFS more leeway, as with GT5P you have the option of driving the car in real life and judging by what you know the car is like, instead of what you think the car would be like. Also I'm trying to make an argument against comparing the two. Atleast until LFS gets some real roadcars that people can judge based on firsthand knowledge.
That doesn't address the fact that you don't need first hand experience to know what shouldn't happen and to spot flaws and inaccuracies in the physics engine. At the end of the day there is a lot of information available regarding LFS physics engine, it's extremelly intricate but there's almost nothing about GT5:P's. I'm trying not to get into GT5:P's physics because while there are still some flaws (there are in all games and sims) they don't have nearly the impact that the flaws in GT4 had. But the point still stands, you don't need first hand experience to be able to say that's not right or to exploint the physics engine to degrees that have stupid results (ie the wheelie trick).

LFS has at least 3 licensed cars, the raceabout, the MRT and the BMW F1 car. It's not so much a question of can we compare thoes cars with real life, Gran Turismo hasn't featured any car I've ever oned (at least not the spec I have) so does that mean I can't be a judge of it's realism? No, absolutely not. I've driven plenty of cars and I know enough about physics to calculate whats possible and what isn't in many instances such as can x car go y fast with z power, the cars in GT4 could go a lot faster than would physically be possible and you don't need first hand experience to work that out.
 
I'm not talking about GT4, we all know it had big problems. I know you can spot certain flaws without firsthand experience. What I'm saying is that if you want to judge as accurately as possible you need first hand experience with the real car, which isn't possible with LFS for most people.
 
I don't dispute that first hand experience is the best gauge of accuracy, but without first hand experience of a particular car I'm fairly confident I could pick out the most accurate simulation by looking at the flaws and using logic and maths to be the judge. You can apply general real world experience and that same logic and such when playing Live For Speed, the game was head an shoulders above anything else in terms of realism not too long ago. The gap has closed somewhat but I still believe LFS is top of the tree when it comes to accurate physics in a driving game.

Jus because I've never seen a ball hit a flat object at 45 degrees in a weighless environment does not mean I don't know or can't work out what should happen. You can apply that logic to a lot of things. First hand experience does count for a lot, I'm not disputing that, bu just because most of the cars in LFS arn't licensed that doesn't mean it's impossible to tell how accurate it's hysics engine is.
 
My thoughts on GT5P so far, after one solid day of gameplay, and test driving 2 cars ( i drove others but only for one event for each)

the Integra and the Mustang...
The physics really really improved.. they now seem like a PC racing sim game.. the integra was rather really easy to drive when you dont over power it. It really feels that wind took bigger step in GT5P, at Dytona the with the integra when going top speeds, when you hit a small ditch and you steer, you can as almost feel the car lift a bit, and when i watched the replay, it seen that 2 tires slightly rise above the pavement because of the powerful air drag. so physics really really improved.

the mustang.. really feels heavy with lots of power and torque.. and you do feel that you have a solid axel in the back making it really hard to control when one back tire dirives over the curve (those red and white ones, i dont know how you call em) and then the other tire brakes loose and spins the car out if you over power it.. i was just watching the top gear review of the car and they were right.. its not a good handling sports car but it is cheap and powerful.

in both cars the suspensions really felt GOOD , after watching replays of my own the wheels mover UP and DOWN like they suppose to and drive correctly over curbs as they would in real life.

I yet have to try out some MR cars, super cars.

In my opinion, i really get more feel of the car then driving one in GTR2 or rFactor. and its a lot easier to keep them in control but at same time when you over do something cars easily spin out. In GT cars feel like they suppose to, some are heavy and bluky, yet some have hard suspension and swift. When i try out GTR2 they have few car classes but to me only difference i feel is Power and handling... but i guess those are racing cars and almost have the same Suspension settings... HARD and made for racing. The tracks seems really rally real, and a lot better then competition on PC

but at the same time GTR2 and rFactor, Race 07 feature more racing cars, Gran turismo feature normal sports and street cars. so its really hard to say one is better from the other.. like someone stated above.. if PC racing sims like LFS or GTR features street cars, then we would be able to compare the two.

as for now, get both and stop complain. those games are worth every penny
 
if PC racing sims like LFS or GTR features street cars, then we would be able to compare the two.

Ok so lfs has the raceabout that only a small group of people have driven. Either way the "road" cars in lfs are templates of sorts that could easily emulate many different real car platforms give or take some minor differences and the blatant lack of electronic aids (sans fz5). Just because you don't see a "brand" car that you might have driven doesn't mean that these car models don't reflect alot on how similar "real life" car platforms might react on the tarmac.

I second your enthusiasm with gt5p, but its still a bit arcade compared to the pc sims you bring up. :)
 
rFactor have a lot of good street car mods like the "C6 Corvette" and "Caterham"; both are done with "car factory" and uses "realfeel" feedback plug-in.
 
Well, the article is hardly what I would call a critical review. It starts with the basic premise that the GT series is the greatest thing since sliced bread & goes on from there.

It would be really interesting to read a comparison of real-world driving & a variety of console & PC sims such as GT5/P, Forza 2, Enthusia, F1:CE, Rfactor, GTR2, LFS etc. by a or (preferably) a number of professional drivers. It would seem like an excellent project for Top Gear for eg. (yeah, I know they did that Laguna Seca piece - but it was not really comparative).
 
i think top gear will be a little biased in their outcome since they have collaborated with PD to include the test track.I'm also a big fan of the best motoring crew...but they're also on GTTV.This only leaves one man to test whice one game is the most realistic............................................................................................................................................................................................ME!, no actually i recon it would have to be Tiff Needell from 5th Gear.
 
no actually i recon it would have to be Tiff Needell from 5th Gear.


Who has also played GT4 with KY previously, and has also talked very positively about GT4.

So he also can't be trusted.......... :lol:
 
Back