[360, PS3] Need For Speed: Shift

S

standing ovation

Regular
Shift_SS03_Jun26.jpg


Judging from its E3 2009 showing, Need For Speed: SHIFT has sparked the interest of mainstream consumers. But among sim racing purists it has been a faulty ignition wire—a fire starter—of debate that has many reassessing their notion of simulation.

What is a simulator? Is it a realistic array of adjustment options and treacherously steep learning curves? Or is it a theoretical construct, a collection of mathematical models that mimic the interaction of mass and energy. Is it both?
 
Will the difficulty setting be as accurate as other sim games?

Ian Bell [SimBin co-founder and former CEO; Slightly Mad Studios’ Boss]: SHIFT is built on the most advanced physics engine we've made, and is by far the most advanced and realistic physics engine of any of the mainstream racing games. We're talking about an engine that can run unlimited threads, detailed physics parameters running at around 400 Hz on consoles. Things like tire deformation based on g-forces, tire pressures, rubber thickness, which only the most diehard sim fans will notice, but we know it's there.

Speaking of physics engines, will this game's physics be more in line with GRID, Forza/GT, or GTR2?

IB: You can't really compare SHIFT's physics to any of those games. The physics are tuned for realistic behavior on top of the most advanced simulation engine in this generation. We'd rather have you play SHIFT and tell us how it feels. We're confident that we'll be seen as the most accomplished racer of the current generation when the demo is released.

We see that iRacing and Codemasters are using laser scanning technology in order to reproduce tracks: did your team look into it?

IB: Our track layouts are extremely realistic. This is an area where we are confident we are surpassing the competition... It's not enough to simply get measurements and photos. If you want to recreate the experience, you have to know the subtleties of the track's surface contained in these data treasures [race telemetry - Ed.] that even laser scanning can't deliver. It does no good to know about a 5mm bump in the track if you have no data to show how it makes a car react. And the reaction, the experience, is what truly matters.
Click to expand...

Source: DrivingItalia
 
standing ovation said:
We see that iRacing and Codemasters are using laser scanning technology in order to reproduce tracks: did your team look into it?

IB: Our track layouts are extremely realistic. This is an area where we are confident we are surpassing the competition... It's not enough to simply get measurements and photos. If you want to recreate the experience, you have to know the subtleties of the track's surface contained in these data treasures [race telemetry - Ed.] that even laser scanning can't deliver. It does no good to know about a 5mm bump in the track if you have no data to show how it makes a car react. And the reaction, the experience, is what truly matters.
Click to expand...
That's just nonsense! The track is a track, a series of tarmac ups and downs. The simulation of the cars should recreate the response to those bumps.
 
If I were to play a race, this is the kind of PR talk that gets me excited. For something like racing, I either want arcade racing where you smash the hell out of everything, like Burnout or Flat Out, or I want hardcore sim. There isn't much of a middle ground. People are always tied up in the graphics on racers, but I care more about the interaction. I want to see cars behave like cars, or I want to see cars smashing down barns and exploding.
 
Shifty Geezer said:
That's just nonsense! The track is a track, a series of tarmac ups and downs. The simulation of the cars should recreate the response to those bumps.
Click to expand...
You can't do that level of simulation just from theoreticals, and even if you did it would be slow and a waste of computation.

Models based on real data are far more accurate than simulations. Responses can be fit with parameterized curves and recreated very efficiently and accurately. Even with the tire deformation stuff they're talking about, it's really just a model that add some more input parameters (G-forces and maybe wheel angle) to the regular ones (spring compression and wheel momentum) into a new non-linear model that determines friction at each road contact point.

So what they're saying is that they took some measurement on these bumps. They could well have different friction characteristics from the rest of the track due to accumulated rubber or more wear.
 
But whatever measurements you get from one car and its handling won't be applicable to a different car. Any parameterization will be a 'hack' to recreate a best-guess at how a car would respond. Now I can accept that makes a better game, but it'd also be less of a true simulation.
 
That's all a simulator is doing really—guessing at what could have happened in the real world. But whereas the model is a simplistic view of reality, empirical data is not. Making an educated guess based upon fact is what simulation is all about.

And this is not unfamiliar territory. Before it delved into software development, the SimBin mod team had become renowned for doing this very thing: calibrating simple models with the results of an infinitely more complicated reality.

Ian is bringing this expertise to the table.
 
Mintmaster said:
Who says they're measuring with one car, or even with a car at all?
Click to expand...
Hmm, I'm clearly missing the idea here. Whatever measurements they take, how are they applicable to different vehicles? Let's say they take a corner at 50 mph in a Toyota Yaris and hit that 5 mm bump. How does it affect the car? Let's say they measure the effect. What about when the Yaris is going 70 over the same bump? Or a GT40? At 140 mph? Or a Lotus Exige?

You can't measure and apply every force on every car in every circumstance. The idea of a simulation is to mathematically model the forces that affect the vehicle's internals so the model (chassis position, ultimately) responds to the virtual world the same as the real car in the real world in all cirmcumstances. Knowing 'how a corner feels' doesn't fit into any mathematical model.

I can agree that we don't have the tech to model tyre deformation over bumps at different temperatures of different composites etc. and we are limited to approximation hacks, but I don't see what data they can get from gaining race telemetry that will help create a more realistic model. It'll be too tied to whatever platform the telemetery is taken from.
 
I wonder if it's similar to T10's 'tweaks'..

For example, they get race car drivers that do lots of miles on the circuit they are modelling, and he drives the game, and comments about the difference to the real thing in terms of bumps, camber changes etc that you really 'feel' when driving on the circuit, then they add these to the game..

If for NFS Shift, they used race telemetry data to show where the main bumps are and look for common events and car behaviours at a particular part of the track, perhaps they just fudge in these nuances in a similar fashion, and so the track may feel more authentic without the complex physics, this would assume a lot of approximations applied generically to all cars, but I think it only adds to the game, so it's acceptable.
 
Shifty Geezer said:
I can agree that we don't have the tech to model tyre deformation over bumps at different temperatures of different composites etc. and we are limited to approximation hacks, but I don't see what data they can get from gaining race telemetry that will help create a more realistic model. It'll be too tied to whatever platform the telemetery is taken from.
Click to expand...

If I read the interview correctly, they are trying to determine the "nuances of a track's character" using satellite imagery/GPS data, onboard cameras, AND race telemetry.

How this information characterizes a circuit (or makes a caricature of it:smile2:) is anyone's guess. What we do know is that, ultimately, it helps make their real-time simulation more believable.

You can't measure and apply every force on every car in every circumstance. The idea of a simulation is to mathematically model the forces that affect the vehicle's internals so the model (chassis position, ultimately) responds to the virtual world the same as the real car in the real world in all cirmcumstances. Knowing 'how a corner feels' doesn't fit into any mathematical model.
Click to expand...

If reality were a book, a real-time simulation would be an incredibly brief summary. Non-real-time systems, models, are the ever-growing population of chapters and subchapters, each becoming increasingly specific and voluminous. Of course models are better guessers, but they are too impractical for real-time interaction.

A model may begin to pick up on the psychology of driving, which is all Criterion Games—err, Simply Mad Studios—is really after. And the easiest way is to 'hack' emotion into a simulation rather than to try to squeeze it out of a bunch of interwoven models.

"There is a huge difference between what happens in the real world and what happens in video games—even the most 'realistic' simulator has to bend real world physics to make the game more fun," [Gareth Wilson, Bizarre Creations’ Design Manager] said.
Click to expand...

Source: BBC News
 
I don't know if this is the correct thread for this but I hear Criterion are doing the next Need For Speed and it's going to see a return to the original test drive/need 4 speed format.
 
Shifty Geezer said:
Hmm, I'm clearly missing the idea here. Whatever measurements they take, how are they applicable to different vehicles?
Click to expand...
You change the parameters in your various models so that the simulations match reality with a few test drives. Even better, you can also have a jig with an interchangeable wheel and some force meters.

The basic laws of car motion are easy to simulate with 10 year old computers in realtime. The more complicated stuff is when you want to accurately model how the road applies forces to the car. This is usually done with some models with have some parameters associated with the curves that take, for example, wheel slip and give you normal/friction force ratio for one particular type of road surface and tire combo. You then have different parameters to fit this same model to different cars. Your adjustments to a few cars based on some real tests can be applied to the parameters for all cars if done correctly.

This bump you speak of now adds a couple new time dependent parameters. Alternatively, it may be found to affect the forces on the car identically to a smaller bump with a bigger footprint, so you process your track elevation accordingly.

Anyway, I think a lot of this is marketing and hype. A lot of things that you can feel while driving a real car holding a real steering wheel are useless to simulate in a video game.
 
It's important to drive the tracks and get a feel of what the diff bumps actually do. There might be similar bumps and dips but their effects could be completely different based on the location in the corner, angle(s) of approach, elevation (uphill or downhill corner), flat or compression dip inducing? All these things need to be learned from actually driving the track.

Here's a good example. I raced bikes and have drive around Sears Point/Infenion Raceway. Going up the into the 3,3A complex there is a compression dip. This dip gives you the same compression effect regardless of a being on a bike or car! In contrast, there is a similar dip at thunderhill Turn 5. However, since you're now going downhill, it behaves completely different.

The effects of imperfections on a track are best determined by driving/riding feedback + telemitry. Not some theoretical computer models.
 
Speaking of Laguna Seca ... :LOL:

I shot out like a sprinter, zipping around Turn 2, hugging the inside of the track into Turn 3.


The bicycle they loaned me was amazing. Coming out of Turn 4, I began to feel a slight incline as I motored my way into Turn 5. No big deal. Pedal to the medal. Let's knock this out.


Then the unexpected hit me before I could look up. Coming out of Turn 6, it became evident in a heartbeat that this experience was not going to be that simple.


I'm not sure what gear I was in, but it wasn't in low.


"Did you forget that you have a gear shift on the left side of the handlebars as well?" I was asked.


But during this torturous climb, my mind was wandering. I wasn't going to stop, but I wanted to. I spent as much time climbing a quarter mile as I did covering the remaining 1.8 miles, or so it seemed.


I didn't expect to climb the hill at the speed of sound. But I didn't expect each push of the pedal to feel like it was going to be my last.


When it became evident that I wasn't going to be able to figure out how to work the gears, the joy of being on the Laguna Seca track waned and surviving the hill became my main concern.


"There's a reward at the top," I kept telling myself. I could see bikers ahead of me disappearing as if they were falling off the hill. But they were actually going through the Corkscrew. I thought, "Will I live long enough to feel that sensation?"


By the time I hit Turns 7 and 8, my legs were numb. The damage to them will last into the night, but I headed down the hill through Turn 9, flying on two narrow tires. It did cross my mind that if I blow out a tire, I'm going to look like a crash-test dummy rolling around on the pavement.


As these fancily dressed cyclists kept flying by me with no fear, I thought, "Am I racing in the Tour de France or what I imagined was a media-only bicycle ride around Laguna Seca?"


As I came around Turn 11 and towards the pits, I wondered, "Will I be able to even get off this bike and walk?"


I can only remember this feeling one other time and that's when I walked up 221 feet worth of stairs at Bunker Hill Monument in Boston on a 90-degree day.


This was definitely a once-in-a-lifetime experience. Would I do it again? Absolutely. Just not today.
Click to expand...

Source: Monterey Herald
 
You must log in or register to reply here.
Back
Top