Jump to content
DIRT 5 - Chat about the game and get support here. Read more... ×
Christmas Period - Codemasters Staff and Support Read more... ×

Archived

This topic is now archived and is closed to further replies.

[DiRT 4] Constructive feedback on the subject of car-feel and physics: The Ultimate Thread

Recommended Posts

fab1701 said:
KevM said:
In-game 'G' doesn't exist.  It's just numbers on a screen.  Comparing real telemetry to in-game telemetry is like arguing that Donald Trump weighs more than Homer Simpson!
Technically you're right, but you have to compare it, because only then it's possible to figure out how good a model of the real world, in this case a simulation, actually is. If there is a difference, even if it's very tiny, then it's the reason to ask 'why is that?' and try and improve it.
But the in-game G Meter can display any number programmed, so you could have the G Meter reading identically to real-life and change nothing about the game handling itself, just by applying a multiplier 


Share this post


Link to post
Share on other sites
KevM said:
fab1701 said:
KevM said:
In-game 'G' doesn't exist.  It's just numbers on a screen.  Comparing real telemetry to in-game telemetry is like arguing that Donald Trump weighs more than Homer Simpson!
Technically you're right, but you have to compare it, because only then it's possible to figure out how good a model of the real world, in this case a simulation, actually is. If there is a difference, even if it's very tiny, then it's the reason to ask 'why is that?' and try and improve it.
But the in-game G Meter can display any number programmed, so you could have the G Meter reading identically to real-life and change nothing about the game handling itself, just by applying a multiplier 


EDIT: g-forces shown are NOT the forces the telemetry transmits!

Yes it could, but braking from 100km/h to 0km/h with the impreza 2001 on gravel at dirtfish (right besides that big stack of wood) shows that the required time to full stop is about 2.4s and the average g-force in that timeframe is around -1.18g. That fits perfectly:

-27.77m/s (= -100km/h, because we're decelerating in Z) : 2.4s = -11.57m/s²

-11.57m/s² : 9.81m/s² = -1.18 g

Here is the data:
https://www.dropbox.com/s/nvoej9qt2z2fhq/Braking_Test.txt?dl=0

And visualized ;):


EDIT: g-forces come from velocity-data!!

Share this post


Link to post
Share on other sites
fab1701 said:
KevM said:
fab1701 said:
KevM said:
In-game 'G' doesn't exist.  It's just numbers on a screen.  Comparing real telemetry to in-game telemetry is like arguing that Donald Trump weighs more than Homer Simpson!
Technically you're right, but you have to compare it, because only then it's possible to figure out how good a model of the real world, in this case a simulation, actually is. If there is a difference, even if it's very tiny, then it's the reason to ask 'why is that?' and try and improve it.
But the in-game G Meter can display any number programmed, so you could have the G Meter reading identically to real-life and change nothing about the game handling itself, just by applying a multiplier 


Yes it could, but braking from 100km/h to 0km/h with the impreza 2001 on gravel at dirtfish (right besides that big stack of wood) shows that the required time to full stop is about 2.4s and the average g-force in that timeframe is around -1.18g. That fits perfectly:

-27.77m/s (= -100km/h, because we're decelerating in Z) : 2.4s = -11.57m/s²

-11.57m/s² : 9.81m/s² = -1.18 g

Here is the data:
https://www.dropbox.com/s/nvoej9qat2z2fhq/Braking_Test.txt?dl=0

And visualized ;):


Can you do the same test on dry gravel out on the stages? 

Share this post


Link to post
Share on other sites
That's fine, but what actually makes it 100-0kmph apart from a number on the corner of the screen?

We can't measure actual speed, distance, friction, weight etc.  It's all virtual.  It's nice that the numbers calculate, but does the car literally travel enough car lengths of itself (known length) in a certain time, to equate to the speed & distance displayed on the speedo?  No way of checking really

Share this post


Link to post
Share on other sites
KevM said:
 It's nice that the numbers calculate, but does the car literally travel enough car lengths of itself (known length) in a certain time, to equate to the speed & distance displayed on the speedo?  No way of checking really
Game is numbers. So... I don't even have words to describe what you just said. 

How do you think sophisticated F1 simulators are done? Hint: numbers, numbers and calculations. 

P.S. This is based on UDP data and not speedo. It's different. 

Share this post


Link to post
Share on other sites
versedi said:
KevM said:
 It's nice that the numbers calculate, but does the car literally travel enough car lengths of itself (known length) in a certain time, to equate to the speed & distance displayed on the speedo?  No way of checking really
Game is numbers. So... I don't even have words to describe what you just said. 

How do you think sophisticated F1 simulators are done? Hint: numbers, numbers and calculations. 

P.S. This is based on UDP data and not speedo. It's different. 
What he meant is what I was wondering one page back- if game calculates that car traveled 10 meters, you could as well change parameters to calculate for it to travel 100 meters. That's why he mentioned we'd need to calculate car lenght because if a car model is accurate and proportional you could use that to calculate actual distance in game. Not a number. But I don't consider that an issue any more, because fab1701 already verified that ingame distance is consistent with car weight length. So car- enviroment scale is accurate, therefore ingame 10 meters are really 10 meters. No less, no more. Otherwise other measurements such as speed would be off.

Share this post


Link to post
Share on other sites
versedi said:
KevM said:
 It's nice that the numbers calculate, but does the car literally travel enough car lengths of itself (known length) in a certain time, to equate to the speed & distance displayed on the speedo?  No way of checking really
Game is numbers. So... I don't even have words to describe what you just said. 

How do you think sophisticated F1 simulators are done? Hint: numbers. 
You are thinking far too linearly to get what I am saying.  

PS on your edit:  F1 simulators use real measurements of real cars, real distance & real track lengths.  It's numbers based on real world data.

How do you measure a generated stage accurately lad?  The only thing that's accurate & to real-life scale, is the car model we chose in the menu. Do you think Codemasters have laid a car model they know the length of, bumper-to-bumper the entire length of the track & then multiplied the length of the car by the amount of cars it takes to cover the track distance & then calibrated the in-game speedo exactly to those finite values?  Maybe they have, but I very-much doubt it.

Let me try and explain....
In the game, on gravel I can literally brake from flat in 6th, for a hairpin, when I can see the cone sitting on the hairpin apex.  Mathematics say the G Meter is reading fairly close to the mark, but I'm not sure the actual scale of distance in Dirt4 is accurate 

Share this post


Link to post
Share on other sites
Turns out, I made a mistake and used the wrong data-set from a different test. The g-forces presented are values calculated from the velocity-data, because the forces in Y-axis aren't transmitted. Will get this sorted, but it shows, that it's relatively easy to get realistic g-forces just by the cars movement data still... :D

Share this post


Link to post
Share on other sites
With som minor tweaks the NR4 felt pretty nice in the Swedish Weekly. This car remains one of the better feeling cars in the game.

SS4 was actually really enjoyable. Love the sky as well! 

https://youtu.be/7Xrl-3m1c8Q

Share this post


Link to post
Share on other sites
fab1701 said:
Here we go, right values, same result ;) (sorry about the confusion, got lost in my code somewhere :D ):

This time 100km/h to 0km/h in 2.5s, average g-force = -1.13 g.

-27.77m/s² : 2.5s = -11.11m/s²

-11.11m/s² : 9.81m/s² = -1.13 g

Data:
https://www.dropbox.com/s/nvoej9qat2z2fhq/Braking_Test.txt?dl=0



I made the same test without an ABS and the braking force set to the max:
https://docs.google.com/spreadsheets/d/1XC1YztS0woBNymABVWSAckRUlptr8uZY4UxKRaRBCn0/pubhtml

With the highest braking force the car didn't manage to keep the wheel speed at 0 km/h. The braking distance would be shorter with better brakes.


Share this post


Link to post
Share on other sites
hawku0 said:

I made the same test without an ABS and the braking force set to the max:
https://docs.google.com/spreadsheets/d/1XC1YztS0woBNymABVWSAckRUlptr8uZY4UxKRaRBCn0/pubhtml

With the highest braking force the car didn't manage to keep the wheel speed at 0 km/h. The braking distance would be shorter with better brakes.


Very cool! This is even more sophisticated!

Share this post


Link to post
Share on other sites
hawku0 said:
fab1701 said:
Here we go, right values, same result ;) (sorry about the confusion, got lost in my code somewhere :D ):

This time 100km/h to 0km/h in 2.5s, average g-force = -1.13 g.

-27.77m/s² : 2.5s = -11.11m/s²

-11.11m/s² : 9.81m/s² = -1.13 g

Data:
https://www.dropbox.com/s/nvoej9qat2z2fhq/Braking_Test.txt?dl=0



I made the same test without an ABS and the braking force set to the max:
https://docs.google.com/spreadsheets/d/1XC1YztS0woBNymABVWSAckRUlptr8uZY4UxKRaRBCn0/pubhtml

With the highest braking force the car didn't manage to keep the wheel speed at 0 km/h. The braking distance would be shorter with better brakes.


Surely that's a clear indication that their is waaaay too much grip?  If you can't lock a wheel on gravel with max brake force, the grip/friction/downforce is much too high?

Share this post


Link to post
Share on other sites
KevM said:
hawku0 said:
fab1701 said:
Here we go, right values, same result ;) (sorry about the confusion, got lost in my code somewhere :D ):

This time 100km/h to 0km/h in 2.5s, average g-force = -1.13 g.

-27.77m/s² : 2.5s = -11.11m/s²

-11.11m/s² : 9.81m/s² = -1.13 g

Data:
https://www.dropbox.com/s/nvoej9qat2z2fhq/Braking_Test.txt?dl=0



I made the same test without an ABS and the braking force set to the max:
https://docs.google.com/spreadsheets/d/1XC1YztS0woBNymABVWSAckRUlptr8uZY4UxKRaRBCn0/pubhtml

With the highest braking force the car didn't manage to keep the wheel speed at 0 km/h. The braking distance would be shorter with better brakes.


Surely that's a clear indication that their is waaaay too much grip?  If you can't lock a wheel on gravel with max brake force, the grip/friction/downforce is much too high?
I wasn't sure if I understood that post you quoted, but what you wrote was first think that came to my mind.

Share this post


Link to post
Share on other sites
hawku0 said:
fab1701 said:
Here we go, right values, same result ;) (sorry about the confusion, got lost in my code somewhere :D ):

This time 100km/h to 0km/h in 2.5s, average g-force = -1.13 g.

-27.77m/s² : 2.5s = -11.11m/s²

-11.11m/s² : 9.81m/s² = -1.13 g

Data:
https://www.dropbox.com/s/nvoej9qat2z2fhq/Braking_Test.txt?dl=0



I made the same test without an ABS and the braking force set to the max:
https://docs.google.com/spreadsheets/d/1XC1YztS0woBNymABVWSAckRUlptr8uZY4UxKRaRBCn0/pubhtml

With the highest braking force the car didn't manage to keep the wheel speed at 0 km/h. The braking distance would be shorter with better brakes.


In your tests you're getting some lateral forces which might be helping slow the car more because it goes sideways. Can you do a brake test where the car brakes in a straight line and the g forces are only forward?

Share this post


Link to post
Share on other sites

KevM said:
hawku0 said:
fab1701 said:
Here we go, right values, same result ;) (sorry about the confusion, got lost in my code somewhere :D ):

This time 100km/h to 0km/h in 2.5s, average g-force = -1.13 g.

-27.77m/s² : 2.5s = -11.11m/s²

-11.11m/s² : 9.81m/s² = -1.13 g

Data:
https://www.dropbox.com/s/nvoej9qat2z2fhq/Braking_Test.txt?dl=0



I made the same test without an ABS and the braking force set to the max:
https://docs.google.com/spreadsheets/d/1XC1YztS0woBNymABVWSAckRUlptr8uZY4UxKRaRBCn0/pubhtml

With the highest braking force the car didn't manage to keep the wheel speed at 0 km/h. The braking distance would be shorter with better brakes.


Surely that's a clear indication that their is waaaay too much grip?  If you can't lock a wheel on gravel with max brake force, the grip/friction/downforce is much too high?
But people, have said that they can lock wheels in dirt 4. I think I locked wheels myself a few times.

Share this post


Link to post
Share on other sites
Ok, so arguing against well presented graphically represented actual UDP data, you 'think' you locked your wheels?

Okaaaay.... 

:)

Share this post


Link to post
Share on other sites
Long story short if wheels cannot be locked on braking, then no wonder few cars have problem with engine chocking on taking off the start line, or powersliding corners (like square on junctions). I hardly ever watch my own replays, so I never actually was paying attention if wheels are locked visually.

Share this post


Link to post
Share on other sites
KevM said:
Ok, so arguing against well presented graphically represented actual UDP data, you 'think' you locked your wheels?

Okaaaay.... 

:)
I Ve only looked at the the last few posts so I'm not aware if someone proved that it's impossible to lock wheels. Can you show me please. I will have another look. 

Share this post


Link to post
Share on other sites
The wheels sure audibly and visually lock the wheels. And I guess the car gets unsettled, you can feel that.
Maybe they don't actually lock in the mathematics as they should.

Share this post


Link to post
Share on other sites
I am just looking through YT replay stuff, and I cannot see wheels locking other than with handbrake. If I find anything else I will let you know guys.

Share this post


Link to post
Share on other sites
I just tested this in Michigan on a short stage. You can lock your wheels, you feel the loss of contorl and you can even hear the car sliding when the brakes are locked. 
You can see the effect in the replay, but something strange happens if you are still at speed and lock the wheels up. They stop spinning overall but rotate incrementally, as if  they gripped to the surface and the brakes are just not enough to stop the rotation completely. No video unfortunately but easily repeatable.

Share this post


Link to post
Share on other sites
So they cannot lock at high speed, they just rotate slowly? That confirms forward grip is too high. 

Share this post


Link to post
Share on other sites
kobeshow said:
I just tested this in Michigan on a short stage. You can lock your wheels, you feel the loss of contorl and you can even hear the car sliding when the brakes are locked. 
You can see the effect in the replay, but something strange happens if you are still at speed and lock the wheels up. They stop spinning overall but rotate incrementally, as if  they gripped to the surface and the brakes are just not enough to stop the rotation completely. No video unfortunately but easily repeatable.
By rotate incrementally do you mean like where ABS's stop/start rotation to try and gain traction?  

Share this post


Link to post
Share on other sites

×