Beamng vs reality: Real world personal experiences

Hey guys and girls, I know that currently there seems to be two camps on these forums;Those who love BeamNG Drive and those who are quite critical of its physics simulation.

As a bit of a personal experiment to iron out some of these misconceptions and opinions posed as fact, I thought i would make some comparison videos, using only my own footage so I can only argue what I can prove and have experienced.

I hope you enjoy

 

Ok, now go drive that littel Citroen at 3mph in a parkinglot and watch it bounce around on the 12 segment tires... No, wait....

Ok, try this instead: Go drive your little Citroen with the same jerky wheel movements you do in the game. Please do this in a big parkinglot though, because you WILL lose control.

And then, drive your little Citroen on a road that is not butter smooth (compared to the lethal potholed hell they call roads on dry rock island) and tell me how it feels....

My point is, you aren't comparing the game to real life, the two are completely, and utterly different circumstances. Look how effortlessly you control the real car and keep it in your lane on a narrow winding road doing 45mph. Go do that even on the wide nice roads that are in the game andtell me how easy it is.

I can make my own "comparison" with a completely different conclusion to yours.

I appreciate the effort, though. I just disagree with pretty much everything you just said.
(before you go on, I have had my license for 20 years and have driven cars similar to the ones in the game both on road and track, and dirt and tarmac rallies and autoslalom) as well as "spiritedly" on public roads.

 

I enjoy your spirited opinion, but theres no need for your derogatory narrative. This is precisely the reason for my post, im intending to produce more videos, showing some of the areas you critisized, so hopefully we can see better how it compares.

At no point do i suggest this is perfect, but it is interesting to compare with reality

 

Not really about the comparison of BeamNG to real life, but were you driving in Cornwall? I went there for vacation last year, and those roads look like the same sort of (barely) 2 lane roads.

 

I'll do a quick video of the little hatchback on the tarmac track on one of the islands, compared to a Golf on a similar track in real life. I just gotta swallow dinner first.

Sorry if my post came across as derogatory, that was not at all my intention!

I love exploring physics in games, and like I posted, I appreciate your effort.

 

Looks like any other British country lane with a 60mph speed limit. I live here and travel everywhere.

 

I appologise for biting, I am only intending to explore the simularities... and differences. It would be great to see others personal opinons, after all that is all I was intending to pose.

Would be interesting to get a few people doing the same in different conditions, i have no access to off roaders so somone willing to strap on a camera while playing outdoors would be interesting. I would love to have had my mr2 on the road and drive the same road but I took her off the road a while ago

The road is in Herefordshire, theres a lot of simularities with the rural roads around here and down there in Cornwall. They are great fun, but you really have to be alert while driving them!

 

On the real world one, The steering wheel is on the wrong side of that car

 

Right is right.... left can never be right

 

I drove on a highway for some time at 85 mph (real life) and then tried that speed in Beamng and it feels realistic enough. Maybe in real life I have a bit more grip, but it's close enough I think.

 

BeamNG gets really floaty at speeds over 35-40 mph. Just trying to go straight is a chore, ulike in a real car where you can go pretty much as fast as you like on a smooth road and still keep control and feel planted. Try it on the grid map. Once up to speed, make some minute steering corrections and watch the car float away...

The steering input needed in BeamNG is also very exaggerated. In a real car on a tight track you are very rarely turning the wheel past 90 degrees, let alone having to swap hands.

In the Op video you can see even at a pretty sharp bend the wheel only briefly pass 90 degree. And at no point in that video is the car anywhere NEAR it limit of grip, where in the game the car is weaving and having to be corrected all the time. Example is at the 6minute mark in the OP video.

Now, I made a little video of the hatchback on the industrial area track. Note the initial slow roll where the car bounce around, and how much I have to work the wheel to even keep the car in a straight line. I drive with a Thrustmaster T500RS wheel set to 1080degrees rotation, and manual gears with clutch.

As reference I post a video of me driving on a gokart track in my Golf.
(The video is still processing at the time of writing. if it doesn't work give it a few minutes.

[MEDIA=youtube]http://youtu.be/OYd0qPEWg8I[/MEDIA]

 

just for continuity sake, the second gameplay section was driven deliberately faster than the first, so i was mostly out of control.

Also I play BeamNG with a ps3 controler, which is much harder to get smooth than a steeringwheel..... im looking to get a wheel so i can eliminate that as a potential additional cause for they crazy weaving steering input.

 

Nice work on the video OP, was interesting to watch and compare.
Being from the US, in the first segment of your actual driving, I was really afraid you were going to crash into the first passing car, being on the opposite side of the road and all! haha

 

haha, yeh that was something I had already considered, I was going to flip my video to make it more visibly comparable.... but I think from your reaction I did the right thing :-D

I just watched through Atledreier's videos and the steering rack looks to be much different rate to the standard version of the car that I used. So I see what you mean about how much steering input is required, but the stock version of the car seems to better match the steering rate of my car hence the reason I chose the standard one over the sport or rally version plus my car has less power and no turbo.

What's also interesting in relation to the cars is the typical vehicles from each of our personal locations. In the UK there's mostly FWD 4 cylinder cars, with turbo versions being the upgrade. We obviously have straight 6, V6 and a variety of large capacity engines being the premium options but I would have thought most of the cars on the UK roads are between 1.0 and 2.5 litre engines. I'm sure this is quite different in other areas around the world, so expectations of vehicle handling and performance may be effected by these differences in experience?

Another thing I haven't checked is how much work has gone into the chassis dynamics and suspension/wheel geometry. One of the original critiques of my OP was in regard to small inputs at high speed having unexpected catastrophic effects on controlling the car. I wonder if the wheel alignment of these cars mimics real world toe, caster, camber etc. and whether the wheel in travel within its sprung range complies to changes in those geometries as would be the case in reality. I suppose this may be down to the initial design of a cars suspension and whether considerations were made at that early stage to fully model suspension systems "correctly" to specifications in the real world?

Late night ponderings that I probably should look into before posing questions. I guess I need to do some research into the debugs and data available!

 

Thanks for this comparison you made phipck. This is a very interesting tread.

That's something I noticed form the first time I played BeamNG (don't have the alpha yet).
Yesterday when I was driving home from work and payed attention on how much I had to turn the steering wheel. Most of the time I had to turn it only a few degrees (45 to 90) to drive through normal curves. For cornering a 90 to 180 degree steering is enough.
The video on top of this topic gives a good representation of this.

At higher speeds, turning the wheel has a lot more impact on your car. Maybe that's the reason for the cars being not easy to keep in a straight line in BeamNG. Every small correction leads to a massive steering turn which brings the car out of control.
For users which are using a keyboard or controller (no steering wheel), it might be an idea to make the controlling of the steering wheel a lot less sensitive as speed increases.

The only time a fast steering is required during higher speeds, is to correct the car from over-steer.
A while ago I followed a sliding course.
All cars where having a bad rear tire on the right and we had to make a left turn on wet tarmac at 50km/h (+/- 30M/h).
I was impressed on how fast you had to turn the wheel to prevent the car form turning around his axis.
Maybe it is an option to foresee a certain button to press on keyboard/controller together with the direction keys for fast steering.

Although the speed at the steering wheel turns, is much too fast.
If I reduce the simulation speed to 40%, then I get some realistic steering (while using the keyboard and 'dynamic steering' enabled).


A summary of ideas to make controlling the cars more realistic and easier to control with keyboard/controller:

• Reducing the speed of the steering-wheel turning in global (40%).
• Make the steering speed lower as driving speed increases.
• Foresee a way to switch to fast steering (by pressing a button together with the steering control)

These ideas are for normal driving, not for racing

 

That's a rather informative wee video, good work.

 

I have access to a camera and I am doing track driving this friday on the track in my video. If anyone want to explore some aspect of car behaviour I might be able to comply. That said, that would be in a light, low cog, rwd car with a LSD, so not comparable to anything in BeamNG at the moment. I also have footage of highspeed driving on track both in the Golf and the sportscar (Toyota GT86) if needed for comparison.

 

Thanks all for the positive input.

I know the spec of your truck isnt the same as one in beamng but it would still be interesting to get real forum members own footage to demonstrate their experience and provide evidence so nobody can shout it down.

It would be great to see (and something im intending doing) footage filmed from simular camera angles to those used in Beamng. So of course drivers seat like i did, but also wheel cam (to see wheel travel) then the same or simular filmed in beamng.

Im intending to make a series of experiments, driving at fixed speeds in a range of circumstances like 10mph speed bump, then 20mph speed bump etc for both Beam and reality with the same camera angles in both. It would be great to see where this engine shines and where things need some tweeking, im not closed minded about the potential to find issues, especially if that leads to a solution from the devs.

Either way it gives me an excuse to make silly videos and pretend im doing something scientific

 

This video has some sequences where yo clearly see wheel travel. Now, bear in mind this is a low car with low weight and low COG, so corner roll is pretty much non-existent. What you see is pretty much all bump and terrain induced travel. Look at 16:15 for instance, here we drive at around 100kph on what I'd consider a little bumpy road.

I have access to a large parkinglot where I can do some experiments if needed, and logging equipment where we can see G-forces, speeds and so on.

[MEDIA=youtube]http://youtu.be/rWnO5Q_h5v4[/MEDIA]

 

In the real world, things like this dont happen unless you have HUGE tires.
http://www.beamng.com/threads/2277-Tire-Issue