Is this the most realistic driving/racing sim according to some of you guys?

Here are some reasons they use Iracing and RF2 Because beams FFB is absolute garbage at the moment and there is almost no easy way to set up AI to race 200 laps without crashing in the first few.
Then you got the problem with the graphics being unconvincing. And the fact that car mods take forever to make for Beam while Iracing and RF2 just supports them from default well at least not RF2 because you need to pay for most of them. sorry for bad grammar I am really tired right now.

 

You must've tried really hard to find a reason to be negative about almost everything other than what the thread was about.

 

I was not trying to be negative I was just trying to give the thread poster some examples as to why they use Iracing and RF2

 

When you first sentence is calling indie teams work "garbage" it's hard to tell.

Anyways, iRacing has only recently made AI and I don't know what rF's situation is.
Recently when iRacing was used for virtual events, I heard that lots of drivers got really frustrated over some tire model oddities.

Graphics also aren't really a comparison point because if something does better at simulating than being visually appealing it's still a better choice. For example, if you look at professional flight simulators, the graphics are pretty basic.

iRacing and RF2 vehicles, while you don't have to physically set up the entire body, I'd imagine there's still tons of work that goes into making them actually realistic. I mean, just see how much needs to be simulated:

 

I was not saying any of the other games are better than Beam And I am not saying the indie teams work is garbage its the people who made the engine that was not aware of the bad FFB as it was not really made for racing games. At least as far as I know And RF2 Did not just recently make AI it has literally had it since RF1 that was released some time in the late 2000s.
Edit Also I did not say that the other games vehicles are not hard to make its just that they already have them by now and Beam does not so it would take to long to make all of them. And I was not saying that The graphics for a sim game that is shown on TV for sports have to be amazing. I meant they would need to look atleast realistic and sharp and those are two things that while Beam does sort of do but not as good as RF2 and Iracing thanks to the game engine being out of date with graphics Just like it is with FFB.

 

Those reason are not false, BeamNG is so realistic in its operation that it can not be accurately simulated by the limits of today's consumer hardware, if the Dev team lifted the limits imposed by our relativity slow as a snail consumer hardware (Yes, including the new 10900k and 7980XE ), I imagine it would surpass iRacing and RF2 in realism but then you'd need a chip nearly equivalent to the Guinness World Record holder " 8.723 GHz AMD Piledriver-based FX-8370 chip" with the parallelization of the "Ampere™ Altra™ offers up to 80 cores" but once hardware catches up to BeamNG in a few years it will become more relevant to professionals, developers and enthusiast alike and may very well be the engine of choice for future racing games, its only achilles heel is that its too "ahead of its time".

 

Based off of what I am reading from your comment I might be wrong again I am very tired right now. You are saying the game engine just needs to be changed a bit to improve it and I was not saying that it did not I think that is all it really needed I was trying to say that with a few updates to the engine it will be a lot more realistic than all the other sim games and so on.

 

*cough* Crysis *cough cough*

 

When you say "it needs to be changed a bit to improve" it sound like you want "downgrades to accommodate today's slower by comparison hardware. To run like iRacing and FR2 nodes and beams would have to be severely limited or removed altogether, in fact you can currently turn off collision in the settings and your fps go way up. I think that would be going backwards, the Devs would be better served continuing with the way BeamNG is now with incremental improvements and waiting for the future to catch up.

To clear up any confusion, BeamNG is not poorly optimized, it is simply optimized to run in a revolutionary way that current hardware can't handle.

 

I did not mean they needed to be done now or soon that is for them to do. And they can do it when they want it done. And The slow update style they have now actually is good It makes you want to keep playing if you are say bored of the game to see what the next big update will be. And the time that you spend waiting you end up finding cool things to do in beam and it becomes really fun again. If it did not have the slow update system the game would probably lose popularity really quickly.

 

It simulates in a much more realistic way than anything else so i would say it is the most realistic

 

That is true, this game is really not a game yet, its more of a early release sandbox type of game, It could be compared to Photoshop or 3DsMax as it is not really a game but more of a proof of concept with game elements. It can get boring as it is.

 

The thing is, is that the more powerful a standard computer gets, the more detail the dev's can pump into each vehicle. Since the game engine is 100% physics based and not based on a physical fake mathematical algorithm that makes things look and feel real... but are actually, computationally, real... the faster the average computer is, the more nodes and beams the devs get to play with.

The more nodes and beams the devs get to play with, the more accurately they can represent a true life vehicle... and the more accurately they can represent that, the better the driving dynamics will be.

It really is that simple.

If you look at a current in game tire... its made out of little nodes as pointed out earlier. This basically ends up making each of those nodes have to represent an entire chunk of the tire. This can be complicated as tires are complicated to describe correctly mathematically.

Now... imagine that we could double the node count in the tires... thus doubling the nodes in concat with the ground at any one time... and thus giving each node less tire to have to simulate...

Basically we are doing finite element analysis for each node, and the more nodes we have, the smaller the element size becomes... which in turn ends up increasing our accuracy.

Take this far enough and you end up simulating so many points on a tire that the tire basically becomes indistinguishable from the real thing... but how many points does something like that take?

My guess is to get to a point where we wouldn't be able to tell the difference between most simulations and beam, we would have to quadruple the amount of nodes in a tire... thats just a guess though, so don't quote me on that... but thats all just to more accurately simulate a tire. Now each node in a tire adds roughly (I think) 5 beams at the moment... if you quadruple the node count... then you would in turn... have to... take whatever that node count is... and multiply it by 5... ouch...

So, at the moment, to get around this, the devs have been hard at work for years to come up with ways to simulate tires with this limited amount of nodes... and by golly have they gotten really darn close. The issue is that last little bit... as 20% of the results (a full 100% tire model) requires 80% of the effort. Its easy to get close... but much MUCH harder to get it exact.

To add to this... the devs are creating a tire model that the world literally hasn't seen before (well... ok... they have, but only in RoR... and that wasn't nearly as developed) from the ground up. The math for a full blown tire model that just simulates one contact point on the ground has been around for decades now and been steadily improved and rehashed hundreds of times since... Beam's tire model though... it started here... for the first time... with them.

So... it may seem like they are taking forever to get it right or come out with everything or whatever... but remember... when you hear people like me saying, "It takes a while because they are building it from the ground up"... we aren't joking. They have had to literally take the rule book, rip it up, and rewrite the entire thing.

That my friends... is gonna take time... so be patient... greatness is coming.

 

add to that that more nodes=less weight per node, means you're pushing closer to instability with the 2000Hz physics engine.
I wonder if, at some point wayy down the track, they might overhaul it to work at, say, 4000Hz?

 

That is a good point as well... 4000Hz seems like it would be almost impossible, but then again, when RoR first came out and was running at 2000Hz on what nowadays seems like compleat potatoes, so who really knows what's achievable.

I also have no idea if some of their code would be affected by such a jump... like does it rely on knowing that the calculations will be done at 2000Hz, or can it scale with speed?

My guess is the collision code would not like a frequency change... but I also would bet that a better collision system could be developed with higher frequencies...

but then again this is all just speculation...

 

4000hz or calculations twice as often would require more than the power as this doesn't scale linearly

 

I'd say the handling has been feeling quite realistic lately, but it doesn't translate to the average end user so much due to a lack of steering aids.

 

With this engine, one can give water and snow textures, and make buildings and sand act realisticly, but there is no computer that can handle that.

 

Guys, all of your problems with the handling are litteraly answered here

 

just curious, what would happen if they tried to run BeamNG at 4000hz, no computer could run it?