Smoke Generation

[McGherkin]

Well, the smoke generation in LFS starts with teeny particles at the contact patch... In reality a lot of the smoke gets sucked up and fills the wheelarch. Especially when travelling slowly. So, why not have smoke generate (at roughly the size of the wheel) at the front of the wheelarch and drift backwards a bit? Of course, if the car is reversing then it would be the opposite, but I think it'd be a much better solution. It'd look like the wheelarches are filling, and probably look a lot better, especially when drifting.

e.g. the smoke is generated 1 wheel radius ahead of the contact patch (or behind it in reverse), at the size of the wheel diameter. Either way, it wafts towards the contact patch (even when the car is stationary).

I think it'd be a lot more realistic. What do you lot think?

[Flame CZE]

It's a lot more difficult to make it than to say it :)

[McGherkin]

Hmm, true. But it shouldn't take up much system resources for a huge improvement in the smoke detail... And surely it can't be that hard to program, since the smoke can be wafted by the wind...

The only other way you can really do it is to have hundreds of tiny clouds of smoke forming around the tyre circumference - probably even worse to program and it'd kill the computer that has to run it.

[FPVaaron]

Hmm, true. But it shouldn't take up much system resources for a huge improvement in the smoke detail... And surely it can't be that hard to program, since the smoke can be wafted by the wind...

The only other way you can really do it is to have hundreds of tiny clouds of smoke forming around the tyre circumference - probably even worse to program and it'd kill the computer that has to run it.


The problem I think is that the tires only smoke on the contact patch with the road, if it were to pan out a bit and smoke by the heat of the rubber rather than the friction point alone then it would fill the guards.

[Whiskey]

You guys are missing some important points here.

1) LFS smokes generation has been improved in 0.6 version. Now it generates smoke in each point above certain temperature, not only in contact patch.

2) LFS has no wind. There is only a variable to change the top speed and aerodinamics. There is no way to interfere with smoke (at LFS's actual state), dust, or even sideways cars.

3) LFS crash/body physics are like a shell. Only the bodywork has a colision detection system (not that this is hard to change, but maybe not trivial because its consequences).

[FPVaaron]

Perhaps making the smoke less transparent would be a quick fix?

[morpha]

2) LFS has no wind. There is only a variable to change the top speed and aerodinamics. There is no way to interfere with smoke (at LFS's actual state), dust, or even sideways cars.

That's incorrect, wind does affect particles (smoke, dust), wind mills and flags.

[Flame CZE]

2) LFS has no wind. There is only a variable to change the top speed and aerodinamics. There is no way to interfere with smoke (at LFS's actual state), dust, or even sideways cars.I disagree with you in this one.

For example, if you do a burnout with high wind on, you see the smoke moving around according to the wind direction and speed.

[McGherkin]

Yes, the smoke is generated at the cintact patch, but surely it is only located by a set of coordinates.


I don't see what the collision models have to do with it. The smoke is not interacting with the car, it is simply being generated at a different point and size.

Whiskey, please go do a wheelspin in LFS and tell me where the smoke comes from.

[Flame CZE]

Whiskey, please go do a wheelspin in LFS and tell me where the smoke comes from.He said "it has been improved in 0.6 version", he meant the next patch which is not out yet.

[McGherkin]

Oh, ok.

[hazaky]

GTA IV has a very nicely done and simple smoke generation in my opinion. :shrug:

[Whiskey]

I disagree with you in this one.

For example, if you do a burnout with high wind on, you see the smoke moving around according to the wind direction and speed.

Ok, wind moves smoke/dust, but only in the same way it affects top speed.
Wind, or cars draft, does not affect the shape of the smoke cloud.


@McGherkin: if you want the smoke to be trapped in the wheelarch you need the smoke to colide with it, as simple as that.

[Lible]

GTA IV has a very nicely done and simple smoke generation in my opinion. :shrug:Yeah, and it's so stable :nod:

[McGherkin]

Ok, wind moves smoke/dust, but only in the same way it affects top speed.
Wind, or cars draft, does not affect the shape of the smoke cloud.


@McGherkin: if you want the smoke to be trapped in the wheelarch you need the smoke to colide with it, as simple as that.


Or the smoke is just generated larger than the wheelarch. I'll mock up a pic.

[McGherkin]

Edited post.

[Whiskey]

But if you are doing donuts, the outside wheelarch should be full of smoke, and with that simplified system it won't show any smoke, because it will be inside the car. That's why I say a smoke collision system would be needed to achieve something plausible.

Plus, generating such a big 'puff' of smoke will only lead to unrealist smoke in all other scenarios (blocking a tire under braking, a really hot tire...)

[Mountaindewzilla]

Fortunately for the perverts that actually want to see smoke coming from their tires, Scawen has already mentioned improving the smoke.

I imagine that the next patch will have smoke coming from every burning hot section of tire, whether or not that section is touching the ground.

I don't think that the computing power required to handle smoke collisions is ubiquitous among the LFS userbase. :(
Once we have smoke collisions, we'll naturally want sophisticated fluid dynamics :tilt:

If collisions were to get attention, I imagine most LFS users would prefer that car-launcher-barriers were fixed.

As someone that spends most of their time in LFS drifting, I can certainly relate to people that want the smoke improved, but there are a lot of other things that I would like to see first.

[McGherkin]

Well, if the smoke is going to be generated at all the hot points of the tyre, I'm going to need a new computer.

And the outside wheelarch will fill with smoke because the smoke is being formed on the wheel, which is farther from the centre than the back of the wheelarch. You'd be able to see the face of the wheel, and a bit of tread, then the rest of the wheelarch would be full of smoke which would continue into the car. I think another diagram would illustrate, but since I don't have access to a computer, it probably won't be until tomorrow.

A really hot tyre creates smoke across all the hot sections, which is usually the tyre as a whole, not to mention that smoke rises by convection into the wheelarch. Also, when braking, smoke is sucked up into the wheelarches by virtue of the air in there being relatively calm. So, it isn't that unrealistic IMO.

[Skagen]

The only "issue" I have with the smoke in LFS is the fact it takes way too long to form when using slicks. I.e, watching this clip (http://www.youtube.com/watch?v=Y33r41_1jjA), you will se the generation of smoke is more or less instantainiously when the wheels locks up. And the thickness of the smoke is way less progressive than it is in LFS.

I actually believe realistic smoke is an important part of the simulator, as it's often a visual cue that the guy in front of you either is breaking harder than normal, or he's in trouble and might be going off the track. :)

[Mr_Lonely]

another note is that the smoke generator should generate random shape of the smoke,

at this stage lfs smoke just look like an big cone after the car..

real smoke is more alive and gets more "shape'y'"..

[Breizh]

http://www.youtube.com/watch?v=H7QMzHrD8pM

[Flame CZE]

http://www.youtube.com/watch?v=H7QMzHrD8pM
Isn't it only for nVidia cards with PhysX? I'm not sure.

[hazaky]

Isn't it only for nVidia cards with PhysX? I'm not sure.
If there's no PhysX acceleration, its done by CPU afaik.

[Skagen]

It does not have to be a turbulance simulation, and certainly not so computational complex it needs third party hardware (or a very fast CPU) to yield good results.

As long the smoke is efficient and somewhat accurate in terms of conditions for generating smoke, Im more than pleased. :)

[Breizh]

Isn't it only for nVidia cards with PhysX? I'm not sure.

I have no idea either, and not arguing either way. Just a clear visual of smoke better than LFS' so people can easily refer to it when they argue for relatively more or less realistic smoke.