Realistic smoke

The texture is a little uniform, but shader would need a lot of work to make it look good with anything else, since they're all being drawn in the same orientation. Maybe replacement of the single smoke frame with an array of them, selecting one at random for each particle. And simulating actual smoke is a huge pain in the behind... I've seen papers on it and they're happy if they can get 4-5 frames per second on high-end CUDA processors. Usually with a static frame of reference and motionless source, convecting the smoke upwards (eg. a cigarette). To put that into Racer, it'd need to simulate the air movement all around the car, on a fine enough grid to look good (probably 0.1m or so), which is a pretty big deal when you consider how large the track is. Though I've seen some work on sub-voxel turbulence before the render pass, so it might be possible to raise that to 0.5m without totally destroying the sense of smoke. Unfortunately I saved all those papers on my laptop which is non-running at the moment.

On the other hand, it would make car aerodynamics easier if we were simulating air movement :p no need to guess at drag coefficients and wing locations, you could measure directly how much air it's pushing out of its way.

Maybe in another 5 years...

They just need to white out when overlaid, but have a bluey tinge so when they are not so dense they look bluey (as per photos)

Then you could have an array of images to cycle through with some turbid smoke animation, that can play through, but have a different start offset for each particle for randomness.

Or just a better base particle texture, it could have more detail in there and still not look repeaty, imo...

FSR's settings are deffo the right way forward for a smokey look/behaviour, just the actual particle imagery itself that needs tweaking.

Dave
 
Then you could have an array of images to cycle through with some turbid smoke animation, that can play through, but have a different start offset for each particle for randomness.

Or just a better base particle texture, it could have more detail in there and still not look repeaty, imo...

Nail. On head. You have hit. Bang.
 
Here is my version...
screenshot002.jpgscreenshot004.jpgscreenshot003.jpg
Alex Forbin
 
Sweet texture! Care to share?

Thanks, it's more than a texture though. You will need to replace your data/render/particles.ini as well.
The settings as they are work fine on my system. Depending on hardware your user experience may vary. ;)


Alex Forbin
 

Attachments

  • 4binsmoke.rar
    30.5 KB · Views: 249
Cool stuff Alex, thanks !

Even better, here's Shift smoke which is also perfect as you will see...
Use the texture with Alex custom ini file...
 

Attachments

  • smoke_fake_lighting.7z
    4.4 KB · Views: 213
Well, allthough I like Alex's smoke, I begin to question the general knowledge of smoke. After 8 years of being involved in drifting as a driver and a judge, I've seen my share of smoke, and even though Alex's smoke looks georgeous in pics, I can't say that I like the way it behaves.

The thing is, smoke is just air. It doesn't move in any direction unless the air does. Being hot it might ascend a little, but usually it just dissolves in the same place it originated. Having velocity makes no sense, neither does having that much texture. To the human eye, smoke is basically a huge diffusor/lightbox. It spreads light everywhere and is basically colorless (there are some variations in color from different tyres though).

Basically the smoke should just be created as small particles, then grow and fade after a while. The thicker the smoke, the longer it takes of course...

As for what could be done to make it look better in Racer, I think Mitch is onto something with the soft particles. The rest is up to us ;)
 
As for what could be done to make it look better in Racer, I think Mitch is onto something with the soft particles. The rest is up to us ;)

yeah, i think it will share his discoveries with us, he sayd that the soft particles doesent affect the frame rates, and this is very good too!
 
I agree, though not entirely.
What Alex's smoke is trying to replicate IMO is what yours is lacking. The smaller vorticies are what make smoke look "real" to most people. It's the mix of large and small particles that your implementation lacks I think...that and shadowing (whatever happened to that smoke implementation anyway?)

Smoke generally does just grow and fade (dissipate) but it's the differing densities that racer's lacking atm.

Note: I haven't tried Alex's particle.ini settings yet...though from the sounds of it they have velocity which I think is incorrect. I'm simply talking about the opacity and shading of the smoke.

Hmmm...might have a play tomorrow and see what I can cook up. Might have to play with fumeFX too - could get some nice results from that...
 
Well, that's where a better smoke texture (more like the one in 050f perhaps?) comes in. I actually have more randomness in size in my particles.ini, but due to the smoke being white, the smaller particles are hidden within the larger ones. At the moment I use mostly my settings but with a modified version of Alex's texture.

My opinion is that Alex's smoke looks more real in pics, but mine behaves more real in game. I'll try to combine them to look real in both :)
 
Hmmm... I don't know if it's the exhaust, or turbulence from the wheels, but some of the smoke seem to have velocity. The rest would probably just be wind.

 
Well, my 2p.

The smoke generation needs to be tied to slip ratio/angle, like the current tyre skid sounds.

Smoke only has velocity if the air it's in has velocity, which is probable if the tyre is spinning really fast at low road speeds (high slip ratio)

We just need more variables passed into the smoke particles to make it look right.

Slip ratio and angle would allow this. Burnout, massive slip ratio, so have velocity based on slip ratio, and high slip ratio means a longer life, so they grow more at their fixed growth rate.
Growth rate is probably fixed as that is just brownian motion vs time I suppose. More time, more disersed. More particles at higher slip r/a, and it looks thicker and is around for longer.
Velocity should damp with time, which with some randomness, would generate nice turbulence as the fast moving small particles are clearly seen moving in the turbulence, but that visual appearance is lost as particles get big relative to the noise scale.


All fairly obvious stuff if you've used things like particles in 3DS Max... all really basic implementations that have been in 3DS Max for over a decade too, so they are not high-end resource hogs.
Drop in a wind space warp, no wind strength, but have some turbulence.
Drop in a damping space warp, 3d, subtle damping so smoke slows down with time.
Particles grow at a fixed rate over a fixed time. Lifetime then impacts the life of the particle and how big it gets before it dies.
That generally is a really nice way to make smoke for a fire for example, heat rising giving it upward velocity, lots of obvious movement and turbulence as particles are small and at speed, becoming less apparent as particles get big and slow down.


Just pass more info to particles, and then add a global brownian motion to all the smoke (smoke is in air, and air has a normal 'random' movement to it even on a calm windless day)...


No point fudging things the wrong way when we can just get loads closer simulating natural things and then tweaking the values.

Hmmm

Dave
 
Wow ! great work here guys !!!

Alex Forbin, would you mind if i use your settings and texture as the new default for smoke ? I love it !
I had me a blast just spinning around ..
 
Wow ! great work here guys !!!

Alex Forbin, would you mind if i use your settings and texture as the new default for smoke ? I love it !
I had me a blast just spinning around ..

Sure Mitch, glad to help. :)
And thanks.

I guess you've read the other posts but just to summarize. It would be really helpful to have the following...

A growth curve to allow the particle to start out very small and grow at a non-uniform rate.

Some form of tire heating calc (SA*time?) that controls the alpha, more heat=less transp.

The "x" velocity needs to be toward or away from the centerline of the car instead of simply right/left.

A simple "suck" parameter. LOL This would be just a simple formula that says the closer a particle is to the
car the more likely it is to follow it.

Soft particles! I know you're already working on this one anyway.

Thanks,

Alex Forbin
 

Latest News

How long have you been simracing

  • < 1 year

    Votes: 88 12.6%
  • < 2 years

    Votes: 63 9.0%
  • < 3 years

    Votes: 72 10.3%
  • < 4 years

    Votes: 44 6.3%
  • < 5 years

    Votes: 98 14.0%
  • < 10 years

    Votes: 96 13.7%
  • < 15 years

    Votes: 60 8.6%
  • < 20 years

    Votes: 40 5.7%
  • < 25 years

    Votes: 33 4.7%
  • Ok, I am a dinosaur

    Votes: 105 15.0%
Back
Top