Quote from tristancliffe :

I just replied to a PM from tinvek about this.

I believe that the devil is in the detail. The laser scanning will ensure we get a track that looks a lot like Rockingham, and maybe that is enough for 99% of players. They want immersion, but perhaps don't care about realism.

But what separates a track that looks like Rockingham from a track that DRIVES like Rockingham is the detail - the dips behind a kerb that mean you can't cut it at all, the drain covers in a kerb that try to spit you around, the sudden changes in camber that might be missed by the laser scanner, or smoothed by the post-processing (look at the first corner of the infield - a double apex hairpin (where you usually miss the first apex by a car's width). As you come off the oval onto the infield there is a nasty, sudden change of camber. Without that the corner would be dull. With it, you never quite know what the car will do. How accurately they model that will define whether sector 1 of the infield tracks is good or a bit crap).

It's all well and good modelling the grandstands and pitlane in amazing detail (and it does look amazing to me), or even modelling the kerb shapes in detail, but it's the details even finer than that that make a track. Drivers don't walk a track to look at the big picture - you can do that in a car or from onboard videos - they do it to look at the details.

Will LFS have these details? Will Eric have gone that extra mile and talked to drivers (more experienced, better ones than me I mean) to find out what makes a track a track, and what defines the lines through a corner. Will he have bought/read the UK Circuit Guides handbook thingy? Will he have looked at Datalogging (and not just of a perfect lap) to see what is happening to the cars as they circulate.

I'd much rather have that surface fidelity than any fancy modelling grandstands, because the surface fidelity is what makes a real track real. With a fictional track you can avoid a lot of it, because the fictional track isn't trying to recreate something that actually exists. Therefore the fictional track IS the track - just a track that doesn't have dips in the kerbs at an apex drain!



I have a few on my website(s), but I'm uploading some Rockingham videos to Vimeo. http://vimeo.com/tristancliffe/videos

The first one (and the last one of this week) is the wet practice session. All from 2008. I ended up on pole, by quite a margin!!!
http://vimeo.com/6733690 - it has actually finished uploading, but there is a conversion process that occurs. I don't know how long that will take...

Edit - just looked at the instance of Firefox that is doing the uploading - conversion is due to finish in 45 minutes. No, 31 minutes. No, 37 minutes. You get the idea. Check in an hour. I'll upload more next week. Wish me luck at Croft!!!!

From what I have red about laserscanning is that it notices differences from 2 mm in surface. So I don't think a chance in camber would be missed and the kerbs are laserscanned with the track.

Yes, the laser scanning can detect small differences, but they don't check all of the surface of the track - You've seen the pictures. Close the camera car they probably scan every 30cm, further away once a meter. Then they move up and do it again, so it overlaps a bit, and then they do the fiddly bits. But what if they don't pick up on an important fiddly bit? It'll just be forgotten, but that could be the critical thing that makes the corner what it is.

As for tyre models, I disagree. Regardless of the tyre model, in an ideal world we'd still prefer to race on a real track that was real - hell, that might even highlight flaws in the tyre model that could be improved that we don't see with a less accurate track.

And I'm well aware that making a 10GB track file and expecting us to buy super computers from Sauber to play it is too much (at the moment). There has to be compromise. I just hope that the important details are found, used and retained, even if that means reducing the complexity of the grandstand modelling.

By the way, when you're hitting bumps at 150mph, a bump that is 2mm too high goes from being perhaps unnoticeable to being a critical factor in how the track is driven. 2mm accuracy - for those that think laser scanning results in a realistic track from a driving point of view - isn't enough. Least of all when the points taken are so far apart.

The laser scanning ensures the track is close to reality, but doesn't mean that it's an accurate recreation of the track surface.

I think it was 2cm. Hopefully the laser scanning will have picked up things such as changes of camber.

From memory the Kerbs at the Rock can be quite vicious, epsecially the one on the exit of Tarzan (where the pit entrance for the infield circuit is), that shatters your spine if you go over it and bend wheels out of shape.

Quote from marzman :

From what I have red about laserscanning is that it notices differences from 2 mm in surface. So I don't think a chance in camber would be missed and the kerbs are laserscanned with the track.

You can't have that much detail in the virtual track though, as framerates would plunge. Looking at the shots on the main LFS page, the scanner seems to have a resolution of around 1 data-point every 10cm. This should be enough to capture 99% of every little camber detail - it depends on how sensibly Eric performs the post-processing - turning millions of points of data into merely thousands by getting rid of the unneccesary ones. The trick is knowing whether a point is neccesary, for example the middle of a long, flat straight won't need as much detail as the point where you come off the oval banking.

Tristan, I believe Rockingham is in safe hands - looking at Eric's other circuits, in particular his most detailed work - the chicane at Blackwood - I think he has the experience to do his research and keep the detail in where it's needed, and lose it where it's not to keep things running smoothly.

Now... where to kidnap a laser-scanning team for myself.....?

And that is why there is no difference between a fictional and real track on a pc. 95% percent of us are not very likely to go driving there so we'll never know how exact it really is.
I'll never know that bump A should be less hard than B or that Curb C has different paint or that this odd tire stack is supposed to be 10cm more away from the curb. You can't get an infinite exact model, tracks also use to change over the time...

And I don't care to be honest.

Quote from three_jump :

And that is why there is no difference between a fictional and real track on a pc. 95% percent of us are not very likely to go driving there so we'll never know how exact it really is.
I'll never know that bump A should be less hard than B or that Curb C has different paint or that this odd tire stack is supposed to be 10cm more away from the curb. You can't get an infinite exact model, tracks also use to change over the time...

And I don't care to be honest.

True, most of the bumps in a circuit are not "by design" but because of aging, also a track can get new asfalt what changes everything.

Quote from tristancliffe :

By the way, when you're hitting bumps at 150mph, a bump that is 2mm too high goes from being perhaps unnoticeable to being a critical factor in how the track is driven. 2mm accuracy - for those that think laser scanning results in a realistic track from a driving point of view - isn't enough. Least of all when the points taken are so far apart.

Most of the measurement error is in horizontal plane. If the laser is 2m above the ground, just in 10 metres from the laser, vertical measurement error will be 2mm*2m/(2m^2+10m^2)^0.5=0.19mm

Quote from tristancliffe :

But what separates a track that looks like Rockingham from a track that DRIVES like Rockingham is the detail - the dips behind a kerb that mean you can't cut it at all, the drain covers in a kerb that try to spit you around, the sudden changes in camber that might be missed by the laser scanner, or smoothed by the post-processing (look at the first corner of the infield - a double apex hairpin (where you usually miss the first apex by a car's width). As you come off the oval onto the infield there is a nasty, sudden change of camber. Without that the corner would be dull. With it, you never quite know what the car will do. How accurately they model that will define whether sector 1 of the infield tracks is good or a bit crap).

Interesting detail. Thanks for explanation.

AFAIK, in LFS as well as in NKPro collision meshes can be separate from visual meshes where necessary. But how are kerbs made in LFS anyway?

Quote from tristancliffe :

I'm sure no one is interested, but I've started uploading my videos from Rockingham from last year. My upload speed isn't very high, so I'll probably only get one done today (wet practice), and then I'll start with the others next week. Off to Croft tomorrow for a race meeting, so won't get any up over the weekend. Maybe Eric will watch them and see something in them he hasn't spotted before, thus making the track even better in LFS - I'm still concerned that the end result won't be very much like Rockingham in reality.

Maybe they can invite you to the closed-beta team?

Quote from tristancliffe :

Yes, the laser scanning can detect small differences, but they don't check all of the surface of the track - You've seen the pictures. Close the camera car they probably scan every 30cm, further away once a meter. Then they move up and do it again, so it overlaps a bit, and then they do the fiddly bits. But what if they don't pick up on an important fiddly bit? It'll just be forgotten, but that could be the critical thing that makes the corner what it is.

As for tyre models, I disagree. Regardless of the tyre model, in an ideal world we'd still prefer to race on a real track that was real - hell, that might even highlight flaws in the tyre model that could be improved that we don't see with a less accurate track.

And I'm well aware that making a 10GB track file and expecting us to buy super computers from Sauber to play it is too much (at the moment). There has to be compromise. I just hope that the important details are found, used and retained, even if that means reducing the complexity of the grandstand modelling.

By the way, when you're hitting bumps at 150mph, a bump that is 2mm too high goes from being perhaps unnoticeable to being a critical factor in how the track is driven. 2mm accuracy - for those that think laser scanning results in a realistic track from a driving point of view - isn't enough. Least of all when the points taken are so far apart.

The laser scanning ensures the track is close to reality, but doesn't mean that it's an accurate recreation of the track surface.

You're right,that's why i prefer fictional tracks somehow.
I mean,its better to make a fictional track from scratch with bumps and thing rather than laser scanning a track but leaving important bits out of the final result.
Any way,Rockingham should be a nice add to the game and im sure it'll make many people happy.
Not complaining(:razz just my two cents.
Anyone knows about how the Iracing scan method is managed?I mean,i guess they're not going to laser scan the whole track,inch by inch,so there must be some particular working method behind.
BTW:nice cameracars Tristan,love the trackvision thing

Quote from tristancliffe :

I just replied to a PM from tinvek about this.

I believe that the devil is in the detail. The laser scanning will ensure we get a track that looks a lot like Rockingham, and maybe that is enough for 99% of players. They want immersion, but perhaps don't care about realism.

I think that's why (truly) profesional simulators exist.

There are lot's a great simulators out there, planes, boats, trains, R/C, etc... but all of them have a line that separates reality from simulation, and we have to accept it. I enjoy XPlane as a great simulator, and I'm sure is one of the best out there, but I don't think a simple Dual Core Pentium is going to give me what a real plane does. You talk about small track details and surfaces, well, think about air currents, temperatures, presure, gasses, weather...

Maybe you have a more critic view because you can enjoy real racing, and I understand it, but you have to understand too that the majority of us will do well with it, we all share the same enthusiasm for accuracy but we have to accept that a home-simulator will always put the line that separates reality from fiction farther that a professional one.

Just my 2c

Quote from marzman :

True, most of the bumps in a circuit are not "by design" but because of aging, also a track can get new asfalt what changes everything.

That's a great point

Tracks are not inmutable, even if we had a 1:1 version of today's track, next year probably the simulated version would differ form the real one.

While having all the bumps with centimetric precission would be awesome, Im convinced most people here will be very happy to get a new track if it has a reasonably challenging layout and/or is interesting enough to drive

Quote from tristancliffe :

The laser scanning ensures the track is close to reality, but doesn't mean that it's an accurate recreation of the track surface.

While I think it's important to realise this, it's also worth making the point that 'the perfect is the enemy of the good', and that while laser scanning won't get us all the way to a perfect simulation of a real life track surface all by itself, it's still pretty darn good for what it can do. Track recreation seems an ideal application for this technology imo- used in conjunction with an awareness of the kinds of details that you've mentioned, I think you'd be able to get pretty close to an accurate reproduction. Not perfect, but certainly good enough, and certainly better than what we've seen go before it in home simulation.

At first, I was thinking: "Why is everybody so happy about these new additions? When we got the FBM, it was not much different, almost the same as the FOX. The Scirrocco is also quite similar to the XFG and the new track is also similar to the Kyoto one."

But then, it hit me: "Along with FBM came the clutch heating, engine stalling, transmission upgrades and other things. With Scirrocco we will get limited setups and tyre simulation improvements. Rockingham brings us a firts laser scanned track which might also bring a different feeling to the game."

There is more to it than just another fwd hatchback and another track. Although I'm in favor of a more "vertical" track (with lots of elevations), I am still looking forward to the new patch and S3 content. Devs, you have my support and as soon as S3 comes out (whether it has one track or ten, I don't care), I'm buying. Just show me where do I sign

In addition to laser scanning, I'm sure they've also realised that there is an immense amount of telemetry data that they could use from a variety of cars. When a car goes over a bump it will register as suspension movement and it cannot be beyond the wit of man to map this movement and recreate the bumps from it.

Quote from SkyNet :

Why don't you go the iRacing way? Where users have to pay for each track/car. Imho LFS is way too cheap for what it offers.

No not if its on a motnly base if its 1 time pay uyou get ir forever yes maybe

Quote from Storm_Cloud :

In addition to laser scanning, I'm sure they've also realised that there is an immense amount of telemetry data that they could use from a variety of cars. When a car goes over a bump it will register as suspension movement and it cannot be beyond the wit of man to map this movement and recreate the bumps from it.

It may not be beyond a man's wit, but is it beyond a man's wallet?

Or the will to provide the data ...

Quote from Danke :

It's possible that Eric has been working on a fictional track or two since the release of S2 way back when. Then he put it/them on hold when this Rockingham thing came up. So we still could see a fictional track or two eventually in S3.

That's highly unlikely, but like you said, nobody knows for sure. It's just that if a single track takes Eric a year to build, and people think it only takes him three months, we're behind the 'expected track amount' by three tracks a year -> lots of disappointed people after it's realised that there's nothing more to come after years of "development"..

Let's just hope that's not the case

Been to Rockingham once, can't wait to live it again in LFS

I went to see Ascar in 2004 at Rockingham, was great..

Considering there will now be two oval tracks, it would seem a bit strange to not produce a stock car for the sim seeing as they are focusing on all areas of motor racing.

Personally I would love to have Legends Hotrods with the hiyabusa bike engines in LFS which race around rockingham regularly, would be awesome imo.

on the new track
on the extra new car

Still want a live track that builds marbles etc off the line where the grip is worse etc Lets hope S3 brings that in at some point.

Hey Scawen, I know you don't usually answer stuff like this but: Any vague idea about how far off you are with this stuff? Is it a "needs a bit of a tidy" or a "needs a currently-unknowable amount of more tweaking" or a "we're waiting til other important stuff is ready to go too, because the new environment will work better then" type of scenario?

Kev.. it was released last week.

www.lfs.net