Adelaide Clipsal 500 V1.0 - by Digga

I was talking about tracks and sims in general, not about this or any particular car. It was just some general discussion. Trust me, I appreciate as well as enjoy this track. Digga's work is very nice looking as well as very fun and engaging to drive

 

I am the original author of the Adelaide track. Originally created for Sport Car GT, then upgraded to F1 Challenge and now converted to RFactor2, I am glad to see that has been converted and very good one, thank you
The track was created over 10 years ago using a satellite map gave to me by a mapping specialist before google earth was available.
I spent nearly 3 years creating it using actual photos converted digitally, it is still in my opinion the best and most accurate track available. Sure some elevations are not accurate and no bumps but it still is fun to drive on. Maybe one day I will release an updated version when time becomes available. Here is a link to a video of the original track

https://www.youtube.com/watch?v=wLZOKKdCKEQ

 

Thanks for your support Wizzfactor2, can I ask does this mean you have given your blessing to the conversion and it can be made freely available.

 

I'm really enjoying this race track, I love street races. Thank to everybody who made this track possible, from the first version to the current one.

 

Can we see this Isi affiliated like Bathurst?

 

Thank you, wizzfactor2, for all your time and effort spent into creating the original content of which this conversion of Digga's would not have been possible without.

 

Great stuff wizzo Good to see the old track back in action ... I still remember testing with this way back when we were doing the first V8SuperCars mod with ORSM

 

I'm glad to hear it is in fact your version of Adelaide, wizz. Simply because it's always been the best take on it. Though I sorta did have a bit of a idea it is your version, because a couple of buildings gave that away. Just stoked to hear you confirmed it, and the rF2 conversion has your blessing/permission.

 

You can't give permission for something you never had permission for in the first place. Glad Wizzfactor2 is pleased and not offended, or wasn't off doing it himself wasting one party or the other's time, but the only permission that is actually required (from original track creator) was never sought. Just clarifying since I see these discussions repeatedly and it's bit of joke squabbling over secondary rights when the primary ones are ignored. Better to drop the whole topic unless everyone involved is prepared to refuse to use any track that is not properly licensed.

 

For sure laser scanner tracks are not the solution to microbumps. The mesh density to be used would be astronomical and even in that case the physics engine wouldn't run sufficiently fast to take advantage of it. A car travelling at 200 kph makes 55 meters in one second. At 400 Hz the car would advance 14 cm per step. Therefore smaller scale irregularities could not be correctly simulated from a dynamical point of view.

As Lazza remarked, the best approach would be to artificially introduce the effect of those bumps. Apparently (I didn't know) more realistic noise patterns have been introduced in rF2. For sure, a senoidal pattern is insufficient to describe all the range of frequencies that compose a random noise.

However, I am not sure how ISI is tuning this since it can be quite complicated to get a realistic effect. I would expect very different Fourier decompositions depending on the track surface. Even between different parts of a same track. From what I have seen ISI tracks have usually a couple or three road materials regarding tdf definition. I'll check latest ISI releases to see how they are defining this noise and if they customise this by track or if they are using the same standard values depending on new or old tarmac for all tracks.



Enviado desde mi SM-T530 mediante Tapatalk

 

I just checked and latest ISI released tracks (brianza, portugal, bathurst and montreal) are using the same bumpiness parameters.
RoadBumpAmp=0.001 RoadBumpLen=3.0.
This applies to both new and old asphalt (road and asph) materials.

I haven't found any other parameter where the more advanced noise patterns Lazza described are customised. The above mentioned parameters describe a senoidal noise as far as I know. For sure to describe multfrequency noise many more parameters should be needed.

It would be nice if Luc or Tuttle could say a few words on how to fine tune this area of track modding. I didn't say but I also feel bumpiness artificial and not natural. The main bumpiness I feel seems to be provoked by the macrobumpiness caused by track mesh. With 2 meter long tris defining the road it is for sure that microbumps cannot come from the mesh but from the noise generation algorithm. If so it seems that this microbumps are the same regardless the track.

Enviado desde mi SM-T530 mediante Tapatalk

 

Thanks for the answer Luc.

However I still cannot understand how the different types of bumpiness can be adjusted using only two parameters one afecting intensity and the other frequency. With only 2 dof where the practical selection range for both is quite limited I would expect to get only few variations of bumpiness results.

I expected as well some kind of variation among new and old asphalt and between different tracks. Not only a slight change in grip. I understand that since you didn't deny my suggestion that all ISI tracks have the same "added bumpiness" you are confirming it.

If so it would seem that not much attention is being paid to this aspect or maybe that ISI has checked that the difference in perceived bumpiness between different tracks is caused by the major bumpiness which can be finely simulated with polygons that are several meters long.

I would expect that the best way to analise/tune this is to compare the suspension position telemetry data between reality and simulator.

In order to validate a model, the Fourier transformation of the suspension position vs time, should show the same density distribution among different frequencies as the real car does.

I would expect components with significant intensity quite less than 1 Khz being the main components at much lower frequency. In any case high frequencies cannot be accurately analysed unless we could capture data at much faster speed than the frequency to be analysed.

Comparisons should be made preferably in straights where tire loads would be stable if bumps didn't exist.


Enviado desde mi GT-I9505 mediante Tapatalk

 

It looks like no info wants to be provided about how to setup added bumpiness.

Either that or maybe it doesn't want to be confirmed that this aspect of simulation hasn't been sufficiently addressed since the same parameters are applying to all ISI tracks.

Enviado desde mi SM-G130HN mediante Tapatalk

 

No idea if helpful or not but I've seen detailed tutorials around regarding how accomplish various effects through the 3D modeling.

 

Your assumption on the polys being several meters long is a bit wrong, RF1 typically used a vertex length distance of 2.5 meters, however since then its moved much closer usually 1 to 1.5 meters, pending on the track creator, and in some cases tighter in turns since curving causes inside polys to become closer and outside to become looser so tight corners can have a distance of between .5 to .75 meters in many cases. This is good enough to simulate physically most bumps in surface that would affect a car. In reality I think the tighest LiDar data I have actually seen was around 1/3 to 1/5 meter between points, I'm not sure the density of what Iracing uses. As far as truly capturing reality, we can fuss over microbumps all day long but no sim that I know of simulates the potential of loose surfaces on a hard surface other than marbles affecting some grip. In reality there are still random factors like loose gravels, dirt, rubber, etc that may come between the tire and the surface constantly. At some point you have to face it that Simulation can never actually recreate reality, there are just too many variables. You just have to wait as simulations keep becoming more complex and keep getting closer, I personally think it is pretty staggering how far we've come already. If for some reason someone had even denser data, say .1 meter LiDar data and could create a surface from that, you could use that data to create a physical drivable mesh while keeping the visual mesh lower poly much in the way Iracing works and likely come out with even more precise feel. As far as RF2 physics I do know RF2 physics will react to a .05 meter alteration in the z value which is tighter than RF1 did, sometimes annoyingly so when you are trying to smooth out a imperfection.

 

i dont care for if theres a crack or minute bump missing, this track is one of my favourites in rf2, including ISI content. Looks superb, great ai......... I would like to see ISI 3PA this as bathurst.

By the way this with a grid of the Apex porsche = fun

thanks so ooo much, helping me fall back in love with rf2!

 

This!!!!!!!

I didn't fall out of love with rf2 though

 

IMHO one prob with micro bump laser scanning is that track surface tend to change sometimes (resurfacing, etc.)

i.e. the latest rF2 Lime Rock is more latest than the iRacing one because ISI did it after the LRP was resurfaced and the iRacing version is very much older than that.

 

I think the Ai warms its tyres like at Oregen, can anyone else confirm ? ( on the warm up lap )