W.I.P. Leeds Loops WIP

[Mr Whippy]

Continued from here:
http://www.assettocorsa.net/forum/index.php?threads/leeds-loops-wip.32352/

Sadly lots of the images are missing but I've uploaded a few of them here in a sorta sensible order.

Initial layout was literally the loop right at the far left including the pits. But it was like 15s a lap in an F1 car. This one was a bit longer, but still a bit dull.

So it ended up that this was probably the best compromise on workload and an interesting layout and length.

Splines drawn over the photogrammetry re-captured HD G3D point cloud data outlining the asphalt.

Splines all finished and starting to fill in with a roughly 1m quad mesh.

Not the best way to fill in it all in in hindsight. More prep work on a copy of the base splines might have been better (cutting and normalising chunks of spline to near 1m intervals)
There are swings and roundabouts to every workflow on a track. Cutting and normalising might take almost as much time as doing things manually, and given the cut and normalised splines are 'binned' afterwards it is also a waste of time if you later go back and tweak the base splines... the cut and normalised bits need remaking and merging again etc etc.

Almost done.

Close up of a bit. Mmmm, flowing mesh. Nice to have uniformity for when it gets conformed to Z data for the road crown, cambers, extruding out for pavements and so on.

Here's a shot of the final road surface mesh.

And one without the wireframe. Here I've just done a 4096x2048 px planar texture over the entire road and then baked the rgb from the point clouds of the G3D data and used it as a texture just to give me an idea of where things are etc.
This is the version in Free Racer (not in AC yet) that I can drive around on just to check how smooth and correct things feel so far.

 

[Mr Whippy]

Here is the conforming mesh. It's lower density at this level, and has just enough detail to capture the camber/crown and hills/slopes
This can then be conformed to my point cloud data using a statistical approach (multi-sampling, area sampling etc, so each vertex gets a good average of the location it 'lands' when conforming)

The mesh is then sub-divided to add smoothness and detail.

The visual mesh (1m grid) is then conformed to that conforming mesh.

I can also play with conforming the 1m grid to the point cloud data (as bumpy as it turns out due to noise), and then morph the two meshes together (ie, use vertex paint to say where I want it smoother or the rougher data, or use a fixed blend amount etc)

I can test all these quite quickly until I'm happy with the way it drives and looks. I think something like 20% noise at the sides, and more like 40% noise in the middle was feeling and looking natural... so all using real data, but picking and choosing the right amount for it to look and feel 'right' (right is so subjective without a HQ laser scan and even WITH a HQ scan )

Buildings so far.

Roughing in buildings from the HQ point cloud (from G3D images)
As you can see the data is pretty good but not perfect.
A lot of the time I'm still referencing the 2D HQ aerial images panned around to fit the right location, and then getting the roof line outline perfect, or exact proportions of things... then using the 3D data to get the Z info correct and the general positioning.

Soooo much work to do.

Main tasks are just keep roughing out 'correct' but not heavily detailed building meshes. Get all those in. Then kerbs/pavements up to the buildings so the track feels water tight.

Then I can start thinking about track layouts and all that stuff and make it race-able, refine that side of things until it feels like a solid quality race course.

Then hopefully interested parties can pick up buildings and up-res/texture them and it can become a combined effort (assuming there is interest) in finishing it up as a HQ city race course.

Then all the props, trees, etc etc.

 

[luchian]

Hey, this is nice. Shows yet another approach on how one might tackle a (real) city track.

So we have 4-5 wip tracks here, that might endup as great content. Until then, I'm sure we can learn a lot from each other.

Thanks for sharing.

PS: when the time comes, I'd like to join the buildings modelling task force . (don't worry Pixeldude, will not abandon LRP )

Sent from my phone using Tapatalk

 

[Willy Wale]

Then hopefully interested parties can pick up buildings and up-res/texture them and it can become a combined effort (assuming there is interest) in finishing it up as a HQ city race course.

I won't volunteer for texture duties as I'm not sure I have the skillz, but I will be making lamp-posts, bollards, rails, traffic lights, litter bins, and post boxes so we might be able to save each other some effort.

 

[r@m]

Good to see so much activity...track makers are rarer than rocking horse shit

....if you build it...they will come

 

[Mr Whippy]

I won't volunteer for texture duties as I'm not sure I have the skillz, but I will be making lamp-posts, bollards, rails, traffic lights, litter bins, and post boxes so we might be able to save each other some effort.

Good point. UK roads will share lots of similar furniture.

I need to get a bit further on the Leeds track roughing out, and then I can basically block out all the furniture and see what is required.
I'll probably do everything with nice HD/LP meshes, doing PBR workflow and bakes with normal maps etc, mainly so all the assets can be sold elsewhere too (ie, UDK/Unity workflows)

I'm most looking forward to doing the bins. Part of me just wants to be able to drive around the track running over bins and watching them go flying haha

 

[Norman Leslie]

Mr Whippy, have you abandoned this poject?

 

[Mr Whippy]

Not abandoned.

Just had two kids, now 5 and 2.5, and then all this covid stuff, and other family stuff.
Needless to say I keep loading Leeds Loops up and refreshing my memory, having a blast around in a Gallardo or Carrera GT, and it’s a track I want to finish up more, and even if I can’t finish it, get it to a point I can release it and source files so others can do as they wish.


And the point cloud plug-in I used worked from Max 2010 to 2014.
Max’s inbuilt point cloud support is really poor in comparison.
So I’m stuck on Max 2010 for any point cloud work.
But newer versions of Max are so much faster in viewports it’s crazy (mesh work), so it’s a difficult process now working in Max 2010.

I did a project in it in 2018, still completely do-able with a billion point LiDAR scan (no issues with the speed of that!)… but the mesh work chugged so I had to manage my layers a lot.
BUT, right in the last few days their authentication server went down.
I had to finish up in Max 2017 using their point cloud support which was horrible.
Since then the server has been patchy so I haven’t been able to rely on the software since.
Sadly they won’t just give out a persistent license or key or just ‘freeware’ it, so pretty crappy.

Though I’ve got a friend who could probably make it ‘work’ for me, but even he’s need paying… and it’d be for an old plug-in…


BUT, clouds2max VX (or something) is now a new and expensive version again, but as yet no point draping (mesh draped onto point clouds) or texture baking point cloud colours into textures, which was a big feature I used in track production a bunch of times… but it’s supported in Max 2017 onwards!
But how long that’ll remain well supported is another question.

A part of me wants to get into Blender more and I’d hoped by now we’d have seen an open point cloud tool in Blender to help push me over.


Right now it’s a balancing act on buying/hoping clouds2max RX gets draping/baking support, or just running two versions of Max and maybe some FBX to/fro workflow figuring out so I can hop back and forth, which is always a thing on hobby tracks… pontificating and changing previously done stuff later.


Also getting slightly more time now the kids are sleeping on a night.



For now though the big job was just roughing in all the main nearby buildings so I could at least seal up the roads/pavement/surrounding buildings.

For now that’s actually quite a fun task.
It’s nice to just put on some music and get into the zone and enjoy the process.

 

[Mr Whippy]

I took a few hours today to install Max2010 and clouds2max.

What a faff wrt Max2010.

Firstly mentalray wasn’t happy loading, only guess work but just renamed the mental ray stdplugs files.
It’d kinda work and load up each time by invoking the compatibility mode, but that was silly. It just wouldn’t work without using the wizard each time.

Then UI scaling. 4k screen. Fiddled with settings. Now runs in what looks like 2560px wide but it’s solid.

Also permissions issues, so ran it in “D” drive with no user mode (ie stores all files in app folder, much nicer, almost portable)

clouds2max installed in seconds, authentication server up, all good.
This is an awesome plug-in. Just how people make tracks from LiDAR without it I have no idea?!

loaded Leeds Loops project, all worked fine. Yay!


I’ll make some progress tonight and update.

Also I think I want to change the layout a bit, so I’ll post my ideas.
Not a huge change but probably it’ll make it loads nicer to race on.

 

[Mr Whippy]

Yikes, Leeds City Centre has changed a lot in the last 5 years.

Swathes of what were wide road areas are now narrower, pedestrianised, or shared cycle lanes.

Here is the original layout I'd run with.

It wasn't great in some regards. T6/7 essentially cut up loads of pavement to make them work, because they had narrow pinch points.
S/F area was also cut up down the middle to give width... but all in all, from what I've seen done on FE circuits I've built for sim users, these changes would likely be made for the benefit of a circuit being run at the city.

The downsides were the SF run was really short, the pits were a short-cut, so it wasn't ideal.


Now follows a boring analysis of the changes.

T1:
The junction has been changed a lot, and the lead in from the SF straight. Basically pavements pulled in on the North side, the taper has gone for turning traffic (a legacy of when this was a big two-way main road in the 1990s) etc. So now the SF is narrower, and it was already short. The T1 is also now even less interesting and exciting.
I could just ignore all these changes and do whatever I want, but sometimes it's the interesting reality of a city that forms interesting tracks. To just slap a track over everything will result in an uninteresting and less believable course I think.

Old


Current


T2. Same.
T3. Same.

T4.
This was already a tight corner, but hey, city track. But now the exist inside is pushed across for a wider pavement, so it's even tighter. However, the entire run down from T4 is now all just pedestrianised and landscaped with trees etc. There would never be an FE track here.

Old


Current


T5.
Because this runs off the T4-T5 changes, it's now not the same high-speed kink/chicane space.
The T5 exit is now also tighter on the exit, and then a protected cycle path is present... I assume this runs off the T4-T5 section, so people can cycle right down into the city centre away from major traffic.
There is also a pinch point to essentially a big car-width, half way between T5/6 for a busy pedestrian crossing.
This entire run made this circuit work one way or another, and offers a decent high-speed section.
I think if we ignore the protected cycle lane kerbs and the pinch point, this could be workable as an FE run. I think there would probably be a way to make this work in a 'realistic' sense. Maybe some kind of removable kerbs for the cycle lanes... enforcing the pinch point with removable kerbs and bollards?!

Any way, the track has to run between T4 and T5 because there is no other way down to the City Square again.

Old


Current


T6.
This is actually improved. The island I just ran over before, is now opened up and tapered both ways... so down here the track can turn both ways!

Old


Current


T7.
Like T6, T7 is now opened up, so no need to run over the island. All good.

Old


Current


T8. Same.

T8.5 (sorry, my bad numbering haha)
This point has been changed a bit. The kerbs to the right are further out with a wider pavement. But they didn't impact anything.
The kerbs on the outside are tapered right back, and the kerbs on the inside are tapered back too, so more room for a faster run, or chicanes etc... or other layout changes.

Old


Current


T9.
The entry is the same, but the kerb on the apex/exit is way across now as the pavement has been made wider on the SF straight. Also the opposing side is a wider pavement on the SF straight.
It'd still work, but it's just not SF sufficient any more.

Old


Current



So yeah. I could just pretend it's 2017... but the entire concept of a FE/city type course is that it works in the city centre... so just turning it all into big fat roads again kinda goes against the entire angle of that I think.

It also had shortcomings with regards to the SF straight.

So a new layout to follow while I'm still at this early stage.




On a more general off-topic note.

It's amazing how much money is being spent in a large regional city like Leeds. I do wonder what the future holds for cities.

They had planned to build trams (which actually ran there half a century ago, and were ripped up for cars), but then they've never got around to spending the money on putting them back.
Now we are WFH, and people live a lot more in the suburbs... Amazon, Deliveroo type lifestyles.
Who is going into cities? Do we need the trams any more?
Less roads, less cars, less public transport?

Expensive car journeys and parking, difficult driving... a good chance places won't have what you want in stock to buy any way.
Expensive public transport that hasn't been upgraded in any meaningful way for decades is nasty.
Leeds train station is now lovely, but most routes and rolling stock running into it are hideous. Only the Ilkley to Leeds line is something I'd want to take myself or my family on.
Out of town shopping was already killing city centres. Now internet shopping is killing out of town shopping.
Cities are re-inventing themselves, and to be fair having cars flowing through them, especially the middle of them, was never sensible.
In the 50s and 60s much of Leeds was being designed to be a two-level city, with roads/network underneath and a nicer pedestrain layer above it... but obviously that never happened.

I just don't know what I see for cities. I like the idea of making a track there. I like the experience to go occasionally for something special... but absolutely nothing attracts me to working there, living there, shopping there, eating out there, at any point.

Clearly money is being spent there by developers and the council. I just wonder why. It feels richer and more exuberant than ever, compared to the Leeds of the 90s that I remember... but then there appears to be less and less money going there than I can remember.

People talk about money still being spent somewhere, and yes that is true, but it'll be suburbs, smaller towns and commuter hubs around where people are living.

Interesting times ahead... but my gut feeling is big glorious cities days are numbered, unless we tax online shopping to bring it to partiy with the cost of shopping in a city (inc travel)... and WFH is taxed/disincentivised, again, to support the existence of cities.


What do other people see as the future of a city? What will we do?

Cities aren't intrinsic to our world or society... you could argue their unsustainable relics of our old way of life... do they need to die/shrink away, slowly returned to green spaces and lower density?

All very interesting.

 

[Mr Whippy]

This is a new layout idea.

There is a fair bit of room for pits tents. The FE spec tents fit in the space nicely. They're backed up against buildings in a few places, but generally it's spacious to feel realistic.

There is space for 140m of grid spaces back to the pit exit point.

The run to T1 is about 150-175m which is a good run vs before.

The pits are a longer route. Good. But they're not on/over the SF line... not sure that's critical on a street course? Some other FE courses are like this.

Same overall length, give or take. 1500m ish. Quite short for even an FE level track, but long enough I think.


Except T6-T7, the layout pretty much works on the current street scene...

I have the points clouds from when I captured in 2017, for the areas required... sans just a patchy area between T4 and T5, but I know that is just a smooth run over the gap.


The pits entry will be really good at T5, outside of bend, high speed entry off T4-T5 run up.
Pits exit is nice and off-driving line, and should be 'safe' within the space... hopefully.


Also at each corner point, there is sufficient space to make them sharper, faster, have chicanes etc, to add interest, and a few areas are wide enough to have mid-run chicanes etc.

Lastly, the route probably takes in more 'nicer' sights and areas of the city, which is no bad thing.

Hmmmmm