Tracks Deutschlandring

I thought I should probably start a WIP thread for this track, as I feel I may have spammed the "What are you working on" thread a little too much lately - sorry :whistling:

I initially started work on it back in July 2016, right after I finished work on Bilster Berg. I was looking around for something new to work on when I found a link to 'race tracks of the world', including proposed tracks, unfinished tracks, disused tracks (http://theracingline.net/racingcircuits/racingcircuits/). I love the idea of bringing something to life in Assetto Corsa, and bringing something back from the dead doubly so :) When I saw the layout, and read the description of how the track would have looked I was hooked!

In the 1930s, the government of Nazi-Germany constructed a circuit that was to replace the infamous Nordschleife of the Nürburgring. The new track allowed for much faster speeds and, being exactly 10km in length, was to have much quicker lap times. This in turn meant more fun for the one million spectactors for which the circuit was to provide accomodation. In those days, motor racing drew huge amounts of people. It wasn’t unusual to have several hundred thousands of spectators at a single Grand Prix. This is reflected in the Deutschlandring design. Parking space for over three hundred thousand cars was planned, all of which could leave the area within two hours, thanks to several direct connections to the surrounding road network.

The circuit extended from an old hillclimb track near the town of Hohnstein, just a short drive south of Dresden, into a large oval. The designers got their inspiration from a number of legendary racing circuits around the world, including Mellaha at Tripoli, Linas-Monthléry, Monza and of course the Nürburgring. And there was no doubt that propaganda played a key role. This was to be the track to end all tracks. On April 27th, 1939, the circuit was baptised Deutschlandring. Enthusiastic supporters of the 1000-year empire quickly extended this into Grossdeutschlandring. But ironically, the circuit experienced the same fate as the Reich from which it had sprung. After the tarmac was laid but before the circuit infrastructure was constructed (pits, grandstands, hospitality buildings), war broke out and nothing came of racing. The upcoming German Grand Prix was cancelled, Mercedes and Auto-Union mothballed their racing cars and nobody ever raced the Deutschlandring. Today, it’s part of the public road network.

Untitled-24.jpg

deutschlandring-map.jpg
zeitungopeningcdubq_1.png

As the circuit was never completed, my version will obviously have to be semi-fictional (with respect to the track-side objects), although I will try and make the track surface as close as I can to the real track (which still exists and is now a public road). I have found no Lidar data, and the only DEM data I can find is of low resolution (90m) and only suitable for the surrounding terrain. The track will have to be modelled using topographical markings on maps, studying modern videos of people driving around the roads, and what feels right to me.

And so the work begins...
 
Interesting history, I've never heard of this circuit before. Sounds like it will be fun to drive.

I'll look forward to seing the "very elaborate pits of the most luxurious type" too :p
 
Wait, thats just 40km from where i live, how did i not know about this track? :confused:
Have to drive there someday, really looking forward to this one :thumbsup:
 
Just so I can get up to speed, this was my process to date:

1. To find out if the track was going to be worthwhile creating for AC I did a few day's research, taking screenshots from Google Earth and Michelin hiking maps. I used these to digitise the centre-line of the road as best I could, in Adobe Illustrator
2. I located the relevant area on http://jthatch.com/Terrain2STL/ and downloaded a low-resolution mesh of the terrain
3. I took my spline and draped it over the low-res terrain, lofted it to form a driveable road in AC, and went for a test drive.
D-Ring_02_1.jpg

Even though it was terribly inaccurate, and very bumpy, I could tell pretty much straight away that this track was worth pursuing. That was when the long hard job of researching the Deutschlandring began, looking for photos, maps, videos of people driving round the roads that are left today.

Fast forward about 4 weeks and I have quite a driveable track. I don't think I need to go into such details as how I digitised the road, or draped the spline over the terrain - those techniques have been documented many times already. One thing I did try for this track was to use a video to work out the track width at various points around the circuit. As I said, I couldn't find any high-res data for Germany (for free), and you can't even view the road on Google Streetview.

I knew that the 'base width' of the track was 12m, and I had found one good, steady, clear video taken from inside a car (so very little body roll compared to the bike videos I found). I set the video up on my screen, and placed a ruler in front of my monitor so I could measure the width of the road. As I knew the 'standard' width was 12m I was able to scale from that measurement to work out how wide the road was at key points like hairpin bends, or bridges. It took a while but I sat with a piece of paper and a pen and wrote down around 100 measurements around the track, and while it may not have been the most accurate data, it was some kind of data to work with.
IMG_20170222_224643029_HDR.jpg

IMG_20170222_224602629_HDR.jpg

So, in the top photo, the road at that point is 12m wide, and measures 240mm on-screen (scale 1:50). In the second photo, dropping down into the valley, the width has reduced to 195mm on-screen. As I have a scale of 1:50, and the camera hasn't changed its position relative to the road, that means that the road here measures 9.75m. I did this same 'measure-calculate-edit' all the way round my track, and it made a world of difference to the driving experience, changing it from being quite simple, almost mundane, to suddenly becoming much more of a challenge, much more technical. It's not exactly LIDAR but it worked for me. It was also useful in trying to gauge the camber changes at various key points too.
 
Still catching up to where I am currently at...

The road was created in 3ds max as a Loft object, using the spline I digitised in Adobe Illustrator (for the path) and 3 splines that I created in 3ds max (for the shape, or cross-section). The 3 shapes are for the cross-section of the track at various points around the circuit.Some parts are completely flat across the width, some parts are raised in the centre (to aid drainage) and some parts are raised at the edges (hairpins mainly, for safety).
D-Ring_CS.jpg

At this point in the project I used a single low-poly road for both the visual mesh and physical mesh, just to get the track in-game and driveable. The spline I created earlier has 192 verts, or a vert every 52m (on average) on a track that is 10,000m long, and that is fine at this stage where I am trying to create the correct elevation changes.

With 192 verts and a shape that uses 7 verts my lofted road contains 4,600 tris, well within the 64,000 tri limit that AC can handle.
D-Ring_road_01.jpg

I increase the path step in 3ds max until my road starts to smooth out, but stays under 64,000 tris. This gives me a driveable road to test with, but one that is still very rough, especially in areas of elevation change, where the steps become very noticeable.
D-Ring_road_02.jpg

I then copied this lofted road object, changed the cross section shape for one to create the grass verges either side. As both objects use the same spline, any adjustments I make to the spline will be automatically updated in the both the road object and the grass verges object. This means less work for me, and less room for error, making everything automated.

To make the width and camber changes on the road (and grass verges) I use the loft deformations (scale and twist respectively). This is where the notes I made from measuring the road width in the video come into play. I had written down how wide the road was, and the time in the video. Then I edited the scale of my road at the points that corresponded with the times in the video. It's a tedious job but gave me acceptable results. The position on the track runs from left to right, and the scale (or width) at each point is the Y-axis. My cross-section shape was 12m wide, and the Y-axis is a percentage, where 100% is 12m (75% is 9m etc).
D-Ring_width.jpg

Again, from studying the video, I also made notes about the camber around the track, and used the Twist deformer to twist the track by varying amounts to give my road camber. Again, the road runs left to right and the Y-axis is % twist.
D-Ring_camber.jpg

The road at this stage, with correct elevation, varying widths and varying camber. Not shown, but the grass verges either side of the road are also changing in elevation, width and camber, updating automatically with any changes I make.
D-Ring_road_03.jpg

So, I now have a road that follows the correct path, is the correct width and has the correct camber along its length, but this is just the beginning of the track. The next post will explain how I add some character to it, and prepare to split the road into a low-poly visual mesh, and a (much) higher poly physical mesh.
 
The beta version that was released used exactly the same physics mesh as the private, finished version. However, I'm pretty sure I could do a better job on the track now, with what I learned making Thomson Road.

I will just have to apply that knowledge and experience to the Deutschlandring instead, and the track that will come after ;)
 
Back
Top