RELEASED-wip Kyalami 2016

[Prototype]

@luchian I *think* my visual mesh is getting there. I did some tests on the physical mesh, and I'm getting ok results, but this will definitely need more work on my part, figuring out the vertex groups etc and adding the kind of variation that we need here …

I needed a break from the track surface building stuff for a bit, after trying SO many ways of doing the same thing, trying to figure out the best way forward to make my life easier when I get to actually building the entire track.

.. and of course, learning a new software package like blender has its challenges !

---------------------------------

Just a note on the Google mesh captures - this has been an absolute pleasure to work with, having so much reference, being able to tweak the buildings etc to match things up. Details like the pitch of the thatch roofs for example, such a luxury to have that info in there, and not have to guestimate those kinds of things. All this should add up to a more authentic experience when you're driving around the track.

Really good feeling knowing all my walls are the correct height.
All the painted road sections are correct and unique, and not just generic around the track. etc
---------------------------------


QUESTION:

Ive run into an issue.

I rebuilt the galvanised frames above, in Blender out of a bunch of cubes.
It was looking good in the KS editor.

ABOVE: Looking good. Everything looks as it should, and its taking light correctly.

---

But when I welded all the verts together and made it one single mesh object, the mesh then came into the KS editor with all the normal faces flipped. In other words I could see the insides of the frame box sections, but not the outsides. However, in Blender, the mesh showed up BLUE in face orientation. And normals seem to point outwards as they should.

When I flip the normals - shown below, it indcates in red in Blender, but in KS Editor it draws the faces facing outwards, (as if it were blue in blender). The other blue faces all work correctly.

So RED Blender faces now draw correctly in Editor, BUT it takes light "incorrectly."
In other words, it takes light the way it should if the faces were facing inwards.
BUT its drawing the faces facing outwards.

-----------------------------------

Track surface and wall taking shadow as expected.
But Fence framework is doing the opposite.
Shadows when its supposed to be lit, and vice versa.

Im sure this is just me not knowing Blender well enough yet.
Any help would be hugely appreciated.
Ive spent hours trying to fix this, and not getting anywhere.

 

[luchian]

Very weird issue . I am not sure how to go about it. Just a few (shy) ideas (as stated previously I am not a Blender expert either):

1/ Did you do a cleaning after welding all ? (aka remove doubles):
TAB to go to Edit mode.
A to select all
W and then select remove doubles.

2/ Is the shading set to Flat ?
(maybe try an edge split as well)

3/ Maybe make a try to open all in Blender 2.79 and do a test export ?
(I did not migrate to 2.80 - which is a big change - so not sure how it behaves and if all the bugs are ironed out).

4/Maybe you can send me a test file and I'll have a look. Sometimes I get ideas just by clicking like a monkey and trying stuff..

 

[Prototype]

Thanks @luchian

In 2.80 the remove doubles feature has been removed and replaced with
Mesh>Cleanup>MergebyDistance (It seams)

This definitely got rid of some verts! Thanks for the tip.

Tried your other suggestions but still same results.

I have a suspicion it has to do with the messy way Ive built the mesh.
Using mirror tool, dissolving edges etc. I think something just got out of whack?

Ill send you a file regardless! If you do have some time to take a look, thanks!
Will pm you link.

 

[LilSKi]

Did you reset scale? When mirroring you often have to reset scale and then flip or reset normals.

 

[luchian]

Hmm, good point^^.

Also, sometimes, I got to get the correct lighting by using a shader that allows a normal map texture.

 

[Prototype]

Did you reset scale? When mirroring you often have to reset scale and then flip or reset normals.

BAM! Yup.
Working now.
Thank you!

I reset all transforms and working as expected.
Good to know! Another lesson learned.

 

[Prototype]

Almost ready for some proper testing … :-D

 

[luchian]

This blue is very trippy

Looks good ! (road mesh is too dense for no bumps and stuff, but I assume it's the first test, to feel the camber and elevations).

 

[Prototype]

Yup this is this main mesh that Ill build the visual mesh and the physical meshes from.

The inner and outer beziers will be used for any additional fine tuning … but Ive done pass after pass getting this mesh to build as close to the Google mesh as possible … so hopefully there isn't going to be too much adjustment needed … My aim was to make this as accurate as I could. And also to build it as cleanly as I could. There isn't a single part of this track that's the same width as any other part, and the elevations constantly change.

 

[Prototype]

Some really interesting subtle stuff happening … the way the camber flicks back and forth SO slightly as you exit this left hand corner into the next left hand kink. The lateral load on the car is subtly shifted and unsettled through this section. I never really "got it" until now … and now thinking back at actually driving that section for example … the minute shifts you do as you exit there makes sense … super cool.

I checked, and double checked and triple checked whenever I got to issues like this.
Will have some friends of mine who have also driven the track, test on this mesh with me, and hopefully I can get second opinions on stuff like this … but feeling bullish for the most part ….

 

[LilSKi]

Kinda sucks to build the track again don't it? Been there done that with Bridgehampton. I have a love hate relationship with better data

 

[Prototype]

Kinda sucks to build the track again don't it? Been there done that with Bridgehampton. I have a love hate relationship with better data

Ha ha yeah,when I started this thing last November I didn't expect to still be working on it 7 months later...

So what did you initially use for Bridgehampton?

 

[Prototype]

Full hermit mode for 2 days.
Mesh is driving nicely.
Tentative lap trying not to fall off!

 

[LilSKi]

I originally had this 5' topo map that I built it on.

It was a great base to build on and I used the drawn outline for the track which I later found had some inaccuracies in it.

I later was given some lidar for T1-2 as those still exist today and then I found a nice high res aerial which better defined the layout.

This image here shows the differences between them. Orange is new and green was old. It drastically changed some turns for the better. I could now drive the track like I saw people do in old videos.

I also had learned a lot doing Riverside in studying video to pick up small elevation and camber changes and applied that to the new Bridge.

 

[Prototype]

Very cool. That vintage contour map is a thing of beauty.

 

[EKO Sim-Racing]

I originally had this 5' topo map that I built it on.
View attachment 3367

It was a great base to build on and I used the drawn outline for the track which I later found had some inaccuracies in it.

I later was given some lidar for T1-2 as those still exist today and then I found a nice high res aerial which better defined the layout.
View attachment 3368

This image here shows the differences between them. Orange is new and green was old. It drastically changed some turns for the better. I could now drive the track like I saw people do in old videos.

View attachment 3369

I also had learned a lot doing Riverside in studying video to pick up small elevation and camber changes and applied that to the new Bridge.

My question is this: how on earth did you get the data from Riverside, when the track was demolished and constructed over many years ago ?

 

[LilSKi]

My question is this: how on earth did you get the data from Riverside, when the track was demolished and constructed over many years ago ?

Riverside was based on a 20' topo and a 1980 aerial photo. The 20' elevation points I would "set in stone" and would not move them. I then studied hours and hours and hours and hours and hours of onboard video to fine tune and tweak the elevations between those 20' points. It took a very very long time to nail that layout down at about 4 month. It is a primary reason why I prefer lidar data now.

 

[EKO Sim-Racing]

Riverside was based on a 20' topo and a 1980 aerial photo. The 20' elevation points I would "set in stone" and would not move them. I then studied hours and hours and hours and hours and hours of onboard video to fine tune and tweak the elevations between those 20' points. It took a very very long time to nail that layout down at about 4 month. It is a primary reason why I prefer lidar data now.

I can imagine all that time spent !!
I mean, I refused the LiDAR data I have for Phoenix, because too time costly and the track is sort of flat (not really, but let's consider it like that), and I did all the bumps and street water evacuation cambers by hand.... took me 3 months and I still have some retouches to perform.

By hand it takes a LONG time.

 

[luchian]

I then studied hours and hours and hours and hours and hours of onboard video to fine tune and tweak the elevations

So what, now you expect a donation "bro" ?
I have also spent like seconds and seconds and seconds to download and install.. I even gave you a 3-star rating. You should basically thank "me".

#fairtrading #internet #sorryforofftopicprotoype

 

[Prototype]

So what, now you expect a donation "bro" ?
I have also spent like seconds and seconds and seconds to download and install.. I even gave you a 3-star rating. You should basically thank "me".

#fairtrading #internet #sorryforofftopicprotoype

HA HA yeah, "Users". Got to suck up to them when you're selling games to them. Doing all this work and then just giving it away for free … and then you still get entitlement … IDK .. this concept is kinda insane.