In that case do you need to manuall stitch the physics curb+road meshes together?
General Proper technique in track making plus tips
- Thread starter LilSki
- Start date
In that case do you need to manuall stitch the physics curb+road meshes together?
No there can be tiny holes and you will never notice them.
Thanks for your long answer. So I will add the tail on all my curbs. I already thought to it before but i believed it was a not a clean way.
Just two questions.
How can we have a look on how Kunos track are made?
When I have quad poly I can unwrap curved shape (follow active quad) but I can't use follow active quad when there are triangles. I also try to map from top view and Smart UV project but no results. How do you unwrap this part (see your picture) along the arrow to have a repeated texture ?
Sorry for my poor english but now I don't write english very often.
Just two questions.
How can we have a look on how Kunos track are made?
When I have quad poly I can unwrap curved shape (follow active quad) but I can't use follow active quad when there are triangles. I also try to map from top view and Smart UV project but no results. How do you unwrap this part (see your picture) along the arrow to have a repeated texture ?
Sorry for my poor english but now I don't write english very often.
The way I map my tracks I only care about quads on the road surface and the initial edge between the track and the rest of the ground. If you look right off the curb it is still quads in that one section. The rest will be a regular projected from view map and applied to a google earth image so quads are not required.
Here is an example of a UV map of my NJMP Lightning.
And here you can see how the edge between the road and the grass is still in quads.This edge is what I use to blend between the road and the grass.
By no means is this the only way it is just how I have done it on all of my tracks. There are also a few ways you can examine original tracks but I will not explain them in this thread.
Ok I see. All grass is mapped one time. I understand more picture showing grass mask I can see on other topic. So this texture should be very high resolution (8K may be). It helps me a lot.
So now, I will rework my kerbs and continue to try to get clean road texture then get cool shader because I am really far far away to get the same results that yout photorealistic tuto (look my first post).
So now, I will rework my kerbs and continue to try to get clean road texture then get cool shader because I am really far far away to get the same results that yout photorealistic tuto (look my first post).
busterwolfeman
NetracingEurope.org
When I delete the ac_crew.dds from the .kn5 the game reports an error and will not load the dds.
However, I have a problem when editing track .kn5 files with 3DSimEd3. It happens that suddenly the ac_crew.dds will not load from the car folders. Everythink looks ok when I have a view at the pit objects and materials in 3DSimEd3, but they will not load the ac_crew.dds from the carfolder. This may be an issue with 3DSimEd3 (latest version), maybe the .kn5 have to be imported with activated object nesting, maybe it´s the thing that 3DSimEd3 renames all object names to lower case, maybe I shouldn´t save the scene as .3se (its own format). I don´t know it yet....
Anyone here with better expirience using 3DSimEd3?
Edit: Just did some testing. For the record: I imported the 5.kn from ks_barcelona to 3DSimEd3 and did the following:
1. Imported with deactivated objects nesting, saved it to the .3se format, loaded it again from this file and exported it to the 5.kn again. Result: No pitcrew at all.
2. Imported with activated objects nesting, saved it to the .3se format, loaded it again from this file and exported it to the 5.kn again. Result: Pitcrew with the ac_crew.dds from the 5.kn file, the ac_crew.dds from the carfolder has been ignored.
3. Imported with activated objects nesting and exported it right again as 5.kn file. Result: All is good, the ac_crew.dds from the carfolder has been loaded.
So it is most possible an issue with when the file will be saved as 3se. I will post this in the official forums of 3DSimEd3 as well.
Last edited:
I never said anything about deleting ac_crew.dds from the kn5. You rename or delete it from the texture folder which is something you use when you build tracks properly with the editor. I do not use 3dsimed for anything other than opening and exporting tracks. I do not care to support its use in my thread.When I delete the ac_crew.dds from the .kn5 the game reports an error and will not load the dds.
However, I have a problem when editing track .kn5 files with 3DSimEd3. It happens that suddenly the ac_crew.dds will not load from the car folders. Everythink looks ok when I have a view at the pit objects and materials in 3DSimEd3, but they will not load the ac_crew.dds from the carfolder. This may be an issue with 3DSimEd3 (latest version), maybe the .kn5 have to be imported with activated object nesting, maybe it´s the thing that 3DSimEd3 renames all object names to lower case, maybe I shouldn´t save the scene as .3se (its own format). I don´t know it yet....
Anyone here with better expirience using 3DSimEd3?
Edit: Just did some testing. For the record: I imported the 5.kn from ks_barcelona to 3DSimEd3 and did the following:
1. Imported with deactivated objects nesting, saved it to the .3se format, loaded it again from this file and exported it to the 5.kn again. Result: No pitcrew at all.
2. Imported with activated objects nesting, saved it to the .3se format, loaded it again from this file and exported it to the 5.kn again. Result: Pitcrew with the ac_crew.dds from the 5.kn file, the ac_crew.dds from the carfolder has been ignored.
3. Imported with activated objects nesting and exported it right again as 5.kn file. Result: All is good, the ac_crew.dds from the carfolder has been loaded.
So it is most possible an issue with when the file will be saved as 3se. I will post this in the official forums of 3DSimEd3 as well.
busterwolfeman
NetracingEurope.org
I searched the net about my problem and found this thread (which is public and no ones "property" btw). I thought that maybe some of the participated guys here knows about my problem. This is what forums are for...
I got you right with the track building process using the texture folder. But I thought your post is quite old meanwhile, and maybe AC changed some behavior. So AC acts different by loading from a folder compared to the .kn5, OK and thanks for the info.
Anyway, I isolated the problem as you can read above and wanted to leave this information here. Maybe it´s useful for anyone else.
BTW: I use a paid and licensed version of 3DSimEd3 for adding pits and grids to tracks for league use. If you just want to extract .kn5 tracks there are other free utilities out there...
I got you right with the track building process using the texture folder. But I thought your post is quite old meanwhile, and maybe AC changed some behavior. So AC acts different by loading from a folder compared to the .kn5, OK and thanks for the info.
Anyway, I isolated the problem as you can read above and wanted to leave this information here. Maybe it´s useful for anyone else.
BTW: I use a paid and licensed version of 3DSimEd3 for adding pits and grids to tracks for league use. If you just want to extract .kn5 tracks there are other free utilities out there...
I understand it's a public forum but it took a lot of time to put together this information and I don't care for it to become full of questions about 3dsimed. And yes there are other tools out there but for a while before content manager it was the only way to get shader information to find out how everything works.
Also if any of this information was outdated I would have updated it. Everything in this thread still works as posted.
I also update tracks for the league I am in but as I mentioned I only use it for the export process and have really come to hate how it is so good at screwing things up. I use it to get to blender as fast as I can.
Also if any of this information was outdated I would have updated it. Everything in this thread still works as posted.
I also update tracks for the league I am in but as I mentioned I only use it for the export process and have really come to hate how it is so good at screwing things up. I use it to get to blender as fast as I can.
I could have posted in the old dumb question thread, but i guess here is a bit more appropriate.
Looking back at the trees i did for LA Canyons, i used Forest Pro plugin and a small script to put all my normals up after it was scattered (pretty long process on the millions of polygons... and to redo each time i'd change anything)
I remember testing kstree_group_ naming, and it does look a bit better than normals up, nicely affected by the lights as you showed on screenshots.
But i couldn't work with separated objects on such a big scale of terrain, that would probably be insane.
I know you use Blender, so i guess that's a 3DSmax question more than anything else... although i'd be curious if solutions exist in blender.
kstree_group_ requires the trees to be separated objects. In 3dsmax, there are few tools to paint randomly on a surface, but it starts to be tricky to spread a large amount of separated objects on a given surface.
Alternatively, I tried to edit the normals on my source objects, but then there's no plugin that will retain the normals info once it's scattered...
Forest Pro doesn't, multiscatter doesn't... tyFlow doesn't either. pFlow neither.
I posted on tyFlow forum, no reply. On itoo forum (forestpro) and the dev told me in their situation it's the way 3dsmax deals with the normals and the instance system. They would need to rework the way they work completely.
So there. If anyone knows a solution, cause i found none!
Meanwhile I'll continue to spread my millions of trees with ForestPro, with no hope of proper normals, and they shall continue to look average ..
I might paint a layer close to the road using kstree_group. but the distant ones shall remain meh !
ps: cheers lilski for your time detailling this posts
Looking back at the trees i did for LA Canyons, i used Forest Pro plugin and a small script to put all my normals up after it was scattered (pretty long process on the millions of polygons... and to redo each time i'd change anything)
I remember testing kstree_group_ naming, and it does look a bit better than normals up, nicely affected by the lights as you showed on screenshots.
But i couldn't work with separated objects on such a big scale of terrain, that would probably be insane.
I know you use Blender, so i guess that's a 3DSmax question more than anything else... although i'd be curious if solutions exist in blender.
kstree_group_ requires the trees to be separated objects. In 3dsmax, there are few tools to paint randomly on a surface, but it starts to be tricky to spread a large amount of separated objects on a given surface.
Alternatively, I tried to edit the normals on my source objects, but then there's no plugin that will retain the normals info once it's scattered...
Forest Pro doesn't, multiscatter doesn't... tyFlow doesn't either. pFlow neither.
I posted on tyFlow forum, no reply. On itoo forum (forestpro) and the dev told me in their situation it's the way 3dsmax deals with the normals and the instance system. They would need to rework the way they work completely.
So there. If anyone knows a solution, cause i found none!
Meanwhile I'll continue to spread my millions of trees with ForestPro, with no hope of proper normals, and they shall continue to look average ..
I might paint a layer close to the road using kstree_group. but the distant ones shall remain meh !
ps: cheers lilski for your time detailling this posts
Last edited:
You can do something else that gives the same kstreegroup type normals. If all trees are single objects, you can create a sphere and use normal thief. If the middle of the sphere is at the bottom of the tree, the normal effect is the same.I could have posted in the old dumb question thread, but i guess here is a bit more appropriate.
Looking back at the trees i did for LA Canyons, i used Forest Pro plugin and a small script to put all my normals up after it was scattered (pretty long process on the millions of polygons... and to redo each time i'd change anything)
I remember testing kstree_group_ naming, and it does look a bit better than normals up, nicely affected by the lights as you showed on screenshots.
But i couldn't work with separated objects on such a big scale of terrain, that would probably be insane.
I know you use Blender, so i guess that's a 3DSmax question more than anything else... although i'd be curious if solutions exist in blender.
kstree_group_ requires the trees to be separated objects. In 3dsmax, there are few tools to paint randomly on a surface, but it starts to be tricky to spread a large amount of separated objects on a given surface.
Alternatively, I tried to edit the normals on my source objects, but then there's no plugin that will retain the normals info once it's scattered...
Forest Pro doesn't, multiscatter doesn't... tyFlow doesn't either. pFlow neither.
I posted on tyFlow forum, no reply. On itoo forum (forestpro) and the dev told me in their situation it's the way 3dsmax deals with the normals and the instance system. They would need to rework the way they work completely.
So there. If anyone knows a solution, cause i found none!
Meanwhile I'll continue to spread my millions of trees with ForestPro, with no hope of proper normals, and they shall continue to look average ..
I might paint a layer close to the road using kstree_group. but the distant ones shall remain meh !
ps: cheers lilski for your time detailling this posts
I've used 3dsmax for trees in the past with pretty good results IIRC. Collapse the modifier to make it seperate objects and use a mass-rename script or something to rename sections into group letters. KsEditor deals with the normals. It's been a while but I should still have the files. I'l check tomorrow.
His issue is that he’s using a plug-in to spread massive amounts of trees, and they only do combined meshes, so you can’t rename them. And now I realize my earlier message is probably pointless lol
Oh yeah I understood. I meant to say I've had good results without using plugins But come to think of it, it might not give control via weight maps
But come to think of it, it might not give control via weight maps cheers for answers guys
i guess youtube for "normal thief" was predictable
i found the scriptspot will see how that works for attached trees. Since it says "For each vertex in target object - the script will search the closer vertex in master object."
shouldn't be limited to single mesh, if there isn't collisions between my trees...
Stereo, your mesh voodoo is interesting thoughts, i'll give an attempt.
i know for LAC i did attach vertices as much as i could to save some space
Yet again, i think forestPro does reset a lot of things when distributing an object
in the meantime i'll also ask Ilja, since he did the moving trees for the wind whether their grouped or not, he could probably apply that automated normal wizardry there too
i guess youtube for "normal thief" was predictable
i found the scriptspot will see how that works for attached trees. Since it says "For each vertex in target object - the script will search the closer vertex in master object."
shouldn't be limited to single mesh, if there isn't collisions between my trees...
Stereo, your mesh voodoo is interesting thoughts, i'll give an attempt.
i know for LAC i did attach vertices as much as i could to save some space
Yet again, i think forestPro does reset a lot of things when distributing an object
in the meantime i'll also ask Ilja, since he did the moving trees for the wind whether their grouped or not, he could probably apply that automated normal wizardry there too
