The "What Are You Working On?" Thread

Just as a thought I think of building my first 3d model for AC. I have never done anything 3d related and now I want to knikif a car or a track would be easier to start with
Hmm, to start with you should pick a car or track you like a lot. Passion is the most important thing, if you like the car enough you will have the will to persevere through the tricky bits, whereas if it's not something you have a passion for you will more than likely give up when it gets tough;)

Also try to pick something simple and beginner friendly, for instance modeling a complex car (e.g Mazda Furai) is going to be more difficult for a beginner than something with a more basic shape (like a Volvo 740). A car with good blueprints/reference images available will also make life easier!

Anyways, AE86 dash is now partly finished :D
__custom_showroom_1521456912.png

(textures are placeholders, going to make some in due course)
 
Tracks almost doesn't need any work with physics, gadgets, animations, moving parts, sound. Also because they are greater scale precision is less demanding. I still wouldn't do any real important track without LiDAR ever again. Because you'll work your ass off trying to make it precise, and people will still just don't give a damn about it, because it will not have the scan label, and spoilage is real. Unless it is some novelty so precision is less desired.

Track will be more difficult the larger it is and the more variety is there. For example if your track is running through one type of environment it will be easier than it would be if the surroundings of road is constantly changing very much. More objects around also means it will be more work. The more work you'll be able to do with as few splines as you can the easier it should be.

Tracks requires more consistency of your work quality. If you'll mess up at beginning you'll have more pain than you would with a car, unless you'd rush really far into progress with a car... it should still be easier to tweak things out IMO than for a messed up things at the track. I have experienced both. Though some modders here would remodel whole car if they mess up, rather than tweak it out...

If you'd do car, then try learning photo matching early. Most blueprints usually are off. Otherwise you'll see the pictures of a car when you'll be closing your eyes. And possibly go insane. But you'll understand the curves well :D Well if you will understand them for once haha car can look totally perfectly modeled from one angle, and completely off from different angle.

For a car I would also look for less curvy car, with not too complex interior, also would check if it is symmetrical, for example I have been modeling D-Type not long ago, comparing to XJ13 and E-Type it is very asymmetrical, obviously it makes it a bit more difficult.
 
How does the community feel about high poly models? I might be able to get it down to 350-400,000
Stick to Kunos guidelines and you'll be fine, it's all in the dev folder on the AC_Pipeline_PUB_Rev2.0.pdf file. No point in doing a car that'll only run by itself and crap out performance when racing against a full or half grid.
300k-350k is a solid number if you want to have visible engine details, it all comes down to the LOD_B and COCKPIT_LR models which is what the game uses most of the time.

My latest car has about 320k with engine details and moving parts (suspension, engine, etc.)
Showroom_rss_gt_shadow_v8_19-2-2018-10-59-51.jpg

Having the complete LOD_B around 50-60k is what does the trick there.
 
Door panels and seat will build the wrap around of the cockpit. Dashboard and center console last. Luckily everything is sort of flowing inside...the way the dashboard and floor seaml seamlessly connects to the door panel is pretty timeless and very well executed.

aSappDf.png


Guess I coulda picked a car with less buttons :D

dAje5Bb.jpg


Interiors on these cars which could blow the 200.000DM sound barrier in the late 90s got pretty crazy...

e0T9e8s.jpg


g62Zy8q.jpg


Lm8w6QE.jpg


cQDIAWU.jpg


:D
 
Last edited:
Stick to Kunos guidelines and you'll be fine, it's all in the dev folder on the AC_Pipeline_PUB_Rev2.0.pdf file. No point in doing a car that'll only run by itself and crap out performance when racing against a full or half grid.
300k-350k is a solid number if you want to have visible engine details, it all comes down to the LOD_B and COCKPIT_LR models which is what the game uses most of the time.
The big problem for me is the tessellating of the Nurb surfaces. Once this is done and UV mapped I can't reduce the count because it just destroys the details. I have to start from the beginning and re-tessellate which means I will also need to re-bake the AO and create new UV mapping. In a nut shell, there goes my holiday. Maybe this will be a good thing because if I want to continue with my Nurb models I need to really refine my process.
Thanks for all the input.
 
What kind of E-type are you planning to do ? And how seriously ? :)
The blueprints I have are for a 1961 Jaguar E-Type 3.8 Coupe. I'd love this to be a serious project, but I have been working on the rear tonight and the blueprints seem to be much more inaccurate than I first thought. Pair that with my amateur modeling techniques and I'm really not sure how much validity the model even has anymore...
1uLdTlv.png
 
The blueprints I have are for a 1961 Jaguar E-Type 3.8 Coupe. I'd love this to be a serious project, but I have been working on the rear tonight and the blueprints seem to be much more inaccurate than I first thought. Pair that with my amateur modeling techniques and I'm really not sure how much validity the model even has anymore...
1uLdTlv.png

That is really kool ! The hero from this Goodwood MM. Rear is so much easier than front. Shapes looks ok from these direct angle views. The trickiest parts are always "corners", you really have just understand them, and this is hard to do, especially with Sayers curvy aerodynamic designs. From this view lines looks good, rear end too.

I have also modeled an E-type a bit, IDK when I will continue on it, I also have D-type, same state, just a start. My E-type front kinda looks still off, also windshield starts way too early and too close to front... but those Lightweights has different body panels than the one you are doing, almost a different car :)

Maybe we could possibly share/do different assets for it, share knowledge, I'm sure there has to be same parts in those two cars, or could be.

Etype: https://imgur.com/a/QrXat
Dtype: https://i.imgur.com/qmrtPz4.jpg
https://imgur.com/Ts6b3z
 
Out now...

http://www.racedepartment.com/downloads/audi-r8-lms-44-magnus-racing-sebring-12hr-2018.20165/



Hmm.. what next?? :laugh:
Maybe will update the #29 Land skin as I've learned a lot since then. Other than that probably back to some Porsche 911 GT3 Rs again, either from the IMSA grid, or do the #991 from Bathurst which is still on my (ever growing) list! :D Also I have my version of the Jamec Pem/Valvoline Audis from Bathurst which are finished, am deciding if I should do the race winning #37 to add to them or not before releasing.
 
The blueprints I have are for a 1961 Jaguar E-Type 3.8 Coupe. I'd love this to be a serious project, but I have been working on the rear tonight and the blueprints seem to be much more inaccurate than I first thought. Pair that with my amateur modeling techniques and I'm really not sure how much validity the model even has anymore...
1uLdTlv.png
I know is not exactly the same car but I did this model back in 2009, so if you want it I'm willing to donate it for a good cause :cool:
screenshot023.jpg

screenshot022.jpg


Of course it wasn't done with AC in mind at all, in fact is not meant for any game, but maybe it can be used as a base to work on.
 
Busy working on updates for a lot of my skins:
  • Missed the 2K version off the #44 Magnus Racing Audi so released that.
  • Finished the update for the Lanzo DAMS Scream R-GT fixing the door number font and adding the other car
Next up will probably be release the Jamec Pem Audis - I don't think I can be bothered to do the #37 WRT car. I might do it at a later date if they used the same livery at the next round of the Intercontinental GT Challenge and/or Blancpain. After that I think I want to update the #29 Land Audi as I can make it better as I've learned alot since I did it. Then maybe start something new lol
 
Tracks almost doesn't need any work with physics, gadgets, animations, moving parts, sound. Also because they are greater scale precision is less demanding. I still wouldn't do any real important track without LiDAR ever again. Because you'll work your ass off trying to make it precise, and people will still just don't give a damn about it, because it will not have the scan label, and spoilage is real. Unless it is some novelty so precision is less desired.

Track will be more difficult the larger it is and the more variety is there. For example if your track is running through one type of environment it will be easier than it would be if the surroundings of road is constantly changing very much. More objects around also means it will be more work. The more work you'll be able to do with as few splines as you can the easier it should be.

Tracks requires more consistency of your work quality. If you'll mess up at beginning you'll have more pain than you would with a car, unless you'd rush really far into progress with a car... it should still be easier to tweak things out IMO than for a messed up things at the track. I have experienced both. Though some modders here would remodel whole car if they mess up, rather than tweak it out...

If you'd do car, then try learning photo matching early. Most blueprints usually are off. Otherwise you'll see the pictures of a car when you'll be closing your eyes. And possibly go insane. But you'll understand the curves well :D Well if you will understand them for once haha car can look totally perfectly modeled from one angle, and completely off from different angle.

For a car I would also look for less curvy car, with not too complex interior, also would check if it is symmetrical, for example I have been modeling D-Type not long ago, comparing to XJ13 and E-Type it is very asymmetrical, obviously it makes it a bit more difficult.
Would Blender and Gimp be enough or do I need more programs
 

Latest News

What's needed for simracing in 2024?

  • More games, period

  • Better graphics/visuals

  • Advanced physics and handling

  • More cars and tracks

  • AI improvements

  • AI engineering

  • Cross-platform play

  • New game Modes

  • Other, post your idea


Results are only viewable after voting.
Back
Top