The "What Are You Working On?" Thread

When making a car there are couple of trouble points that are difficult to get right. One is the overall shape. This means the nose is typically rounder than it looks and the car is narrower at the front and rear than it is in the middle. This is easy to get wrong which makes the car look boxy. When working using those image planes there are also couple of places where the image planes typically suck. The front A-pillars are typically difficult to smoothly shape to the hood and sides. Better make these shapes as good as you can before you move to making window seams for example.

It is also tempting to make holes to the shape for things like door handles, mirrors, lights and seams but personally I'd avoid making those holes as long as possible. You can still make the door knobs and other bits as separate objects and it is a lot easier to adjust the body shape when you don't need to worry about keeping the annoying little bits smooth.

You can and also should look at how professional modellers have made certain parts of the car. Look at the wireframe around the fenders and around the window edges and see how yours is little different. Pay attention how the faces loop around the shapes and holes how and the width of these faces changes when they are closer to edge.
 
Bremgarten 1954 - up in the usual place....now we really need some more classic 50s cars to go with the Maserati , Have Fun!!

20190108173407_1.jpg
20190108173345_1.jpg
20190108173237_1.jpg
20190108172924_1.jpg
 
When making a car there are couple of trouble points that are difficult to get right. One is the overall shape. This means the nose is typically rounder than it looks and the car is narrower at the front and rear than it is in the middle. This is easy to get wrong which makes the car look boxy. When working using those image planes there are also couple of places where the image planes typically suck. The front A-pillars are typically difficult to smoothly shape to the hood and sides. Better make these shapes as good as you can before you move to making window seams for example.

It is also tempting to make holes to the shape for things like door handles, mirrors, lights and seams but personally I'd avoid making those holes as long as possible. You can still make the door knobs and other bits as separate objects and it is a lot easier to adjust the body shape when you don't need to worry about keeping the annoying little bits smooth.

You can and also should look at how professional modellers have made certain parts of the car. Look at the wireframe around the fenders and around the window edges and see how yours is little different. Pay attention how the faces loop around the shapes and holes how and the width of these faces changes when they are closer to edge.
Thanks for the insight. The blueprints are indeed not that good, even some landmarks are simply wrongly placed, according to pictures. Well, with my ability, the shape will be off slightly no matter what I do, so I will just try my best. Some things are intentionally boxy or low fidelity because I don't know how to define it without needless loops.

Right now I've thought about taking away all the seams and going back to the basic shape. Sculpt everything in one large solid, without any regard to any seams or anything, then separate the 'problem areas' into their own object, work it freely, then re-topo by hand the door back into the body for example. It will be a separate object in the end, but I think it's good to attempt to still maintain flow, isn't it?

excusemewhathe****.PNG


The geometry on the left side of the door seam setup lines, next to the A-pillar. What the hell is that? I can probably avoid that if I approach it correctly and not make it up as I go.
 
As for maintaining flow, depends whether the actual car bothers or not. Usually you can tell by how continuous the reflections are.

458italia.jpg

eg. the 458 door to rear quarter is smooth

E30-M3-Photo-gallery2.jpg

M3 E30 they just did not bother

For the most part you also want the flow to parallel reflections of the horizon, cause reflections go zigzaggy if they cross faces in other directions (but that's not always possible)
 
@Stereo
Interesting! I guess BWM did not get the memo on polyflow :p.

Perhaps I will attempt to maintain flow simply because if you tell a beginner to not care, it will probably end up in a complete cluster**** with 500k tris. Haha.

Is this a better direction?

newtopo.JPG
 
Yeah, although the fewer you can leave the better.
O5vVrvP.png

Added white ticks showing how the overall (away from doorhandles, etc.) spacing of faces is large but then comes closer together towards the ends of the car. If the shape's changing less quickly, use less faces. As far as the A pillar I forgot to make the front window wireframe visible but it tapers closer together (so top and bottom are both 3 columns) rather than having vertical and horizontal rows meet.

You can also see the side markers aren't modeled into the mesh at all since they don't need to be.

Should note I don't model with subdivision 99% of the time though, that changes considerations. At most I use it with very basic shapes that I want to be smooth (eg. wingmirrors are a deformed half-sphere on most cars, it's much easier to make it out of a subdivided cube since that gives you 8 control points to tweak the shape)
 
Last edited:
@Stereo

Good stuff. I'm slowly realizing just how little loops are really needed for most of the stuff. I try to be as clean and simple as possible and still end up over-modeling.

Procured some suspension measurements. I wasn't too off doing it mostly by eye.

Now to tweak the anti to get it right for AC. Thanks for this bug, KS. It's real fun to work with. ;)

20190109083735_1.jpg
 
@Stereo
Interesting! I guess BWM did not get the memo on polyflow :p.

Perhaps I will attempt to maintain flow simply because if you tell a beginner to not care, it will probably end up in a complete cluster**** with 500k tris. Haha.

Is this a better direction?

View attachment 285221

around the wheel arch you need to keep a "cylinder" looking kind of mesh :p
i usually start my cars with a cylinder... to make sure i have a clean wheel arch with spaced edges
example :
upload_2019-1-9_19-55-8.png

it gets challenging when you have hard edges, like middle of the door :)
also this wasn't game ready, but more like mesh smooth iteration 2 and plenty of polygons for rendering!

in other news :whistling:
upload_2019-1-9_19-53-45.png

and the occasion as well to see how not to model a car :D
except if it's meant to stay low poly...
 
Gone for the 'modern vintage' Dunlops used on the car today instead of the 'vintage vintage' Continentals used in 37 simply because I had much better reference. Not fully happy but look OK for now, squeezed both front and rear designs onto one map :)

Screenshot_auto_union_type_c_deutschlandring_9-1-119-21-22-52.jpg

Screenshot_auto_union_type_c_deutschlandring_9-1-119-21-23-2.jpg
You can see me shivering with antici...pation.

:)
 
@garyjpaterson

I love it. Classic cars feel really nice in AC, I'm doing my part with this. :thumbsup:

@Ben O'Bro

How is this?

Cleaning.JPG


I separated the car into an USDM 240Z and JDM S30. Under the Nissan brand, Japan initially received the S30, a lighter, more barebones version, with an L20 single-carb engine producing about 115ps. Tires were also 175/78/14 or 165/78/14 bias-plys! I have made the 165's.

The western market, and later in 1971 or 1970 I believe Japan as well, received the 240Z under the Datsun brand, sporting wider 195mm tires, radials I believe, and the L24 twin-carb engine, producing somewhere around 150ps. The interior also had real mats instead of plastic mats, and had a clock and some other stuff added. I'm not knowledgeable yet about various trim options and whatnot, it's a bit hard to figure it all out, but there's a lot of data out there. I'll figure it all out in the end.

I've got data for both, and I can probably do both quite well physics wise. Even got suspension travels and damping rates figured out. In terms of 3D, if I manage to make a LHD and RHD cockpit with appropriate mats, gauges etc. then I will be very happy. But I think, I will focus on the USDM 240Z because the specs are higher. On the other hand, the S30 would offer a nice base for an S1, with western market engine and European suspension, and it'd be a good base for any future tuning cars...

Hell, I'll try for both!
 
filledline.JPG


I think that's enough for now. Anything I could improve in the front lights area geometry to produce less pain in the future? It doesn't look quite right to me, but it's a first.

My reasoning is to make the general shape first, with a good flow, then graft the details into it. That way the lights, which are inset, will be able to flow. In theory. Once the base mesh is okayed by the experts here, and I start making details, I think it would be time to open a thread and stop spamming. :rolleyes:

Unless someone likes my posts, in that case, uhh, thanks?
 
Only tip from me is not to go into multiple models with your first car...I did too and it quadruples the work and oversight :cautious:
Perhaps, but I am very unreasonable. ;)

At some point I will have to realistically choose which to go for, S30 or 240Z, and what market. Most likely the USDM 240Z will be the most popular and appreciated model. Personally, I far prefer RHD, so...

At least, they sold the 240Z in Japan as well. From what I understand, the specs and visuals are similar to the western cars, just with driver side swapped, different wheels, some misc changes like indicators, perhaps rear lights, that stuff. I might be completely wrong, though. At least the wishbone geometry is identical in all cars. However the front strut design and length is completely different in the European cars, compared to the 240Z and S30, to make it worse! It's all a bit convoluted, this car, but there's *a lot* of data. So perhaps I will make JDM and USDM 240's. Perhaps I will never finish the car, who knows! I will try.


If I am not mistake, THIS is a 1970 JDM Datsun 240Z
Datsun 240Z front.jpg


and THIS is an USDM 1969 Datsun 240Z

NYPress22Oct69.jpg
 
As far as the headlight area goes, you want to make lines parallel to each other, not strictly in X/Y/Z directions. So that the first row of faces is about the same size, next row's about the same, etc.
4z3b7dk.jpg

These lines are following the shape of the windows at both ends and then smoothly transition towards the middle, whereas your hood line curves along the front bumper, then the next one up is far away from it.

The way you do it from the side view is about right, I think your top view blueprint is just not accurate about the shape of the bumper though so maybe not worth trusting that line too much.

Something else to keep in mind is the location of your 3 and 5 quad meeting points:
43xd4i0.png

3 quads is technically the corner of a cube, so they happen at the corner of the car.
5 quads is what happens when one of the 4 quads in a flat intersection is extruded up or down from the surface.
eg. on this hood, the center section of the hood is extruded upwards so you have 3 above 5, with all the edge lines other than the selected ones flowing smoothly around it.
Getting this kinda thing tidy is important for subdivision modeling. It ends up somewhat arbitrary on large curved panels but I think the S30 doesn't have many places where that's a concern, most of its corners are quite sharp (other than the missing one where the hood edges meet in the area scooped out by the headlight)
 
Last edited:
@Stereo

I think I understand. I thought about it too myself. I should try to keep the rectangles roughly the same dimensions, right? Even if from side or front view, the lines curve in a strange manner, in 3D view looking at the faces itself, it is smooth.

The shape of the light area is a bit of a complex shape, with some roundness to it, so we see how long it will take me to model it.

Okay, I lied. But now I need to take a longer break.

frontworking.JPG


I draw and paint at a decent level, so I can "feel" the form of the light area just looking at a few pictures, but I can't just snap my fingers and translate it to 3D. Lack of understanding of good modeling practices is probably the main cause right now. Slowly, it's getting there, at least.

EDIT: Redid the front area. Maybe a better direction?

newfront.JPG
 
Last edited:
Personally, I far prefer RHD, so...
As a Brit with the shifter on the left-hand side of my rig, that's my preference too. It's always a joy (and less of a disconnect in VR) when a RHD car mod is released.

I've sometimes pondered how much work is involved in converting a car from LHD to RHD. It seems relatively simple on the face of it, but I bet it's not: flip the main dashboard structure, steering wheel and rear-view mirror, and reposition non-flipped items like the instrument cluster, pedals, logos etc.
Most production cars seems to be mainly symmetrical down a centreline, dashboard excluded (with obvious exceptions like the TVR Sagaris and it's track-day helmet bulge in the roof).

@Shaun Clarke is one of the few modders who sometimes offers both variants, so maybe he can actually say how hard it really is. It'd be great if more modders offered this choice, assuming it's not a huge amount of work.
 

Latest News

What is the reason for your passion for sim racing?

  • Watching real motorsport

    Votes: 483 69.3%
  • Physics and mechanics

    Votes: 295 42.3%
  • Competition and adrenaline

    Votes: 319 45.8%
  • Practice for real racing

    Votes: 148 21.2%
  • Community and simracers

    Votes: 190 27.3%
Back
Top