The "What Are You Working On?" Thread

[AadHofman]

Sick. I've spent a few years working on GT-R's in sims. Lots of measurements and countless hours scouring through documents and speaking to owners and whatnot for anecdote and data. You got yourself into a real rabbit hole now. Probably took me at least a few hundred man hours to sort out the electric systems.

After the R33 I'm involved with releases, I could share my ATTESA and HICAS simulation with you, if you want. They're really very realistic, and although it remains a mystery what *exactly* they did to the ATTESA in the Group A or N1 cars, it can be roughly worked out if you figure everything else out. The TAISAN STP R32 is one I've always wanted to make, given how much we know about the specs. Maybe, some day...

Sharing ATTESA and HICAS would be highly appreciated indeed. I've got pretty basic AWD2 controllers at the moment, because anything more complicated turned the car into pure RWD, don't understand why. 4ws I just copied from r34, and haven't fiddled it with at all yet.

Regarding other work, I made a new suspension geometry for Toyota Supra, as it is provided by the suspension program I'm using ("measured with digital micrometers on a lazer level lift"). It is quite different compared to Kunos', and it drives better, which was a nice surprise.

 

[Kyuubeey]

@AadHofman

I know you understand, but I am reluctant to share anything *before* the main GT-R mod I am involved with comes out. Once it comes out and my systems go public for everyone to download, I personally don't have qualms with sharing everything and anything provided you credit me. I'll even help you with data and some anecdote I know and teach you how exactly to use the system. There are some limitations and a basic outline you must understand. However it works on the principles of the *actual* system, as much as I could make it. I could easily turn my R33 ATTESA and HICAS into an R32 ATTESA and HICAS, for example.

I have a history of my work on the GT-R being taken and used, without any credit, so I'd at least like to have one release for myself, first!

Here's my 86 physics in action. I OWN NONE OF THE 3D, I GOT THEM OFF THE NET.

N2 AE86 vs Keiichi Tsuchiya AE86 replica.

Same 195mm semislicks, same 174ps 4AGE.

Equal lenght links, adjusted panhard, negative camber roll center adjuster for the N2 car, all that jazz.

N2's on TRD Blue short stroke 12/10 and the DK car is on TRD Blue short stroke 8/7. It's a replica of Tsuchiya's early 2000's car. Not sure what exact team I want the N2 car to be yet: if someone models me some cars, I will find the data for those specific cars! For now it'll be a generic 190-some hp car on TRD parts mainly.

Oh yeah, the N2 car also has reduced ackerman geometry. I've been led to believe N2 AE86's don't actually run anti-ackerman, at least when they race at Tsukuba, and it didn't feel right to me either. Right now it's on a Quaife 13:1 quick rack with a bit of pro-ackerman. Would need to do more research to figure out some truths. Feels good though.

 

[AadHofman]

Well, I post it here too. Very first draft of Godzilla-sounds:

 

[Leonardo Ratafiá]

@AadHofman

I know you understand, but I am reluctant to share anything *before* the main GT-R mod I am involved with comes out. Once it comes out and my systems go public for everyone to download, I personally don't have qualms with sharing everything and anything provided you credit me. I'll even help you with data and some anecdote I know and teach you how exactly to use the system. There are some limitations and a basic outline you must understand. However it works on the principles of the *actual* system, as much as I could make it. I could easily turn my R33 ATTESA and HICAS into an R32 ATTESA and HICAS, for example.

I have a history of my work on the GT-R being taken and used, without any credit, so I'd at least like to have one release for myself, first!

Here's my 86 physics in action. I OWN NONE OF THE 3D, I GOT THEM OFF THE NET.

N2 AE86 vs Keiichi Tsuchiya AE86 replica.

Same 195mm semislicks, same 174ps 4AGE.

Equal lenght links, adjusted panhard, negative camber roll center adjuster for the N2 car, all that jazz.

N2's on TRD Blue short stroke 12/10 and the DK car is on TRD Blue short stroke 8/7. It's a replica of Tsuchiya's early 2000's car. Not sure what exact team I want the N2 car to be yet: if someone models me some cars, I will find the data for those specific cars! For now it'll be a generic 190-some hp car on TRD parts mainly.

Oh yeah, the N2 car also has reduced ackerman geometry. I've been led to believe N2 AE86's don't actually run anti-ackerman, at least when they race at Tsukuba, and it didn't feel right to me either. Right now it's on a Quaife 13:1 quick rack with a bit of pro-ackerman. Would need to do more research to figure out some truths. Feels good though.

unfortunately there are no scratch made AE86 other than the ugly Kunos ones. All other 86s are rips but they look much better and closer to the real counterpart..
I doubt anyone would ever do an 86 from scratch.

 

[garyjpaterson]

https://i.gyazo.com/89acf9fead3f3c25a8aeac497e3f8846.mp4

 

[Kyuubeey]

unfortunately there are no scratch made AE86 other than the ugly Kunos ones. All other 86s are rips but they look much better and closer to the real counterpart..
I doubt anyone would ever do an 86 from scratch.

The shame is that the Tsuchiya car for example has something like 85% scratch made 3D by various authors around the net. The body is completely scratchmade. I think it's just the interior dash that is KS and hence the car is a rip. But really, if all these authors got together and gave their work to someone who will standardize it and make a full interior from scratch, we would have a mod that can be put on RD. Hence why I'm doing something I never do and posting WIPs of physics somewhere. If anyone wants to go ahead and model them, I'd be up to it to inject these into the car. Ideally they'd be released some day.

EDIT:

And to top it off, something I started at the beginning of the year to practice 3D and get something to put my physics on, but never managed to implement in game. I'll need to also redo some aspects of the physics, but most of the geometry, aero, engine,tires etc. are quite satisfactory.

The crude wheels really do give it character.

Perhaps some year, this might even reach 0.1 Alpha! However it's my first proper car mod, and I'm not very serious about absolutely getting it out in a timely fashion, so prepare to be disappointed. I'd like to complete it though, the physics are quite good and it's a nice base for a variety of styles of build.

 

[Paul Clark]

just finishing new guard rails @LRP

and starting these kudzo foliage designs on the back rails. 1 foliage plane higher poly res than the fence place, with some deformed vertices to make appear in and out the fencing links. and still to make some further 3d additions for hedge width. Suitably realistic I think...

Hi - love you work. This may be cheeky, but with my limited skills and time, I have been creating the Lotus Test track at Hethel in Norfolk . One element I am really not happy about is the guard rail, as IRL they have FIA fencing like your 3 tier , but Race Track Builder only has thin 2 tier. Would you be willing to share an xpack with a variant of your guard rail - then I can improve the feel of the track and post as WIP. Link for current WIP to prove intent appreciate if the answer is a no, but not having the 3d skills or time , I thought I’d ask .

 

[Bakasan]

I doubt anyone would ever do an 86 from scratch.

Careful what you wish for...

A fellow Yamashita Tatsuro fan I see

 

[Mascot]

... I have been creating the Lotus Test track at Hethel in Norfolk

Whatever you do, don't mention this to Lotus.

My first Track Hethel Circuit

 

[Kyuubeey]

Careful what you wish for...
View attachment 284929

A fellow Yamashita Tatsuro fan I see

Fantastic. I actually did physics for a Levin Coupe specifically. How did you know? LOL

 

[Leonardo Ratafiá]

Careful what you wish for...
View attachment 284929

A fellow Yamashita Tatsuro fan I see

uh, yeah, I forgot you were doing that you gotta open a thread!!

 

[Pixelchaser]

Hi - love you work. This may be cheeky, but with my limited skills and time, I have been creating the Lotus Test track at Hethel in Norfolk . One element I am really not happy about is the guard rail, as IRL they have FIA fencing like your 3 tier , but Race Track Builder only has thin 2 tier. Would you be willing to share an xpack with a variant of your guard rail - then I can improve the feel of the track and post as WIP. Link for current WIP to prove intent appreciate if the answer is a no, but not having the 3d skills or time , I thought I’d ask .

sure I can do that. ill pm you shortly in a few days..

in the mean time I had a quick look see at my assets together on LRP. some basic balancing, but lots more to come. and lots of tree pruning , I better call Dr Greenthumb !

 

[Muzzy]

@Pixelchaser

You really should open a dedicated topic for your work! It's time for some track porn!

 

[Pixelchaser]

@Pixelchaser

You really should open a dedicated topic for your work! It's time for some track porn!

possibly. I like the idea of it just appearing one day out of nowhere all finished and that's it. that day is coming soon I promise. I just did some statistical analysis of everything combined together and well id say my polycounts are at about 60% full, texture memory around 42%. plenty room for lots of spectators and buildings.

 

[Kyuubeey]

Although the work now is mainly physics (The one thing I can do!), it'd be nice if I could get the model at least to an "alright" level. I think I would release it quite early in that case. I'm not so stingy about the 3D as I am about the physics, haha.

Question: How should I go about modelling the seams for everything? Do I keep the bumpers and windows their separate object or keep it connected for less draw calls?
Right now I model the car body in a rough manner, then I separate parts that have a seam, make the seam and attempt to retopo manually to get it to flow, then join the object. Is this the right way or can/should bumpers, windows and stuff that might animate just be separate? Even if they are, is it good practice to have it flow anyway?

Right now it's bumpy which I can easily fix with a bit of tweaking but the face density is too much on the body, especially around the seams. Should I try to use triangles "on the ends" to merge small sections into bigger ones? This is with a 1x catmull-clark.
I haven't really done any 3D before, so I think it would be okay to produce a sub-par result (Compared to the 3D gods here) that can still ship in a reasonable time. Learn from it and then do another car, or just redo this one later, with new knowledge. Is that reasonable or should I just keep working this one as much as I can?

Sorry for the frequent and not so exciting updates with little content. It feels like a big job to me to model anything, ahah. The physics however are something in my expertise, and progressing well. Even found damper data for such an old car. ;D

EDIT: Removed some redundant loops.

EDIT 2: And cleaned up the nose even more. No pic because I'm just spamming at this point. Simplification of the mesh will probably make things easier later.

EDIT 3: Curious to see how it'd currently look with some stuff separated and smooth shaders. Passable for babby's first 3D I guess. Needs a bit more definition though.

I'm figuring out the engine as we go, so by now an experienced modeller would probably have most of the exterior done already.

 

[garyjpaterson]

GP racing these days, its all about managing tyres, drivers can't really push, not like the good old days...

Oh, so exactly like the good old days then

 

[Ben O'Bro]

Although the work now is mainly physics (The one thing I can do!), it'd be nice if I could get the model at least to an "alright" level. I think I would release it quite early in that case. I'm not so stingy about the 3D as I am about the physics, haha.

Question: How should I go about modelling the seams for everything? Do I keep the bumpers and windows their separate object or keep it connected for less draw calls?
Right now I model the car body in a rough manner, then I separate parts that have a seam, make the seam and attempt to retopo manually to get it to flow, then join the object. Is this the right way or can/should bumpers, windows and stuff that might animate just be separate? Even if they are, is it good practice to have it flow anyway?

Right now it's bumpy which I can easily fix with a bit of tweaking but the face density is too much on the body, especially around the seams. Should I try to use triangles "on the ends" to merge small sections into bigger ones? This is with a 1x catmull-clark.
I haven't really done any 3D before, so I think it would be okay to produce a sub-par result (Compared to the 3D gods here) that can still ship in a reasonable time. Learn from it and then do another car, or just redo this one later, with new knowledge. Is that reasonable or should I just keep working this one as much as I can?

Sorry for the frequent and not so exciting updates with little content. It feels like a big job to me to model anything, ahah. The physics however are something in my expertise, and progressing well. Even found damper data for such an old car. ;D

View attachment 285054

View attachment 285055

View attachment 285056

EDIT: Removed some redundant loops.

View attachment 285072

EDIT 2: And cleaned up the nose even more. No pic because I'm just spamming at this point. Simplification of the mesh will probably make things easier later.

EDIT 3: Curious to see how it'd currently look with some stuff separated and smooth shaders. Passable for babby's first 3D I guess. Needs a bit more definition though.

I'm figuring out the engine as we go, so by now an experienced modeller would probably have most of the exterior done already.

View attachment 285121

well it's looking pretty good for a first car modeling attempt!
the mesh and the flow is nice, quite balanced.
Some lines should be killed earlier, like the double lines going from the bonnet through the windshield and the roof. Should be ended before going to the windshield

for your questions, don't include the bumpers, as they are really detached object for that specific car
windows on the other hand should be seen in the mesh in order to be detached later in the process for example
Lights should also be cut in the mesh to be located, to be detached and detailled later

As for triangle, some say it's the devil himself showing in mesh modeling
problem is if using meshsmooth, triangle usually call for angles and break the cleanliness of the surface.
For low poly, it's ok to use some. n-gon though should be avoided because the export might connect random vertices together, ending in funky faces.

Now final question, should you try & redo, or go with the flaw straight away
My first car model was amazingly terrible, i did everything wrong, and by the end of it i kinda understand some mistakes, what should and shouldn't be done with mesh flow (at least some of it), and it was way easier to restart using this knowledge.
Seeing how your model is already a good base, i'd say some cleaning would be enough to continue with it
but that's just my opinion

To conclude, i'm renovating a 280z for a year, and now you have a challenge to finish yours before mine!

 

[Kyuubeey]

well it's looking pretty good for a first car modeling attempt!
the mesh and the flow is nice, quite balanced.
Some lines should be killed earlier, like the double lines going from the bonnet through the windshield and the roof. Should be ended before going to the windshield

for your questions, don't include the bumpers, as they are really detached object for that specific car
windows on the other hand should be seen in the mesh in order to be detached later in the process for example
Lights should also be cut in the mesh to be located, to be detached and detailled later

As for triangle, some say it's the devil himself showing in mesh modeling
problem is if using meshsmooth, triangle usually call for angles and break the cleanliness of the surface.
For low poly, it's ok to use some. n-gon though should be avoided because the export might connect random vertices together, ending in funky faces.

Now final question, should you try & redo, or go with the flaw straight away
My first car model was amazingly terrible, i did everything wrong, and by the end of it i kinda understand some mistakes, what should and shouldn't be done with mesh flow (at least some of it), and it was way easier to restart using this knowledge.
Seeing how your model is already a good base, i'd say some cleaning would be enough to continue with it
but that's just my opinion

To conclude, i'm renovating a 280z for a year, and now you have a challenge to finish yours before mine!

Alright, thanks for your feedback. I will continue with it to the end, is what I felt. I did some more cleanup, more things about flow and how to attach the pieces are making sense. I tried to model before, years ago, but it was too daunting. Maybe now I can do something, so I will stick with it to the bitter end, probably.

Your 280 will most likely be done before mine, I don't intend to work on this full time seriously, or I will just get burned out. I learned that lesson many times.

 

[Sergio Loro]

GP racing these days, its all about managing tyres, drivers can't really push, not like the good old days...

Oh, so exactly like the good old days then

Golden Age Classic Tracks project just was born

 

[ppylon]

Working on some Formula SAE/Student stuff. Doing last year's car as a proof of concept (I have a few fairly obvious bugs to work out) but hoping to get the physics right creating this year's car.