The online racing simulator
Quote from TehPaws3D :You forgot about making sure the damage model works unless that's somewhere in what you said.. And since you seem to know more than me, Whats the problem of making cars that are based off licensed cars but not like.. Ford would be forunda or something crazy.. Is that legal or no?

That is a gray-area that a developer takes if he is willing to risk it. Some people will notice things even the devlopers never see - and puts projects at risk.

RFactor uses fictional ccar names by fictional developers....

Lets say that - since my altar ego is Morugan Kodi..... I could turn to Rfactor and say "- Hey - Your Manaufactur for some cars is Kodi - I was born before your game - you owe me money bitch" - so not only do they have fictional manufactures... but the car designs are very very loosely based on the real cars in design to avoid issues.

We had a issue in RCX when relating to RC1 and RC2 - with names such as Vostok, Team Unity, Matrox.......

.... We are entitled to use these names freely so im going to stop talking aout it - It will defeat my own point.

Our cars are inspired by RC1 and RC2 - but are intended to be more realistic - but because of the team names - if we were commercial we would be in the shit if we hadn't asked Sony Liverpool - email ( Formerly Psygnosis ).

Sony Liverpool have already cleared up names... they are awesome.
Some people just don't understand Live For Speed at all
Quote from justasimfan :
As for the models themselves how difficult is to simplify a model ? Simply reduce the polygon number...
The same with textures...I mean you can have 4096x4096 textures simply converted to lower resolutions in seconds with photoshop....

Reducing the size of a texture is not a trivial operation if you want the result to look good. And here we're talking about 2D space. Doing the same in 3D is much more complicated and reducing some 1.5M+ poly model to some 30k poly model by hand could actually take more time than creating a new low-poly model from scratch. And as CodieMorgan already explained, finishing a 3D mesh of an object is where the real work begins.

BTW, I haven't seen any XFG, FZ5 or XRR models in the TurboSquid shop, so I guess your whole idea is to replace these with something similar. No offense, but this is stupid on so many levels. We like our cars and we've known them for years, I certainly don't want them to entirely change shape overnight. Also changes of bodywork should affect how the car behaves, so replacing the model and leaving the physics as it is would feel like a slap in the face of those who actually give a crap about physics.



Can someone please close this thread now?
Quote from justasimfan :As for the models themselves how difficult is to simplify a model ? Simply reduce the polygon number...
The same with textures...I mean you can have 4096x4096 textures simply converted to lower resolutions in seconds with photoshop....

Quote from ATiRAGEPRO :.... Have you ever tried decreasing the polygon of a Freeform Mesh Object? automatic polygon pruning in 3d programs either DON'T WORK, or they leave your object looking like a six-sided turd! Most of the wire frame ends up having to be pruned manually, and there is no bigger pain in the ass than buying an object from sites like renderosity.com, and having to erase most of it's mesh so that it can actually RUN in a game. It is a f***ing nightmare to have to do that. If the Devs actually DID waste their money on those models, there is no way in hell those cars are going to run well on any decent computer, let alone laptops that only run DirectX8.1.

AFAIK the VWS was given to the devs in a hi-poly mesh and eric had to reduce it drastically to use it in-game.

Anyway. I think the OP idea is good but just because I don't see it happening anytime soon (soon = many years) I think it's not an important thing to the current state of LFS as physics should be the priority right now (followed by sounds IMHO :shrug.
Quote from MadCatX :Reducing the size of a texture is not a trivial operation if you want the result to look good. And here we're talking about 2D space. Doing the same in 3D is much more complicated and reducing some 1.5M+ poly model to some 30k poly model by hand could actually take more time than creating a new low-poly model from scratch. And as CodieMorgan already explained, finishing a 3D mesh of an object is where the real work begins.

BTW, I haven't seen any XFG, FZ5 or XRR models in the TurboSquid shop, so I guess your whole idea is to replace these with something similar. No offense, but this is stupid on so many levels. We like our cars and we've known them for years, I certainly don't want them to entirely change shape overnight. Also changes of bodywork should affect how the car behaves, so replacing the model and leaving the physics as it is would feel like a slap in the face of those who actually give a crap about physics.



Can someone please close this thread now?

I know that you cannot simply drag and drop a high polygon finished model in a game before in that case you would live everybody with older computers behind.New graphics cards can handle easily such loads but thats not the point.Definitely some work would need to be done to make simplier polygon models so you can have different LOD levels in the game.

But i hold on tight on my belief that the amount of work needed to simplify models are much much less than actually the time needed to make a model from scratch.

The real cars lfs cars are based from are a 911,a mitsubishi starion,a fiat punto,a celica gt4,a honda dc2 integra,a mini and a lotus super 7(caterham). There are incredible versions of those cars and if you leave the name to the fictionals name already exist(Or use names like conda tegra,celca gt4,etc...) and change the badges(so instead of a prancing horse you have a prancing goat for example) everything is ok.
Quote from justasimfan :i know that you cannot simply drag and drop a high polygon finished model in a game before in that case you would live everybody with older computers behind.new graphics cards can handle easily such loads but thats not the point.definitely some work would need to be done to make simplier polygon models so you can have different lod levels in the game.

But i hold on tight on my belief that the amount of work needed to simplify models are much much less than actually the time needed to make a model from scratch.

The real cars lfs cars are based from are a 911,a mitsubishi starion,a fiat punto,a celica gt4,a honda dc2 integra,a mini and a lotus super 7(caterham). There are incredible versions of those cars and if you leave the name to the fictionals name already exist(or use names like conda tegra,celca gt4,etc...) and change the badges(so instead of a prancing horse you have a prancing goat for example) everything is ok.

just give up!
I guess first of all should be the game experience, i think adding some rain would be the best, better if it can be with variable intensity, we´d have to get to know each car and each track all over again on different conditions. You can also add that probably you´ll have to find new race lines in some tracks, for example in South City, where you can find some "ponds" out there And i´d deffinetly add some tracks or cars, or both! Of course after rain But Graphics? I don´t take much time to see each detail on the car in front of me when im going over 100mph, bumper with bumper, trying to find the right place to dive in for a pass. But thats just me
The problem with adding rain - is the difficulty of cars handling like real cars in wet weather.

Something that cannot be thought of until the tyre model is finished - but when wet weather developments - someday begin, it might mean the tyre model could have more flaws found in it.

Be careful what you wish for unless you are prepared for the wait.
I think that Scawen is well aware of that and I'm convinced that after a year and a half of development he made sure the new model will work in the wet without any serious changes in it.
@MadCat: U can never know until something like that is tested =)


Quote :As for the models themselves how difficult is to simplify a model ? Simply reduce the polygon number...

WTF - THAT is impossible.

Quote :But i hold on tight on my belief that the amount of work needed to simplify models are much much less than actually the time needed to make a model from scratch.

Wrong - that is easier said than done.


Get Car - Start Editing....

After you optimize and update mesh:

1: Error Correction, Rats Nests should be checked - HDRI support if required should be checked.

2: Because the mesh was edited all UVW texture cordinates for all channels except reflections will be invalidated...
.. you now have to remap all UVW coordinates to the precise detail and fitting that it was before - or even redo it completely.

3: Any burn-In Radiosity ( or BSP ) texture layers have to be done

4: Testing - and repeat steps 2 - 3 until correct....

Time - 2 -8 hours for the UVW fixing by itself - nevermind the mesh editing and optimization - a High Poly car will take a quite a while.


Opinion: Its always better to model a car from scratch - you know you own mesh better than anyone else. Especially since you don't have to deal with Import complications from other non-native formats.

And no - you cannot use NURBS in this case..... in case anyone starts thinking about it.


5: Now you have to deal with IK linkages for Suspension parts....

... etc ....


What MadCatX said is right - Reducing a high poly model is not a trivial task.

Utra high detail imports will take - well - ages - if you don't want to **** it up.


Just so you know - Correct UVW unwrapping for texturing, as well as multi channel texturing ( for shaders too ), will sometime take longer than actual modeling itself.



I spend alot of time testing individual ( and generically un-announced ) car models for RCX ----

This is how I test weather there are too many polygons on my models: ( we want at least 20 cars on track - maybe 50, about 100K vertext maximum per car, and a 200K maximum environment before it gets segmented )

The car in the middle is drivable, the rest were swapped with Physics Test Geoms for dropping and testing framerate in out WIP racing engine.


I pushed up to 50 cars, But alot of engien work itself is needed to enhance rendering performance before we code in textures....


This car's mesh is considered mostly complete. The part I hate - is described above - because ill have to go through it later on.


On another note: Try to guess how my original RCX Car #8 counts - how many polygons do you thing are in it? ( Excluding wheels ) You will be pretty surprised.

can we get bigger pic? maybe 20000 ?
Quote from e2mustang :can we get bigger pic? maybe 20000 ?

You can get a bigger version when u find my photobucket - or my facebook. =) Good luck.

There were a total of 14 cars modeled , only 4 models have been used ingame in the 2 year period so far and only 5 of them are 'official' ( because I have the power to mark any 3d content as Official or not )...


JustaSimFan.... ( and the rest that bitch about slow development )

Try to understand this: The following is the same kind of stuff ( but simpler ) to what scawen has to look at daily. Scawen is a great math expert in my eyes. Scawens tyre model - is likely about 10 times more complex i believe.

Slinger is trying to work out realistic Tyre fixes atm due to our cars spinning unpredictably and incorrectly at times over bumps - after stating that "even though RCX is based on 2 previous arcade games, he sees RCX as a simulator in terms of physics".... which is why:

Quoted by Slinger(mat) in the RollcageX Development Project:
Quote :btw: I finally implemented "combined slip" for friction. It was done after a long time of trying out how the ode friction really works (you notice the comments). Anyway the current solution is based on the following:

The friction should be scaled like a circle, and we begin by defining X and Y to be how much X and Y should be scaled:
new MUx = MUx*X
new MUy = MUy*Y

Now, we want the X and Y values to "weight" the original mu values as a circle:
1 = X² + Y²

Lets say we got a way of specifying how much Y depends on X and vice versa:
Y = K*X

which means:
1 = X² + (K*X)²

and we can then figure out the X value:
X = sqrt(1/(1+K²))

And since:
Y=K*X -> X=Y/K
we get:
Y = sqrt(1/(1+1/K²))


So what's left is to figure out how to define "K".... This is something I've been trying to find an answer to the last few months...
*a usual friction would be:
K=Vy/Vx

*a similar variant of the first one:
or possibly if using slip values: K=one_koefficient*slip_angle/slip_ratio

(so we simply rescale based on velocity along x and y).
*the one I'm using now (based on a thing I've found used for ADAMS/TYRE):
for x K = (MUx/ypeak)*sin(slip_angle)/slip_ratio
for y K = (MUy/xpeak)*slip_ratio/sin(slip_angle)

and what I've found to mimic the pac2002 (official pacejka MF I think) solution:
for x K = one_koefficient*slip_angle²
for y K = another_koefficient*slip_ratio²

Even though he is describing it - the "average sim racer" is unlikely to understand this.

Hell - not even I get half of his formulas... gets confusing very fast.


Try it . Its a great challenge - and challenging things such as this are always fun.
This is why Scawen is developing LFS in the first place . because - even though it is "work" - he enjoys programming.

Like any artist - development is 'artistic and creative'. Regardless of what you code, You are always 'creating' something, a part of something, or a friend of something that makes something else do something really coll, or creative.

Generally early programming can be easy to learn - until the time comes when u want to code something as big as this - then the 'one project' becomes your occupation for 4, 5, 9 or 20 years.....

RollcageX is potential to us. And like Scawen says " will be finished when it is done".

It is impossible for a developer to accurately predict how long it will take to update some code.

One bug fix can lead to many - you end up following a chain that leads into another chain , and then into another 2 or 3 chains of linked errors. Then you have some obscure thing floating off were you dont see it - that will eventually become a problem later.

That obscure thing turned out to be the physics in LFS. The Tyre Model failed to demonstrate acceptable realism when a real car was tested on it - and the ABS failed to behave like it should.


So Scawen starts fixing it... but in the process, Scawen finds a bug, or an issue - he follows it until he ends up with various areas in the code that needs to be updated.

In the end the code is updated and some testing is put in place - he finds it isnt "fast" enough for slower PC's....

.. so what happens?

.. Right back into various areas of the code again - to now optimize it .... while fixing other bugs "as he finds them".

Because of all this - Content updates are something that will get in the way of his productivity, and shouldnt be dealt with at this moment - because he has been working on so much code - that has to deal with the addition of a single car....


So Please..... just stop this already. LFS is doing fine as it is. If you can understand the amount of work this is - even for Eric too.

Just because you havent seen much from Eric doesnt mean he is doing nothing. Keep in mind that Rockingham suddenly appeared out of nowhere - and it was already 'mostly' complete.

^ That is proof that - Eric could very be working on something - and just like Scawen , prefers to stay quiet until he has something worth while to show. Most likely.

It makes me sick when idiots run around calling the developers 'lazy'... and say that 'lfs' lacks content - when the fail to keep in mind that - with 3Developers in LFS , versus 10 , or 20 developers in another sim or game... is a major difference , and therefor CANNOT UNDER ANY CIRCUMSTANCES BE COMPARED IN TERMS OF DEVELOPMENT SPEED UNLESS DEVELOPMENT ACTUALLY ****ING STOPS ALTOGETHER.


Got it?


...

And yes - I am relaxed at the moment. I have too much work on my hand.
i work with 3ds max too. what do u use for landscapes?
Internally - there are plenty of options for Terrain Generation.

ATM im using only internal (default) modifiers and material-types to generate terrain.

3DSMax 2010 is good enough for the job. Even if creating the terrain and maps for each required Generation process is ... longer....
Saving polys in a dds? What? Do you even know what dds is? Its a file format for textures. Why would you have 3d polygons in a dds file? 1,000,000 polygons x32 + a track + high res textures can't be handled by todays pcs.
Quote from master_lfs.5101 :Saving polys in a dds? What? Do you even know what dds is? Its a file format for textures. Why would you have 3d polygons in a dds file? 1,000,000 polygons x32 + a track + high res textures can't be handled by todays pcs.

Wait - who said anything about DDS? ..... re-read the post.

Completed Ultra detailed real car models
(94 posts, started )
FGED GREDG RDFGDR GSFDG