Lotus Elise GTR beta release

[humty]

Well here is a beta release of the Lotus Elise GTR.
Give the mod a spin and tell us in this thread what you think of it!

There is a template in the mod but there will be a better version in the final release.
So you can paint your own car but that will likely won’t work in the official release when we finalized the mod.

As this is a beta mod the car is still driverless, all the missing things will be added when we complete the mod.

Please give us as much input as you can, so that we don’t have to release updates after our final release.

As for now we are using the WGTC ingame screens but in the final release those will be replaced by new ones
Furthermore the garage screen will have a few more options you can change when we have our own screens.

The mod is called LETDGTR.
LETD stands for Lotus Elise Track Days.

Get it here:
www.trainzworld.com/upload/LETDGTRbeta06.rar

Have fun
Chris
Rob
Triz

[cooknl]

Hi Guys,

Hope we get some good comments from you as far as things u like to see changed or like about the mod.

ALSO let me remind some of you that has an earlier beta please delete the LETDGTR dir out of your Nascar HEAT mod dir. and replace it with the beta.

I wish you all a lot of fun trying this beasty out.

Rob

[Administrator]

I've just took a quick ride yet, so this is just what jumped my eye in that few laps

1. at ereas marked with an 1. there's gaps in the mesh. The verts or the neighbourig areas are not connected (don't need to), but they are not matching

2. Here you can look thru the car - there's some mesh parts missing that should prevent that.

3. there doesn't seem to be a texture mapped here ?

4. IMO this should be looking darker - no light reflection would come from here in real life. This I'd suggest to either scale down normals down to (almost) zero or map another color (not exhaust pipes, but surrounding area)

5. that brakelight flashes to yellow instead of red ? Should be red ?

6. The wheelwells would look best if mapped to black - or with normals scaled to zero length

7. On the tire rubber textured mesh normals should be scaled down to very little length. Not 0 !, but if they are - like here - =1 (100%), this looks shiny - which it never would on real life rubber surface.

On the front air intake and such parts the parts should be defined as "no blending". Actually I'd suggest to define EVERY part "no blending" that is NOT part of either the shell (= what's to be painted by user) or glas

There's no working cockpit (F1-view) in the beta ?

The engine sound is really nice, but the sound for moving offtrack is too... crisp Doesn't sound like running on gras, rather like on grabble or such. Somewhat weird

[turtlewise06]

Thanks Guys.
Did some laps at RAtlanta and was pretty (twitchy ) but stable .
Cornered well and accelerated well . Never have heard a Lotus
Elise before so dont know what they sound like ! Fine ok with me
Only when watching replay did i notice anything like what Cholerix
was describing . I notice inner fender wells . etc...
Also had an error when trying to play1 replay file crashed to desktop thats it !
Thanks Humty and Cooknl and ? anyone else for the
New toy

[cooknl]

@ Cholerix,

Thanx Karl for the constructive critesism, we will fix all that we can .

@ Turtle,

Thanx for the comments, dont know why ur replay goofed up but have had noprobs with any replays in any stage of pretesting sofar.

Rob

[triz]

Hi guys,

Hey Karl, thnx alot for the constructive views! That will really help moving us along indeed As this is our first attempt there are still many things to learn. So showing us like this what u mean is a perfect way to establish where we need to improve. Ifanyone else has any remarks pls post them, thats what a beta is for

Triz

[Administrator]

Just noticed you have a 43 MB paint.res in the beta, but all the car paints come in the mod folder... so probably the paint.res has just been copied from another mod and has only that mod's paints ?

You'd save 43Mb bandwidth and webspace if you delete that paint.res from the beta, then pack all the carXXX.tex to a new one and put that one to the mod for download

Same for the size of part.res, ui.res and race.res - these are all at least twice the size they should be . Dunno why, but there's probably a lot of file sin these RES archive files that isn't used by the Elise mod in any way. You could downsize the mod to about 1/3 size (~ 30MB) by cleansing these RES files

Furthermore, these files shouldn't be in any mod package at all

- head0-head3.cvs
- paint0-paint3.tex
- subfolders replay/record/ghostcar/paint

The file track.ini should ONLY be in a mod if you want to limit the tracks to use with this mod as that's it's function

The file ELISE1.TEX in ELISE.CAR won't show up anytime - it can simply be deleted as it's without function. It's only use was being mapped to the car mesh in your modeling program. It's just a placeholder for the carXXX.tex files that Heat will look for. If Heat won't find a certain carXXX.tex, the car will show up red - but it will never show up ELISE1.TEX

[Administrator]

Some remark on physics - from some academical point of view, not from testing

You have put in a lot of front downforce - actually 50% more front downforce than on rear. My guess would be that this is to get the car feel balanced.... the reason it probably didn't feel balanced with less front downforce is the tire physics you use in the mod. You'd better go for these - a lot of downforce will iron out some flaws with the base tire physics, but it won't do at all at slow speed for example

So far as I could see, it's probably the same tire physics that are used with quite a lot of mods (we've all copied these ;D) - and these have a serious flaw as I only found on the making of SCGTS mod (yes, GroupC mod has same flaw with tire physics): There front tire's compound attributes differ from rear tires - and define that front tires provide MUCH less grip. Not by their lesser width, but by the tire compound definitions, which doesn't make any sense IMO in terms of real life physics. The front tire's magic_b/c/d/e numbers may make some sense for an oval mod to fake some physical effects, but IMO they are at fault for road racing mods and cause that nasty understeer a lot of mods suffer from

You didn't put in any drag for front spoiler - that means that 100% front spoiler can be run anytime without any slowing down the car. My suggestion would be that increasing front/rear spoiler (to get more downforce) in garage setup should slow down the car (= by additional drag), like it would in real life. Else people would run max downforce all the time

[cooknl]

Hi Karl,
RE: The physics i would like to be able to talk to you about the magic values for tires I did not change these at all because I do not understand what each one is for, If I knew what they vary then I would possibly be able to implement them correctly.

For now I only played with the width, latitude and longtitude variables to get it to handle like it does. And yes u are so right it needs downforce speedlimiting implemented, I do understand how that works, but to be honest I forgot to implement them. .

And as far as the RES files is concernd I was aware of this, but not really sure yet what is exactly needed and whats not, therefore I left all, just to be sure the BETA would run properly for everyone.

Maybe we can get together on Ventrillo one of these evenings and we can talk about some of the things needed.

Thanx again for your contribution here, its really appriciated.

Rob

[Administrator]

that's where the physics tutorial comes in ;D

NHeat's tire model is based on Pacejka's Formula, which is a standard for tire physics models. It's used in most race car simulation, not just games. There is a great tutorial for Pacejka's Formula over there at RACER, a freeware race simulator. Another (scientific) publication to get a basic understanding of how it works can be found here www.control.isy.liu.se/~fredrik/reports/05automotiveDSP.pdf

Basically Pacejka's formula is

y(x) = D sin[ C arctan { Bx - E ( Bx - arctan(Bx))}]

where the variables D,C,B,E are "magic numbers". Adjusting these "magic numbers" alter the modeled tire compound's characteristics. Some useful reference for how to work on the magic numbers can be found here controlfreaks.mne.psu.edu/Publications/Theses/2006_Glumac_Honors.pdf

By playing around with compund definitions, you should always only adjust one magic number at the same time, unless you have a very good understanding of what you're doing - and you should change the numbers only in very small steps (+/- 0.1 or 0.01)

In NHeat the magic numbers of Pacejka's Formula are

* magic_b ( ~ 0.7-0.8, +/- slip angle)
* magic_c ( ~ 1.1-1.5, +/- grip on sliding)
* magic_d ( ~ 1.0, +/- overall grip)
* magic_e ( ~ -0.2)

This is only a very rough description of how these numbers work in NHeat, but since there's no manual which way exactly the devs have implemented Pacejka's Formula, there can be no more than just estimations, trial and error

Easiest approach this case would be to use same magic numbers for front and rear tires and check which way it takes effect

[DaveO]

Been enjoying reading about all this physics stuff. Favorite subject has always been Math

Check out this link about Pacejka Magic Formula. Great read.

www.racer.nl/reference/pacejka.htm