rec.autos.simulators

A quick update to address some issues:

- Shift-R / Shift-P now work far better; they also reset
  the suspension and wheels.
- The car's physics have been tweaked a little bit.
  It's a bit more drivable now.
- No display lists are being used; this didn't seem
  to make much of a difference on my card, and it
  saves a lot of memory (I tried a lap on Monza,
  but it was swapping too much at 256Mb). And no,
  Monza is not downloadable on my site (on others
  though, Sean's conversion from F12K).
- Timing is improved; it should have better resolution
  now. Although you won't see it through the car's
  behavior, the Physics% and Graphics% are much more
  stable now.
- Icons are created in the installation. How handy!
- Curved was missing from v0.39. My mistake.
- The camera pitch follow has been improved, and you should
  be able to see more clearly now where the car is
  going in the default view.

Enjoy!

Ruud van Gaal
Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.racesimcentral.net/
Car simulation: http://www.racesimcentral.net/

1.    I downloaded the v0.39.1 and installed it.  Installation was no
problem however in the start menu/Racer/... all the icons there are not
linked to anything.  I have to start each application by starting Windows
Explorer and going to the directory and double clicking on it to start the
app.

2.    I started the config.exe and set my controller for the default(?)
joystick controller.  I have a home made 4 axis controller running through
my sound card.  when I click on "run" while still in the config application
I get blue screened to death.  However, if I run that and then exit it and
start the sim through the Racer.exe, The sim starts up nicely.

3.    Once the sim does start, the car is sitting in pit lane for a split
second then starts hovering upwards as if it were a helicopter and continues
to climb.  Hitting Shift-R or Shift-P I'm placed in the pits.  Sometimes I
land just right while other times I land and start bouncing and have to
Shift-R again and again till I've landed without bouncing.

4.    The controller issues don't bother me because I know this is a work in
progress.  With my controller, My gas pedal does work for the sim. I can
accelerate nicely but decelerating will put me in to a slide and depending
on how I'm situated I'll either continue to spin forever or I'll spin and
act like that helicopter again and start to fly.  It's as if gravity works
only for objects that are touching the ground but has no effect on anything
one inch or more above the ground.

5.    The supplied oval track has a glitch in what appears to be T1.  A bump
or something that throws the car in to a wild spin upwards in to the sky.
Also, using the access road after the pits, (or is that part of a road
course?) In the second left about the same place with the ovals T1, there is
a bump that throws the car to the inner wall where it again spins the car up
in to the air.

6.     I can't shift out of first gear.  But again I know this is a work in
progress..

    BTW: the car looks great.  and I was surprised that the graphics took
70% and the physics took 3% of the CPU. :-)  If there's anything else you
need from me just let me know.

You forgot to mention the OS and machine type you're using. That might
help a little bit too.

For a default install, I get the following locations:
Target: C:\SIMS\RACER\modeler.exe
Start in: C:\SIMS\RACER

This is InstallShield's macro <TARGETDIR> at work, I would hope it's
quite a stable installer. Anybody else have the same problem?
What do you get in the installer as a default installation directory?
The C:\SIMS\RACER?

That's weird, the blue screen. Anyone else have this problem? Might be
something with the stack size from the MFC program, which exec()-s the
Racer.exe, which does NOT use MFC. But it may then use the same stack.
Just a feeling.

As for the controller, you can (copy &) create your own controllerset
file in data/ctrlsets. Just copy the joystick one and modify according
to your available axes. Give it a new misc.displayname and Config will
pick it up automatically.

As for picking the numbers, which might be a bit confusing, I'd highly
recommend the JoyImm tool which is also on the download page. There
you can see the axes in action, and plug in the appropriate numbers
you see there in the config file to get what you want (this will
allow/disallow single or split axis for throttle/brakes).
No clutch is in working order yet.

Yep, a known problem, which I think I reported during v0.39. It's
still a mystery (I'm crowded with work, from morning to evenings at
the moment), but when I turn on draw_every_step=1 in debug.ini it
seems the first frame is whacky and only then the car warps to the
grid position. Even though it is warped *before* the 1st frame is
drawn (at least that's what I think).

The bouncing thing is a crucial bug. It can jump higher than it
originally landed. But I also have a direction in which I could/should
go to fix this and get some extra bonuses for free. So I'm going to do
it the right way and fix it much better than a hack right now.
Note that to aid further assistance from people, I intend to create a
2nd view, more scientific, which will show for example the lateral G
force, the weight distribution for the 4 wheels, and things like the
energy contained in the body's system.
This should give pointers to me and physicists out there as to what
might be a problem. And ofcourse it's cool to have a lateral G meter
as well. ;-)

I suspect this to be caused by a) weight distribution still being
50%/50% (too much weight is in the front), and now that I think of it,
I should read up on anti-pitch values for suspensions. Since nothing
of that kind is in there right now (there ARE rollcenters, so lateral
roll is already greatly reduced and better), no anti-pitch is present.
So basically, braking puts almost all weight on the front wheels,
which I think (when testing with my own car) is only achieved in
really heavy braking situations.
Another good reason to put in a longitudinal weight transfer meter in
the view. The lines pointing up in the wheels (black ones) indicate
the normal forces. When I brake, the rear ones disappear into the
model, and the front ones come up high. I bet if they're twice the
height as they normally are, the weight is almost exactly 100% in
front.

I can tell you that it works everywhere. I've thought about a crude
collision detector. What happens now is when you hit a wall, it thinks
the road is suddenly 20cm or so higher than normal, so it generates
HUGE forces in the suspension and tire. That's what's kicking your car
into the sky. In fact, let it fly and you'll see the speed meter come
down and go up again as it reaches its highest point.
So IF there is suddenly a huge difference in road height, I can also
see this as a collision and start a hard velocity bounce.
Would be nice (but a bit unrealistic), since the car is hard to keep
straight. I did a Monza lap in 4m25s hehe. Not the best ever, but I
had to avoid the crowd. ;-)
(hm, I'm going to put in fog as default I think, much better)

Yup, I noticed and reported this with v0.39. Need to check what's
under the surface.
I'll have to review the track objects; I don't think every object is
needed. And I'll have to do road flags too to avoid doing road
calculations with the cranes for example (!!!! good enough for a quick
travel to Mars !!!! (c) tm Dolphinity in case NASA is reading this).
;-)

It's high on the list to get controller buttons working for shifts.
For now, use cursor up/down.

You mean they don't add up to 100%? :)
That's because there are 3 other timing areas, somewhere in swapping
mainly IIRC.

Thanks for the elaborate report. I hope that after next week things
will be more quiet for me, and I can patch a lot of things in.

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Car simulation: http://www.marketgraph.nl/gallery/racer/