rec.autos.simulators

GPL physics question.

4408

GPL physics question.

by 4408 » Sun, 31 Mar 2002 04:47:19


> We know that the engine is pretty much is a polynome

A what?
Tony Rickar

GPL physics question.

by Tony Rickar » Sun, 31 Mar 2002 05:04:19

A polynomial is a mathematical expression consisting of a sum of terms, each
term including a variable or variables raised to a power and multiplied by a
coefficient.  The simplest polynomials have one variable.  A one-variable
(univariate) polynomial of degree n has the following form:

See - piece of cake!

Jan Verschuere

GPL physics question.

by Jan Verschuere » Sun, 31 Mar 2002 09:43:38

The implied assumption I find GPL too hard is false as well.

In some ways it is (one doesn't have to suffer the cornering loads, for
example), in some ways it isn't. My point about the cars and the physics
implementation in GPL is that on the one hand they are too forgiving and
vague, while on the other quite prone to unpredictable reactions to upsets.
I have limited real life racing experience, but I'm pretty sure any racing
driver would park the "GPL car" after one lap and walk away.

As simulated racing cars go, however, GPL's implementation was the best of
it's day and still is quite formidable today.

Not easy, no... it shouldn't be, but intuitive. While miles better than any
other sim at it's release it does take far too long for the experienced
simmer to get somewhat up to speed in GPL. It shouldn't take more than a few
familiarization laps to get within sight of one's potential. It might take
ages to actually get there.

I've had fairly little time in NR2002, for example, but I already find I can
pretty much pre-empt the car in most situations. While I still find it hard
to judge how much braking is required to hold an inside line in GPL, it
comes quasi natural in NR2002. I judged where to brake, modulation and
transition into the corner adequately the first time I went for an inside
move into T1 at the Glen just now (the simulated Terry Labonte had no
answer, anyway ;-))

It doesn't have to be exactly the same feeling, as long as it's consistent
with the attitude of the car and with one's perception of what is possible
in reality. The brain will translate automatically.

I ascertain NR2002 does a better job on current hardware (as it should,
being nearly 3.5 years more recent).

It's too easy to take advantage of quirks in the model and the fact nothing
"wears" but fuel, true. It's far easier to be consistent in other sims,
however. I used to attribute that to less simulation sofistication in those
games, which is partly the case, but I've come to realise there's more to
it. Trust me on this, I've been simming for 9 years and have been forced to
re-evaluate my position a number of times over this period. I don't say this
lightly.

Don't get me wrong though... having said all this, I still love GPL to bits.
It's just have had certain misgivings, neigh nitpicks, since practically day
1.

Jan.
=---

na_bike

GPL physics question.

by na_bike » Sun, 31 Mar 2002 15:58:43

On Fri, 29 Mar 2002 09:30:44 -0800, The Carvalho Family


>My $.02, I hear reports that some professional drivers (including Juan Motoya and
>J. Villenueve) use GPL and claim it's the most realistic they've seen. It's
>obviously got something going for it, or there wouldn't be so many of us wasting
>our time talking about it instead of racing! ;-)

Well, ->most<- realistic is by definition a relative term. GPL
certainly was groundbreaking in terms of using Euler equations and
inverse kinematics at high frequency. That gives it a pretty darn
realistic impression of weightshifts and inertia which of course is a
big deal whilst racing. And since at the time GPL was the only sim
that did that it was _the_ most realistic sim and a huge leap forward.

But one cannot expect what JPM and JV say to carry over to every point
or detail. If you could go over the details with any of them I'm sure
we would unearth some very peculiar artifacts. Actually I think it
would be pretty interesting if one could do that with an experienced
driver. Like if they think powersliding through Vialone(a.k.a. Ascari
today) or Gresil(righthander at Rouen) is realistic. Which you must do
since tightening the diff makes the car oversteer in flat-out turns in
GPL. Just to prove my point, JPM also swears by GP2! But no one would
ever try to convince somebody else that GP2 is an actual realistic F1
sim by any standards. It gave a decent impression of driving an F1,
and it was certainly the ->most<- realistic sim at the time(well,
maybe not counting ICR2). But like everyone know you dont even have to
peel a single thing off of it for it to show some blatant errors. I
think to a lesser degree the same applies to GPL.

na_bike

GPL physics question.

by na_bike » Sun, 31 Mar 2002 16:01:19




>> We know that the engine is pretty much is a polynome

>A what?

A simple mathematical equation.
Steve Blankenshi

GPL physics question.

by Steve Blankenshi » Sun, 31 Mar 2002 23:06:34

Don't know about JV, but IIRC, the JPM quote came before the current crop of F1 sims (GP3, F1RC & F1-2001) was released.  So he
really didn't have a lot to choose from at the time, did he?  I remember an interview with his dad, who said he used GP2 to learn
the tracks for F3000, as well as for determining gear ratios.  Really praised GP2's track modeling.

Who knows what he'd say now?  Well, there is that Williams F1 game... ;-)

SB


> Well, ->most<- realistic is by definition a relative term. GPL
> certainly was groundbreaking in terms of using Euler equations and
> inverse kinematics at high frequency. That gives it a pretty darn
> realistic impression of weightshifts and inertia which of course is a
> big deal whilst racing. And since at the time GPL was the only sim
> that did that it was _the_ most realistic sim and a huge leap forward.

> But one cannot expect what JPM and JV say to carry over to every point
> or detail. If you could go over the details with any of them I'm sure
> we would unearth some very peculiar artifacts. Actually I think it
> would be pretty interesting if one could do that with an experienced
> driver. Like if they think powersliding through Vialone(a.k.a. Ascari
> today) or Gresil(righthander at Rouen) is realistic. Which you must do
> since tightening the diff makes the car oversteer in flat-out turns in
> GPL. Just to prove my point, JPM also swears by GP2! But no one would
> ever try to convince somebody else that GP2 is an actual realistic F1
> sim by any standards. It gave a decent impression of driving an F1,
> and it was certainly the ->most<- realistic sim at the time(well,
> maybe not counting ICR2). But like everyone know you dont even have to
> peel a single thing off of it for it to show some blatant errors. I
> think to a lesser degree the same applies to GPL.

Steve Blankenshi

GPL physics question.

by Steve Blankenshi » Sun, 31 Mar 2002 23:16:49

My first impression was that it did feel like a real car, but I had the strongest urge to immediately return to the pits and punch
my engineer for sending me out in such a deathtrap.  "Are you trying to KILL me, or what?  FIX this damned thing!"

;-)

SB

<snip>
<snip>

H..

GPL physics question.

by H.. » Mon, 01 Apr 2002 10:18:34

OK, so polynome = piece of cake.

Got it, thanks.

(curiously, the online merriam-Webster Collegiate Dictionary
did not recognize the word).

Jim Seamu

GPL physics question.

by Jim Seamu » Mon, 01 Apr 2002 10:29:26


I think he meant "the rpm vs torque curve is pretty much polynomial", i.e.
of the form

   torque = a + bx + cx^2 +dx^3......

where a,b,c,d... are constants and x is rpm.

Jan Verschuere

GPL physics question.

by Jan Verschuere » Mon, 01 Apr 2002 10:35:24


> >> <snip>
> OK, so polynome = piece of cake.

> Got it, thanks.

> (curiously, the online merriam-Webster Collegiate
> Dictionary did not recognize the word).

That's because the actual term and spelling is polynomial.

Jan./who, from other posters, got the impression the engines were done via
look-up tables...

Dave Pollatse

GPL physics question.

by Dave Pollatse » Tue, 02 Apr 2002 06:49:04

Radial tires have less camber stiffness than bias-ply tires, which is a
behavior in the linear (low-slip) range which is similar to toe-in.  Radial
tires do, however, develop *peak* lateral friction at fairly high camber
angles (nonlinear range).
-Dave P.






> > I dare you to name a few driving games (beside N4 and N2002) with a
> > better PHYSICS model as GPL.

> F1 2001.  No, I'm not kidding.  It's true some of the track modelling is
> questionable, and like all sims there are people claiming that the grip
> or the aero or something is wrong, but I think the underlying physics
> code is as solid and comprehensive as you will find anywhere.

> > I fully disagree, beside N4 & N2002, GPL has the most sophisticated
> > physics model of todays racing games. There is no flaw, only parts
> > which aren't modelled.

> Sophisticated?  I don't think so.  While I can't prove this, I suspect
> very strongly that GPL uses the "simplified suspension model" with
> virtual links, roll centers, equivalent wheel rates, etc.  I think that
> is why we see "wheel rates" rather than spring rates in the setup
> screen.  I think this is why the cars feel so wallowy and imprecise
> compared to other sims.  I also think doing a full kinematic suspension
> model at 288 hz on the average P200 that was around when GPL was
> released would have been impossible.  More modern sims have taken
> advantage of more modern hardware and are doing more thorough and
> comprehensive physics modelling.  F1 2001 for example appears to have a
> full kinematics model of the suspension.  GPL was excellent for the time
> it was released, but that was three years ago.  Computing power has more
> than quadrupled since then.

> > (I must add that the tyre model in GPL is a bit suspect. It might be
> > the data they used for the tyres, or the tyremodel is too basic, I
> > don't know)

> Agreed.  It seems to be counterproductive in GPL to run more than about
> 1 degree of negative camber.  I don't think that's realistic, but I
> could be wrong.  2 to 3 degrees of negative camber is common with modern
> race cars on radial tires, and my understanding is that radials are less
>  responsive to camber change than bias ply tires.

Tony Rickar

GPL physics question.

by Tony Rickar » Tue, 02 Apr 2002 07:04:03


> Radial tires have less camber stiffness than bias-ply tires, which is a
> behavior in the linear (low-slip) range which is similar to toe-in.  Radial
> tires do, however, develop *peak* lateral friction at fairly high camber
> angles (nonlinear range).

Blimey, Days of Thunder is on in the UK in a minute. I am beginning to relate
more & more with Cole Trickle than with some of the posts on this group these
days.

I am just a driver...

Sadly for my wife she doesn't see the likeness with Tom :-)

Tony

Haqsa

GPL physics question.

by Haqsa » Tue, 02 Apr 2002 08:27:27

Ah (light goes on above my head).  So at a small camber angle, like -1
degree or so, I will get more camber thrust from a bias ply tire, but as
I increase camber angle the bias ply tire will reach its limit and taper
off before the radial does, and the radial eventually exhibits more peak
camber thrust at its optimum angle?  Makes sense now.  Thanks for the
clarification, Dave.


Mats Lofkvis

GPL physics question.

by Mats Lofkvis » Tue, 02 Apr 2002 22:27:49


> "Mats Lofkvist" wrote...
> > <snip>
> > The implied assumption that GPL is to hard is totally
> > wrong imho.

> The implied assumption I find GPL too hard is false as well.

Ok, I assumed wrong there.

The main point of my longish rant was that with high end hardware,
GPL _is_ as easy as one can expect imho. Maybe the major problem with
GPL is that it behaves so different (and is so much harder) with
suboptimal hardware (e.g. the best money could buy in 1998 :-).

      _
Mats Lofkvist

Tony Rickar

GPL physics question.

by Tony Rickar » Tue, 02 Apr 2002 23:17:09


> The main point of my longish rant was that with high end hardware,
> GPL _is_ as easy as one can expect imho. Maybe the major problem with
> GPL is that it behaves so different (and is so much harder) with
> suboptimal hardware (e.g. the best money could buy in 1998 :-).

I always thought GPL scaled very nicely, form its early days in 98 with 640x480
and a rendition card to today's faster processors and graphics cards. Ok so
getting 36 fps looked pretty awful compared to today's PCs but it was quite
playable.

To me that is lot of the charm with GPL, as the third party community has added
and added to it to scale with the hardware. With all the addon bits it looks as
good as any other IMHO.

As for the easy or hard bit it meets my expectations pretty reasonably. I can
run around consistently at high 1:05s at the Glen for instance and the replays
look about right, I am not at the extreme angles discussed elsewhere.

Of course I am probably 2 secs off the really quick guys. It depends what you
compare with - the quick GPL drivers or the footage and lap times of F1 drivers
in 67. If I take the latter then it feels right. If I try to catch up the
quicker guys then yep GPL is too hard!

It has been discussed many times before, but if the real drivers could have slid
through fast bends like some GPLers without fear of the consequences then they
would have gone quicker.

Learning sim racing tends to be the opposite of real racing. Start fast and
learn where you have to slow down. Of course in GPL's case the slowing down
takes that bit longer than modern F1 sims, which tends to make it frustrating at
first.

Tony


rec.autos.simulators is a usenet newsgroup formed in December, 1993. As this group was always unmoderated there may be some spam or off topic articles included. Some links do point back to racesimcentral.net as we could not validate the original address. Please report any pages that you believe warrant deletion from this archive (include the link in your email). RaceSimCentral.net is in no way responsible and does not endorse any of the content herein.