"Damien Smith" <smithwom...@hotmail.com> wrote in message
news:alv1er$1a226$1@ID-110394.news.dfncis.de...
> Mark,
> The reason that most people in this group have a dislike for GP4 mainly
> comes down to physics. Being a simulation group, obviously means that
> physics is the number one priority for most of us. Unfortunately, Geoff
> Crammond is quite ignorant of car behaviour (he's not the only one.)
Sure,
> there's some nice things in GP4 like the beautiful track scenary and the
wet
> weather effects.
I can't agree with the first sentence. Most people in this group dislike GP4
because it's the proper r.a.s. thing to do. I agree that the physics could
be subjective to some extent, but as far as what's available for PC, GP4 is
as accurate as any other. IMO, more so, and to small degree.
>But the physics are primitive. Want an example? Try
> turning all aids off, then, while cornering, try violently lifting your
foot
> off the throttle. Heck, you can even violently pump the throttle with no
> (realistic) consequences. Try doing that in F1 2002 or better still - a
> real (powerful) car.
You know what? That's a good idea. So I did it. I tried it in GP4, the four
games you mention, my 1987 Isuzu Trooper, and my Gold Firefox with a
balanced Suzuki RM125. (We can go through in depth analysis where I may lay
out detailed descriptions of the actual stages of the tire model in both the
sim and real vehicles but for the purpose of this post you'll have to trust
that I have taken all into account before making comparisons.)
Bearing in mind that I have meticulously set up the driver interfaces in all
of these games and the racing kart...
In the computer games I found next to no difference. Which surprised me. The
experiment raised my opinions of N2K2 and F12002. They all displayed
excellent grip below the limit, in which neither flooring or lifting had
much effect. I would expect this to some degree, depending on hp/weight. At
the limit they all displayed the same general characteristics but with
different personalities. Something I would also expect. I found nothing out
of character if we think of the reactions expected in the most broad terms.
None of them had the resolution of sensation that the real vehicles had, and
one of them had the depth of slip angle represented on a much higher level
than the rest, that was GP4. LFS should not be in this discussion, so
anything I say should not be attributed to it. It would take a whole other
conversation to pick out the inaccuracies in LFS's physics model.
Oddly enough, the Trooper and kart acted very much like the sims when below
the limit. The feel in the tires was almost mute with the major difference
being that I had a direct sense of how much strain was being put on the
tires, no matter how low, in any given maneuver in both of the real
vehicles. (Stay tuned to this thought as it is the very best thing GP4 has
to offer.) Once I approached the limit the cars became very much different
and not altogether unaligned with the sim counterparts. Acclerating,
decelerating, and laccelerating were all aided by very tangable personality
traits which could be desribed as elasticity, grittyness, squirmyness,
scuffing, and so on. Overall, the real cars have a much higher resolution.
Unfortunately, I found the GPL cars handle much more like the trooper than
the kart. I won't believe a 67 F1 car was that bad. On the flip side, I
found GPL to be stunningly realistic physically. (As good as GP4, better
than all the rest.) Which leads me to my next point:
If these sims are all close in their representation of reality then what
makes it a realistic experience?
Simple. Drive model or user interface.
I know you don't want to hear that term. But it's the key. Imagine driving
GPL from a spreadsheet. You would enter braking pressure and steering angle
and the program would come up with a response. How hideous would that be?
Well that's what we are doing. Fortunately, the programmers have also given
us a real time interface.
So directness of input and feedback are key in this real time spreadsheet,
yeah?
> I found it strange that you mentioned the tyre model in GP4. This is yet
> another category where F1 2002 simply walks all over it.....???
I'm not sure what the problem is with F12002. But the tire model is far to
slow for modern racing tires, the depth is good but lacks authenticity
compared to GP4 in the way the tire breaks loose and how the chassis
responds to that change, and the 'feel' is nonexistent. I suspect you are
using FF and have not set up GP4 well.
Let's take understeer, for example. In F12002 it comes on as a function of
speed, regardless of steering angle. It is also too slow, too late, and too
forgiving. The forgiving part was by design, I'm convinced, because a lot of
the drive model in F12002 is delayed to accommodate both the ping to FF
wheels and the average driver. Real understeer can be very progressive with
a street tire at street speeds, but with racing tires and a line that's
already at the limit... well... it can be a real surprise. One thing that is
good about F12002 is the fact that it's consistent. So, while certain
physical portrayals may not be accurate to real life, they are learnable.
Once again the 'Translation Factor' is introduced unnecesarily.
For me though, F12002 has too many flaws in the physics/drive model (which
are interconnected and inseparable) that keep me from embracing it as a sim.
Such as: No obstacle under two inches in height has any effect on the car.
Surface differences do not influence grip. (Both of these bug me because I
want my balls to shrivel when I put a wheel off at 170.) An F1 car that can
be floored in 1st and 2nd gear? C'mon...! There is a huge latency of no less
than 200ms between an action happening in the physics engine and when it is
displayed for the user. (They did do a good job of allowing corrective
measures to be retroactive.) For the non-FF user, who feels more detail
than the FF user does by the nature of the interface, there is no indication
of grip whatsoever until the limit has been passed. (I understand this is
probably the source of latency in 'feel'.) The perspective is way off and
seeing things in the distance 'proper' is very hard, leading to unnatural
mistake. Upsets caused by variations in surface texture are nonsensical and
therefore uncorrectable. Setup changes do nothing. The controller setup does
not allow for slope adjustments but rather, has deadzones? (Deadzones have
no place in simulation control systems.) Ultimately, even if the latency was
a non-issue, the translation factor is too high for me to consider F12002 a
sim.
> FWIW - The following have a reputation for good physics :- Nascar 2002,
> GPL, Live For Speed, F1 2002
These reputations are part of the problem. Where GPL earned it's, I have a
distinct suspicion that F12002 is preferred over GP4 because of the name
"Crammond" and prejudices therein.
I was considering going into detail with the flaws in each of the titles
you've mentioned but they are sims of different racing classes and my
comparisons would garner much criticism and the bandwidth is better wasted
on what I find good about GP4 that I don't find in the other sims. (LFS
excluded.) Let me summarize by saying that I find the papy games, even after
the flaws are removed, to be very much chassis driven representations. That
is to say, the information about the tires is derived from the chassis
rather than the tires themselves.
First things first, I think the personality of the tire is on a level that
no other PC game has brought to us before. I can 'feel' the grip
accumulating, which is new, up to the point of optimum slip angle, can feel
it at the limit, have a hard time judging when it's slightly over but can
tell when I'm going to over cook it. The chassis dynamics, physics model,
drive model, and A/V presentation are such that I can drive with
anticipation in the same mental and physical state that I drive a real car
in at the limit. Also a first in PC games. Sure, I can drive GPL with
anticipation but only through translating what it is doing into what
responses will rectify the problems. That is a personal latency that I would
rather not have in my games. (Read - very low translation factor.)
I find the perspective to be head and shoulders above anything that has come
before. I find that I can look more freely where I would in real life. The
distractions that seem to be inherent in every other title I've played are
simply not in this one. This leads again to a more fluid and natural
approach to the mental and physical aspects of driving. (Read - more
realistic.) Peripheral 'senses' are at an all time high in PC simming, I
just 'know' where the wheels are and never give it a second thought. (Or
more dangerously, a second look.)
After that the tracks are great, the chassis responds well to even the most
minor setup change, the graphics and sound are good enough (Which is my
highest rating for graphics and sound.), the AI are fantastic, I can feel
when one or two wheels go light, and I have to state again that the grip is
stunning. Skipping sideways is a real joy.
The only way to 'feel' these thing in any sim is trough a transparent
interface. So that means things like slope control, intelligently applied
controller setup, an open mind, low latency in the game engine itself, and
most unfortunately, no force feedback. The new games have too high a
resolution in responsiveness to allow an electric motor to represent them
and FF by it's nature masks A/V feedback.
> Feel free to ask questions about GP4. I *have* spent some time with it.
> (It's just that in that time I found a few too many shortcomings)
> Cheers
> Damien
Sure, here's one.
What do you look for in the performance traces help you to decide which
differential settings to change? Right now I'm looking at shaft speed
changes as I would ramp angles in GPL, and input torque like I would
...
read more »
