rec.autos.simulators

I've been reading the posts here about rain in GPL and how it would be
too hard and take up too much processor.

Well I'm always up for a good argument so lets have a go... :o)

Firstly, I'm the one who put rain in N2 with the TPTCC patch when
we did version one... all I dis was alter the pallette to make
everything look gloomy and we altered the traction to represent the
loss of grip.

We had loads of e-mails saying "Ooh, its really good - its really
feels like I'm drving in the wet!" - so I must of got something right.

A couple of points come to mind....

1) In most cases, using simple visual effects you can create the
illusion of rain quite well and the psychology of your brain will do
the rest.

2) As a racer myself (and I loved racing in the rain) for me, driving
in the wet was not much different from the dry except for a) loss of
visibility, b) less grip c) uneven traction and d) colder tyres.

O.K, many will argue with me on those points (if your a purist) but
for me, thats all it was...

So, sensibly, I dont see why rain couldn't be put into a sim.

First of all, look at Microprose Grand Prix version 1 with its rain...

How would I put rain in GPL....

1) LACK OF VISIBILITY  - We all know that when it rains you often
get bad visibility. You can see, but not as far in front. Most 3D
cards already have fogging routines built into them so it would be
easy to use these to create a fog effect.

When racing in rain, as long as no-one is in front of you things are
clear (unless there is a mist) and things get shitty when someone
pulls in front of you. So, based on the proximity of cars ahead of
you, the fogging distance and density could be adjusted (like GP1)
to represent that.

Secondly, we already have smoke from the engine so how hard would
it be to create two sources of 'smoke' from where the tires touch the
ground to represent the water being thrown up.

2) LACK OF GRIP - Yes, it would be hard to model everything about
rain but in general, the main thing is the lack of grip. A 'ratio'
could be applied to the whole track to reduce it, but for area's which
are bumpy or waterlogged that could be programmed into the tracks DAT
file. Therefore, localised problems such as aquaplaning or puddles
could be represented.

3) TEMPERATURE - The cooling effect of water on the tyres can
be easily reproduced through a basic formula and as the 'racing line'
is already overlayed on the track within the DAT, could it not be
programmed that over time with cars passing over it, this line drys
out? This would then make the tyres hotter - moving off-line onto the
wet track cools it. We know the program knows when your on-line
because of the autobraking and steering help.

Apart from these points, I dont think you'd need to do rain drops as
in Flight Unlimited 2 because for a start, you'd be wearing goggle and
drops dont tend to form - and with them that close to your face, your
focusing beyond them and dont tend to see them.

I appreciate that GPL is processor intensive and that my suggestions
would still need more processing power, but I dont think we would all
need Cray's to see rain in a sim.

Well thats enough of my ramblings...

- Jed, The Pits

You forgot one thing:

The AI.

Whatever success was had with the AI in GPL would be severely compromised
by  having rain affect their simplified physics model.

When combined with the player physics and graphics issues, which are
non-trivial despite your arguments, I believe it's just too much work to
even be considered for a patch.

Sorry!!!

As Kaemmer says in his introduction to the GPL manual, the missing thing
from the sim is the feel a driver gets through the seat of his pants, and
his other senses. This makes it very hard to 'feel' the lateral forces
building as you enter a corner. In wet conditions these lateral forces would
be even lower before grip was lost - you'd feel like you were crawling round
the circuit.

I'd still like to try it though:-)

- Michael

[snip]

Thank-you, but the sky around Mosport is gloomy enough as it is... :)

As a Canadian I feel like it is about to snow every time I look up at
the sky heading into 5a and 5b... jeez.  You think *rain* is slippery -
hah!   8b

Cheers,
Gian.

Joe

<snip>

I would like to see rain modeled in a sim, but only when it's done
right. Your suggestion is, IMHO, yet another track. (visuals excluded)
The problem with rain is that the conditions are changing all the
time. I think that the calculations done in a dynamic environment are
a tad too much for the PC's available today.

Remco

[A lot of interesting stuff about rain in GPL I have no expertise to
evaluate, but which sounded reasonable to me nonetheless.]

One correction, though:

There's no autobraking in GPL, and no steering help. The only aids
available are some sort of traction control and an ABS system. One
can't assume the program knows about your being on the racing groove,
therefore.

--
Wolfgang Preiss   \ E-mail copies of replies to this posting are welcome.

Joe Conklin
For Central PA Online Gamers:
http://www.geocities.com/siliconvalley/network/5312/
For Grand Prix Legends:
http://members.xoom.com/gplegends/

remove spam to reply

I appreciate that you cant get the feedback but theres always two
sides the argument.

Mine was that you could produce a psuedo-rain effect maybe using the
tricks I described and I know it wasn't perfect - just ideas.

Yes, there are a lot of physics involved, but looking at the range of
parameters available for the AI cars, it may be possible to produce
'wet weather' driving styles for them.

What I was essentially trying to say wasn't that Papyrus/Sierra should
put this in the patch but that, in ANY sim, you could at least produce
the visual effect of rain and with slightly modded physics.

It wouldn't be perfect but hey, until we all have 500Ghz processors
that can calculate it all, we can't expect everything.

- Jed, The Pits

One more issue - time.

Cheers!
John

I agree with your points - my original post was not meant to say "Yes,
it can be done and this is how", mainly to raise some 'possibilities'
of how it could be represented.

I agree, it would be WONDERFUL to race and then find the track getting
wetter as rain builds up but we know thats hard.

To some people, the way we made the 'wet' races in the TPTCC was
enough - their imagination and the alterations did make the same
tracks definatley differnt to drive but I agree that many would
rather see it done properly or not at all.

It sensible discussion like this that I would hope goes on at the
software houses and not "How about rain?", "No", "Why?", "Because we
dont have time/no budget/cant be arsed" sort of conversations...

Its horses for courses really but as a parting shot - although some of
us consider ourselves 100% pure simmers who want it 100% accurate or
not at all consider this:

In racing sims we only really have visual and audio senses to rely on
and, at present, rudimentary force feedback. If you can make those few
senses you use driving a racing sim believe its raining then you've
done your best - you've simulated it.

Without all the input you get from driving for real, I think all but
the hardened would be convinced.

- Jed, The Pits

That was a valid point and the point of my main post was to start a
discussion on how it could be done or what 'techniques' could be used
by programmers to simulate rain. I appreciate with GPL's current
engine it would kill the frame rate, but who knows in the future what
could be possible.

BTW, it is well known that not *every* game on the market is as sleek
and fast as it could be in terms of programming. I used to program
software for ZX Spectrums in the 80's and with 48K of memory, you had
to optimise your code.

I think comercial pressures to get stuff done and out on time means
that coding sometimes is less than optimal and I've seen some
wonderful things compiled into EXEs!

I cant remember exactly, but I think on ICR2, the Michigan track had
the entire "Hi I'm yadda, yadda and this is ICR2" wave compiled into
the track DAT. I dont know how Papyrus's engine handled it, but I I
dread to think how much memory could of been lost if the engine loads
all the files from the DAT into memory regardless of whether they are
used or not.

I've had it too, but certainly in the open***pit cars I've raced its
never a problem - as I said, your focal point tends to make you not
notice them and it would only be more processor ticks that could be
used on other stuff.

- Jed, The Pits

Good for you. :) There are still some things that are better than GPL.
Not many, but they do exist.

Well, the canonical answer to the "why no wet weather in GPL" question
is: "It couldn't be done properly with today's equipment, so Papy
decided not to do it at all." Some people will be - err - 'irritated'
if you dare to question this answer. ;)

You convinced me that a good (though not perfect) approximation could
be done. It seems to me there are several possible reasons why Papy
decided not to go this way. All of those reasons probably are a part
of the 'real' answer why we don't have rain in GPL:

1) It would have taken additional money and time. (And money equals
time in business anyway.)

2) Weather is something you can leave out pretty easily. You can avoid
dealing with a whole complex of difficulties if you treat weather as a
constant instead of a set of interacting variables. There is no other
aspect of simulating driving that is as little hassle to leave out,
and as much of a potential problem to include as weather. If you have
to decide what aspect to leave out, weather is the logical choice.

3) Not too many customers will actually miss rain. If rain is included
in a sim, I at least try what it looks and feels like, then go back to
dry weather and stay there. :)

4) The kind of weather you suggest would contradict the development
"philosophy" of GPL: it would be make-believe rain. While I agree that
it would work [I have been involved in theater productions for some
time and know that f.i. a light effect produces often much better
"water" than real water on stage], it would constitute a completely
different approach than the rest of the GPL engine. The whole physics
engine of GPL works like this: "Let's simulate the parts and see how
they interact. If we simulate them correctly, the effect of the
interaction will be correct too." AFAI see it, there is little
compromise in this model (small bumps in the road and dynamic camber
are the only instances of compromise, AFAIK.) Papy chose to start with
the physics to achieve the effects in all cases, and not vice versa. A
true GPL approach to weather would be something like this: We have a
certain temperature to start with. Now a cloud enters the simulated
world and releases x units of water/square inch. (The intensity varies
with the distance from the center of the cloud, of course.) The water
units are affected by the wind and gravity and fall on the track and
the adjacent landscape. The track cools off with every unit of water
that hits it; water units follow the surface and still obey to
gravity... etc.etc. This kind of rain sim alone would probably be too
much for today's machinery, even without cars... Saying "we know what
rain looks like and what it does to racecars" would work, but it would
be a completely different approach.

Now back to the Ring.

--
Wolfgang Preiss   \ E-mail copies of replies to this posting are welcome.