www.west-racing.com
New renders etc...
I don't think we will see anything actually 'move' until end 2005, I fear...
Notable threadWest Racing's engine renders and the realism of detail
A new West Racing update showing a detailed engine and garage for Racing Legends left most readers puzzled about the priorities of a project with no drivable build. Some suspected the renders served as a portfolio for other work, while Jason Moyer defended the Wests' ambition. The thread became a technical discussion of how much physics detail is useful, with Performance Simulations developer Todd Wasson and Racer developer Ruud van Gaal discussing CPU budgets, simulation step sizes and collision methods.
With J. Todd Wasson (Performance Simulations); Ruud van Gaal (Racer)
Automated summary.
In article <k5Y7b.13968$[email hidden]>,
[email hidden] (Trick Dacy) wrote:
> New renders etc...
I must admit while I remain enthusiastic (in a "it'll give me something to
look forward to when I'm retired" kind of way) about RL, I find their
update puzzling. Is anyone really interested in an interactive garage and
the process of getting the car to the track at this stage? Just gimme a
car on a track guys! It's not as if we'll have a garage of cars to choose
from for at least a decade at this rate.
Still, the pix are very pretty and you've got to admire the
vision and potential even if the reality's several nibbles short of a full
byte of code so far.
Andrew McP
Personally.. I think they are using this whole thing as a way to "show their
stuff" hoping that someone will come to them for other opportunities.
There really can not be too much hope left that it will actually ever amount
to a real game
"Trick Dacy" <[email hidden]> wrote in message
news:k5Y7b.13968$[email hidden]...
> www.west-racing.com
>
> New renders etc...
>
> I don't think we will see anything actually 'move' until end 2005, Ifear...
>
>
>
"Andrew MacPherson" <[email hidden]> wrote in message
news:[email hidden]...
> In article <k5Y7b.13968$[email hidden]>,
> [email hidden] (Trick Dacy) wrote:
>
> > New renders etc...
>
If all those engine parts are actually gonna be functional then i think its
worth the wait :)
In article <7C_7b.33741$[email hidden]>,
[email hidden] (Cliff Roman) wrote:
> hoping that someone will come to them for other opportunities
That's a good point. Their website is a good advert for their design
skills if nothing else.
Andrew McP
I agree fully, Andrew...
But the most important word you've used is 'IF'... ;-)
"frederickson" <[email hidden]> wrote in message
news:DI_7b.1472$[email hidden]...
its
> worth the wait :)
>
>
Quoted text (10 lines)
> > "Andrew MacPherson" <[email hidden]> wrote in message > news:[email hidden]... > > In article <k5Y7b.13968$[email hidden]>, > > [email hidden] (Trick Dacy) wrote: > > > > > New renders etc... > > > > If all those engine parts are actually gonna be functional then i think
"frederickson" <[email hidden]> wrote in message
news:DI_7b.1472$[email hidden]...
its
> worth the wait :)
That's why it's taking so long... they're waiting for the hardware to catch
up to their minimum requirements... P4 4Ghz with 1Gb Radeon11500 :\
KK
Quoted text (10 lines)
> > "Andrew MacPherson" <[email hidden]> wrote in message > news:[email hidden]... > > In article <k5Y7b.13968$[email hidden]>, > > [email hidden] (Trick Dacy) wrote: > > > > > New renders etc... > > > > If all those engine parts are actually gonna be functional then i thinkAndrew MacPherson wrote:
From the website:
I think it also interesting for people to be able
to learn how these cars were put together and how
they worked. I also believe that it adds to the
realism of the sim knowing that you driving more
than just a 3 dimensional shape with pretty
textures and might prompt drivers to take more
care of the machines.
Can we vote on that?
I'm with you Andrew. I really admire the vision and
dedication these folks have. Isn't it time for a beta?
We'd pay for it, and they wouldn't need to sell their
house.
I'm going to be disappointed if the distributor
doesn't have the correct dwell, or if the gap in the
spark plugs isn't correct.
I'm starting to wonder if my 600 MHz P3, with Voodoo
5500, is going to be able to get 30 FPS at 1600x1200x32
with all the graphics options maxed.
I'd like to know the frame rates on today's top of the
line systems, like a 4 processor 2.8GHz Xeon unit.
West brothers and associates, don't let us get you
down. We're just having a little fun. You are already
a legend at SCI. Good luck.
Larry
Quoted text (18 lines)
> > In article <k5Y7b.13968$[email hidden]>, > [email hidden] (Trick Dacy) wrote: > > > New renders etc... > > I must admit while I remain enthusiastic (in a "it'll give me something to > look forward to when I'm retired" kind of way) about RL, I find their > update puzzling. Is anyone really interested in an interactive garage and > the process of getting the car to the track at this stage? Just gimme a > car on a track guys! It's not as if we'll have a garage of cars to choose > from for at least a decade at this rate. > > Still, the pix are very pretty and you've got to admire the > vision and potential even if the reality's several nibbles short of a full > byte of code so far. > > Andrew McPI give them credit.. but I do wish they'd talk about the actual GAME more
often. Mostly they talk about flashy stuff, like the whole garage diorama
thing. THe last update did have a mention of calculating chasis flex, which
blew my mind.. I'd love to see all that implemented.
Mike
http://mikebeauchamp.com
"Trick Dacy" <[email hidden]> wrote in message
news:k5Y7b.13968$[email hidden]...
> www.west-racing.com
>
> New renders etc...
>
> I don't think we will see anything actually 'move' until end 2005, Ifear...
>
>
>
LOL
--
Tim White
"Trick Dacy" <[email hidden]> wrote in message
news:k5Y7b.13968$[email hidden]...
> I don't think we will see anything actually 'move' until end 2005, I fear...
>
"Andrew MacPherson" wrote:
> I must admit while I remain enthusiastic (in a "it'll give me somethingto
> look forward to when I'm retired" kind of way) about RL, I find their
> update puzzling. Is anyone really interested in an interactive garage and
> the process of getting the car to the track at this stage? Just gimme a
> car on a track guys! It's not as if we'll have a garage of cars to choose
> from for at least a decade at this rate.
I am puzzled too! Some interesting ideas to incorporate in further releases
once we are hooked on the physics etc. but the time taken to create a fully
working visual engine seems an unusual priority. I thought shift-r was
enough of a chore after wrecking an engine! If I have to spend nights
mending the darn thing my family are going to despair...
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.515 / Virus Database: 313 - Release Date: 01/09/2003
Tony Rickard wrote:
>
> I am puzzled too! Some interesting ideas to incorporate in further releases
> once we are hooked on the physics etc. but the time taken to create a fully
> working visual engine seems an unusual priority. I thought shift-r was
> enough of a chore after wrecking an engine! If I have to spend nights
> mending the darn thing my family are going to despair...
I hear adjusting the valves clearances and getting the cam timings set
correctly on a 48 valve, 4 cam, V12 can be quite entertaining....
Beers and cheers
(uncle) Goy
"goyl at nettx dot no"
http://www.theuspits.com
"A man is only as old as the woman he feels........"
--Groucho Marx--
> www.west-racing.com
>
> New renders etc...
>
> I don't think we will see anything actually 'move' until end 2005, Ifear...
That's if there's even anything capable of moving it at that time. I
imagine we'll be using 4.5 - 5 GHz PCs at that time. I don't think a
machine of that speed could even fully do F1C justice.
"Goy Larsen" wrote...
> <snip>
> I hear adjusting the valves clearances and getting the
> cam timings set correctly on a 48 valve, 4 cam, V12 can
> be quite entertaining....
Trust me, a friend of mine's older brother deals in 2nd hand luxury cars. He
had a Lambo Countach on the lot at one stage and I can tell from experience
having helped it's an f'ing pain in the behind. If he drove it over a bad
road, that would mess it up and it would start chunking. Then if he happened
to stall it, it wouldn't start again. Oh the fun times he had with that car.
The early Countachs are definately for parade purposes only. Unless you got
nothing better to do than take half the engine apart each time you drive it.
Jan.
=---
In article <k5Y7b.13968$[email hidden]>,
[email hidden] says...
> www.west-racing.com
>
> New renders etc...
>
> I don't think we will see anything actually 'move' until end 2005, I fear...
By far the best car parked in a garage simulator I have ever seen!
Cheers,
Rod.
In article <Hu58b.7703$[email hidden]>,
[email hidden] (Tony Rickard) wrote:
> If I have to spend nights
> mending the darn thing my family are going to despair..
Stop it, you'll put ideas in their head. Of course that may be how they
plan to finance RL longterm... you have to pay someone to come and fix the
car after every crash!
Damn, it'll be cheaper for me to buy a real race car :-)
Andrew McP
PS Actually I do like the idea that RL could keep a rough record of how
bad your crashes are, giving you a guestimate of how much it might have
cost to fix. We could have a RLrank-style "rank per unit repair cost" to
see who's the cheapest driver to run :-)
Do you have any clue what kind of puter you need to generate that kind of
detail and image at a frame rate of 10 per seconds, lets say?.
Even if it is ready to drive in 2005 it won't run on any personnel pc.
Lol
"frederickson" <[email hidden]> wrote in message
news:DI_7b.1472$[email hidden]...
its
> worth the wait :)
>
>
Quoted text (10 lines)
> > "Andrew MacPherson" <[email hidden]> wrote in message > news:[email hidden]... > > In article <k5Y7b.13968$[email hidden]>, > > [email hidden] (Trick Dacy) wrote: > > > > > New renders etc... > > > > If all those engine parts are actually gonna be functional then i thinkRenders <> functional game
I want to believe in RL, I really do, but this is pointless screwing around.
Are we all part of the world's largest practical joke? Do the West's
release these renders, smoke pot then read r.a.s for a good laugh?
There is no way a good render of internal engine parts is an essential part
of a racing simulation. Model the part numerically and move on. In fact,
don't even bother... my enjoyment of RL won't be determined by accurate
replication of a connecting rod. Program in the engine parameters and get
on with the job.
Nope... that's it. As of today I fully beileve that RL is an elaborate
practical joke or at best research on a psychology thesis.
You know... this "update" reminds me of the kind of work I do at my job when
I'm not doing any real work on a project.
Boss: "Is that report done?"
Me: "No, but I've done the contour plots for the data."
Boss: "Oooo... nice graphs."
Oh man that was funnny... :)
Mike
http://mikebeauchamp.com
"RodP" <[email hidden]> wrote in message
news:MPG.19cbaa241081bbcd989e2c@news-server...
fear...
>
> By far the best car parked in a garage simulator I have ever seen!
>
> Cheers,
> Rod.
Quoted text (7 lines)
> In article <k5Y7b.13968$[email hidden]>, > [email hidden] says... > > www.west-racing.com > > > > New renders etc... > > > > I don't think we will see anything actually 'move' until end 2005, IWow.. your post got me thinking, and it does make sense.
About the fancy coloured graphs, they're useless and we know that, but
people do buy into them. A lot of people are impressed by powerpoint
presentations, even though the same data can be shown more efficiently with
a lot less work too. Same with this whole engine thing... So, the question
is.. who are they trying to impress?? I don't think it is us..
My guess is that website could be there PURELY for funding, or in an attempt
to get themselves nice jobs as 3D artists. Those renders are AMAZING, I'll
admit it. I'd hire them to model cars and engine parts if I ever needed
someone to do so.
Or, they are actually making a game. I guess if you did model all these cars
AND you wrote a physics engine that is still in testing phaze, how would you
graphically show on a website what you have done? Some nice render shots of
the cars, or a screenshot of some code from your physics engine?
The physics engine and grapic engine could definately be in development, but
there's no way of really showing it off yet. Whereas the car models can be
nicely rendered in software and shown and be very impressive.
Mike
http://mikebeauchamp.com
"Dale Green" <[email hidden]> wrote in message
news:[email hidden]...
> Renders <> functional game
>
> I want to believe in RL, I really do, but this is pointless screwingaround.
> Are we all part of the world's largest practical joke? Do the West's
> release these renders, smoke pot then read r.a.s for a good laugh?
>
> There is no way a good render of internal engine parts is an essentialpart
when
Quoted text (9 lines)
> of a racing simulation. Model the part numerically and move on. In fact, > don't even bother... my enjoyment of RL won't be determined by accurate > replication of a connecting rod. Program in the engine parameters and get > on with the job. > > Nope... that's it. As of today I fully beileve that RL is an elaborate > practical joke or at best research on a psychology thesis. > > You know... this "update" reminds me of the kind of work I do at my jobQuoted text (7 lines)
> I'm not doing any real work on a project. > > Boss: "Is that report done?" > Me: "No, but I've done the contour plots for the data." > Boss: "Oooo... nice graphs." > >
"Trick Dacy" <[email hidden]> wrote in message
news:k5Y7b.13968$[email hidden]...
> www.west-racing.com
>
> New renders etc...
>
> I don't think we will see anything actually 'move' until end 2005, Ifear...
You'd have thought the artist time would have been better spent on the
mammoth task of making detailed tracks for the game, rather than intricate
modelling of pistons and con-rods.
The engine model *is* a nice piece of work though - if it is all animated it
will be a nice teaching tool to demonstrate how internal combustion engines
work.
- Michael
On Fri, 12 Sep 2003 00:29 +0100 (BST),
[email hidden] (Andrew MacPherson) wrote:
>PS Actually I do like the idea that RL could keep a rough record of how
>bad your crashes are, giving you a guestimate of how much it might have
>cost to fix. We could have a RLrank-style "rank per unit repair cost" to
>see who's the cheapest driver to run :-)
Which will turn soon into a "The highest bill" contest, since you
don't have to pay it yourself anyway <g>
Remco
"MichaelJP"said...
> You'd have thought the artist time would have been better spent on the
> mammoth task of making detailed tracks for the game, rather than intricate
> modelling of pistons and con-rods.
Or start working on other cars for that matter.
Details *are* nice, but this is miles over the top IMHO.
I'd much rather have more cars and tracks than a few super-detailed models.
A sim is to be raced - not looked at! :-)
--
ed_
In article <[email hidden]>, [email hidden]
(Remco Moedt) wrote:
> Which will turn soon into a "The highest bill" contest
Of course :-) But after contemplating this, I think an incentive *not* to
crash would actually be rather interesting. Given enough time on the track
most people can become "fast enough" drivers. But becoming a "fast enough"
driver without breaking thousands of cars might be quite a challenge in
the same way that things like monster rank have become.
And while I'm thinking about it, I was flying around in FS2004 earlier on.
I flew one of the biplanes for the first time and noticed that in the 3d
cockpit view you could see bits of the engine moving at the front...
you know, tappety pistony valve wossits jiggling up and down (or something
techy and engine-esque like that).
It was actually quite fascinating to see that level of detail simulated
and it made me think a little harder about the Wests' ideas.
But I still want something to drive before they go modelling bubbles in
the petrol feed.
Andrew McP
"Andrew MacPherson" <[email hidden]> wrote in message
news:[email hidden]...
I'm only buying RL if they move "Chitty Chitty Bang Bang" up their list of
cars to model. Talk about your moving dookickeys and dynamic wotsits!
Stephen
p.s. I also get all glassy-eyed watching the moving bits and bobs on the
planes.
Quoted text (22 lines)
> In article <[email hidden]>, [email hidden] > (Remco Moedt) wrote: > > > Which will turn soon into a "The highest bill" contest > > Of course :-) But after contemplating this, I think an incentive *not* to > crash would actually be rather interesting. Given enough time on the track > most people can become "fast enough" drivers. But becoming a "fast enough" > driver without breaking thousands of cars might be quite a challenge in > the same way that things like monster rank have become. > > And while I'm thinking about it, I was flying around in FS2004 earlier on. > I flew one of the biplanes for the first time and noticed that in the 3d > cockpit view you could see bits of the engine moving at the front... > you know, tappety pistony valve wossits jiggling up and down (or something > techy and engine-esque like that). > > It was actually quite fascinating to see that level of detail simulated > and it made me think a little harder about the Wests' ideas. > > But I still want something to drive before they go modelling bubbles in > the petrol feed.In article <3f61d244$[email hidden]>, [email hidden] (Stephen
F.) wrote:
> p.s. I also get all glassy-eyed watching the moving bits and bobs on the
> planes
Pathetic isn't it :-) I've no idea which gene is responsible for it, but I
hope they find a cure soon.
Andrew McP
>"Dale Green" <[email hidden]> wrote
> Do the West's release these renders, smoke pot then read r.a.s for a goodlaugh?
I think the order is different, but alas true. Why can't they at least
admit they are doing this for potential business reasons? There's NO reason
to waste developing time at this stage.
--
-- Fran�ois M�nard <ymenard>
-- http://ymenard.cjb.net/
-- This announcement is brought to you by the Shimago-Dominguez
Corporation - helping America into the New World...
Perhaps this is a Rendering Simulation?
It (the development of RL) reminds me of the development by Howard Hughes of
his plywood flying boat (dubbed the "Spruce Goose").
It was a magnificent creation - totally awesome and mightily impressive -
but by the time it undertook its one and only flight the reason for its
initial creation had gone; the war was over.
So it appears to be with RL; by the time the first car turns a wheel in
anger, the rest of the simulation world may well have passed it by.
Bruce.
"Dale Green" <[email hidden]> wrote in message
news:[email hidden]...
> I want to believe in RL, I really do, but this is pointless screwingaround.
>"Bruce Kennewell" <[email hidden]> wrote
> So it appears to be with RL; by the time the first car turns a wheel in
> anger, the rest of the simulation world may well have passed it by.
Indeed, with the emergence of simulations like F1C, with third-party addons
made of pretty high-standard car modeling, physics and tracks, WSC.. err RL,
as time goes by, seems more and more useless.
Even Papy entered that game with the ("someday my patch will come, [] and
how thrilling that moment will be...") addition of 3 new set of physics and
the track tools.
So I ask you, West Bros., what's the point!?
--
-- Fran�ois M�nard <ymenard>
-- http://ymenard.cjb.net/
-- This announcement is brought to you by the Shimago-Dominguez
Corporation - helping America into the New World...
On Sun, 14 Sep 2003 03:40:46 -0400, "ymenard"
<[email hidden]> wrote:
>Indeed, with the emergence of simulations like F1C, with third-party addons
>made of pretty high-standard car modeling, physics and tracks, WSC.. err RL,
>as time goes by, seems more and more useless.
I would hardly compare F1C to a professional-quality driving
simulation, which is what I was under the impression the Westies were
aiming for.
Jason
>"jason moyer" <[email hidden]> wrote
> >Indeed, with the emergence of simulations like F1C, with third-partyaddons
> >made of pretty high-standard car modeling, physics and tracks, WSC.. errRL,
> >as time goes by, seems more and more useless.
>
> I would hardly compare F1C to a professional-quality driving
> simulation, which is what I was under the impression the Westies were
> aiming for.
I disagree. F1C is at least an existing sim. RL is just renders created to
attract business possibilities. F1C has tons of fanatics, you know, me,
you, racing fans, who are creating dozens of mods. They are as much devoted
as the West Bros. are with their renders.
Teddy bears, leather wallets to put in your business card CD, a garage where
you roll your car out.. CMON!!!
--
-- Fran�ois M�nard <ymenard>
-- http://ymenard.cjb.net/
-- This announcement is brought to you by the Shimago-Dominguez
Corporation - helping America into the New World...
On Sun, 14 Sep 2003 17:12:58 -0400, "ymenard"
<[email hidden]> wrote:
>I disagree. F1C is at least an existing sim. RL is just renders created to
>attract business possibilities. F1C has tons of fanatics, you know, me,
>you, racing fans, who are creating dozens of mods. They are as much devoted
>as the West Bros. are with their renders.
I dunno, I still see what the Wests say they're aiming for and what
someone like ISI has done as two completely different things. The
Wests are attempting to redefine the driving simulation genre with a
level of atmosphere and attention to detail that has never been
approached. ISI have released 7 games using the same engine and still
don't have a functional tire model.
FWIW, I'm as disappointed as anyone else that it's been over a year
since the hype machine started rolling with the big website countdown
and we haven't seen anything of substance from them. On the other
hand, I'm willing to give them the benefit of the doubt and assume
that the amount of time it's taking to code Racing Legends is a result
of the complexity of simulation they're aiming for. All we can see is
the work being done by the artist, none of the core program has been
revealed to the public yet. If their engine models things like
chassis flex and multiple contact patches with the ability to wear and
flatspot individual sections of tire then they will have done more
than anyone else in the genre has thusfar, and the wait will be worth
it.
Jason
"Mike Beauchamp" <[email hidden]> wrote in message
news:[email hidden]...
> how would you
> graphically show on a website what you have done? Some nice render shotsof
> the cars, or a screenshot of some code from your physics engine?
>
> The physics engine and grapic engine could definately be in development,but
> there's no way of really showing it off yet. Whereas the car models can be
> nicely rendered in software and shown and be very impressive.
At last - an iota of sensibility.
Doug
On Mon, 15 Sep 2003 10:25:52 +0100, Doug Ellison wrote:
> At last - an iota of sensibility.
They could show nice plots of various effects that
they are able to handle in their physics engine,
but I guess that it would not appeal the majority
of the audience. ;)
Everybody seems to be saying that it would take
a hell of a graphics card to run the game. I don't
agree. With this level of detail in the physics it
is necessary to have a rather powerful processor, unless
they manage to "cheat" or eliminate some calculations
by using lookup tables instead... But on the other hand
I've heard of vehicle simulations with over 70 degrees of
freedom running in real time on a not too fancy computer.
The modeled all of the linkage arms and a powertrain
too I think together with a tire model. It was all done
using "dymola" (http://www.dynasim.com). I haven't tried
the software, but if it is possible to create simulation code
for racing games using it I guess that the modelling time
would decrease by several months! I spoke to the founder
of Dynasim about this a couple of weeks ago and he found
the idea interesting, but it is no market for them I guess.
Too little money...
/Stefan
Yeah, found a nice picture of it...
http://www.dynasim.com/www/dymola_2003_scrn.pdf
/Stefan
"Stefan Larsson" said:
IIRC I seem to have read that the physics in todays sims only account for a
few precentage of the CPU usage.... Graphics on the otherhand need a
helluffa lot more CPU- and VC-power.
--
ed_
Quoted text (7 lines)
> > Everybody seems to be saying that it would take > a hell of a graphics card to run the game. I don't > agree. With this level of detail in the physics it > is necessary to have a rather powerful processor, unless > they manage to "cheat" or eliminate some calculations > by using lookup tables instead..."Mike Beauchamp" <[email hidden]> wrote in message news:<[email hidden]>...
Has everyone forgotten the movies of WSC in development that were
released several years ago? Didn't one of them show the suspension
movement of a bare naked 333sp over a test course? IMO, those movies
were a physics demonstration.
RG
RSDG
Quoted text (11 lines)
> Or, they are actually making a game. I guess if you did model all these cars > AND you wrote a physics engine that is still in testing phaze, how would you > graphically show on a website what you have done? Some nice render shots of > the cars, or a screenshot of some code from your physics engine? > > The physics engine and grapic engine could definately be in development, but > there's no way of really showing it off yet. Whereas the car models can be > nicely rendered in software and shown and be very impressive. > > Mike > http://mikebeauchamp.comQuoted text (22 lines)
>From: "Ed Solheim" [email hidden] >Date: 9/16/2003 8:07 AM Central Daylight Time >Message-id: <3f670bc0$[email hidden]> > > >"Stefan Larsson" said: > >> >> Everybody seems to be saying that it would take >> a hell of a graphics card to run the game. I don't >> agree. With this level of detail in the physics it >> is necessary to have a rather powerful processor, unless >> they manage to "cheat" or eliminate some calculations >> by using lookup tables instead... > >IIRC I seem to have read that the physics in todays sims only account for a >few precentage of the CPU usage.... Graphics on the otherhand need a >helluffa lot more CPU- and VC-power. > >-- >ed_ >On Wed, 17 Sep 2003 08:42:56 +0000, J. Todd Wasson wrote:
I've only seen stills on your webpage.
This depends on the time resolution of the simulation. If you only intend
to simulate the vehicle dynamics... Fine. But if you are going into the
engine to simulate the combustion also (wouldn't that be necessary to be
able to adjust valve timings and such?) you would end up with a much
smaller time scale and the resulting model of the car would probably be
stiff (big difference between different time scales of the model). A
smaller time scale combined with a larger model would certainly increase
the computational load. I read somewhere that to be able to simulate a car
responding to small variations in the surface it is necessary to increase
the time resolution to around 8000 - 10000 Hz... Then add some more cars to the
track... And add the possibility to keep track of the wear of many of the
parts in the car and/or engine. It seems as if it would be quite easy to
get the simulation taking a lot of time...
Using better resolution would also require more detailed data... I.e.
more data shuffling. :P
> Granted, if you threw in a bunch of cars running the same model and added good
> collision detection/response code that percentage will climb, but my point is
> that on today's PCs what we've considered to be really nasty car models over
> the past few years will run basically for free. Collision detection/response
> is most likely the greatest hog in a typical physics engine, but then only when
> a few objects get close to each other.
Agree... This is a very complicated problem if it is going to be accurate.
One cannot test collision between all possible surfaces at the same time.
Hopefully these kinds of calculations will be included in future graphics
cards... I read that in an interview with a developer at Nvidia regarding
his vision of GPU:s in the future.
Depends on if you're into entertainment or engineering. In engineering it
is very important to get it right. In entertainment it would not be as
crucial, but then why use a complicated model at all?
The damage models... If you crash into a wall in GPL only the wheels will
fall off. A more accurate damage model where the hull is getting damaged
and probably aerodynamics is affected too and where the damage (and wear) is
inherited to the next race unless you fix it?
> Hand them a physics model with FEM suspension versus one with straight,
> geometrically calculated suspension kinematics, don't say anything, and see if
> anyone notices a difference. The upper rod shortened 0.0000003 mm in turn two.
> Did you catch that? ;-)
Probably. FEM is overkill for games... But I guess it could be used with
advantage in damage simulation?
> The FEM suspension still won't run on today's PCs in real time. Other
> approaches do easily. Either way you can bet that the tire force and
> aerodynamic input data is not accurate enough to make up for the added
> complexity.
Depends on the number of nodes you are using... And if the model is linear
or nonlinear.
> At this stage you could dive into better tire models, more sophisticated
> aerodynamics simulation approaches, and so forth. Keep in mind though that if
> you have good measured data for any of that than there may not be much point.
True. The next step would maybe be to make the computer control a
hydraulic seat where you actually feel the small details much better?
> What impresses me is how guys like Sebastien Tixier deal with programming good
> dynamics for Playstations' CPU :-)
>
> Ack... Got carried away. It's late and I'm getting delirious.. 8-)
Fine with me... I'm discussing this issue due to my own simulations which
are of crank angle resolution... These don't run i real time... :(
/Stefan
Quoted text (7 lines)
> True. You might have seen the bland, simple graphics in my sig. That physics > model does the whole suspension system except for the actual physical ARB > installation and attachment points, tie rods, and steering rack location (not > much CPU time there really anyway with a good approach). Roll centers and so > forth are essentially moving in real time and all that fun stuff, and it still > runs around 200-250 fps on my P4. (That looks worse than 60, quite frankly, > but that's not the point.)Quoted text (15 lines)
> Stefan, lookup tables like you describe are used to speed things up in some > areas, but to give a sense of graphics/physics CPU usage scale: This is > currently running so fast the physics engine has to stop and wait a few > milliseconds every graphics frame because the clock resolution is only 0.001 > second, and if the engine can't get a few cycles in during that 1 millisecond > (meaning it repeats all of the physics calculations several times, including > doing the entire set of tire force calculations several times for each tire), > it'll time warp. At 200-250 fps it time warps by a factor of around 4, so you > tell me what percentage of the CPU time is being used by the physics engine > every cycle :-P 0.0...%? 0.2%? > > I.e., the physics runs so fast the CPU can't even tell the engine how much time > elapsed since the last time the physics calculations RAN, much less how long it > actually took them to run... Again, that timer's minimum duration is 0.001 > second.Quoted text (11 lines)
> As another example, on our current project we're giving the physics engine > 5-10% of the CPU. Really it's not using more than probably 1-2% at this point, > and I've got just about every bell and whistle in there I can think of > (well....almost.... ;-)). Adding some chassis flex or tire distortion would > not impact anything unless you insisted on duplicating FEM engineering software > line for line. What if you could get within 5% of the same results but on code > that runs 10,000 times faster? If your tire data might be off by any more > than 5% there's no point bothering with the FEM approach because the model is > no more accurate either way. It's cool to say "this has an FEM model", but if > it's not more accurate because of it's inclusion, isn't it just hype? (Of > course you have to take your hat off to whoever wrote it anyway!)Quoted text (19 lines)
> Anyway, the other 95-98% currently is all graphics. The vehicle model does > quite a lot of things the GPL model does not do (and skips a couple things it > did), but what's great is that GPL was intended to run on only a Pentium 166 if > I remember correctly. I'm surprised it ran, quite frankly, let alone with such > pretty graphics for the time. That was brilliantly done, but on a 1Ghz+ system > that technology runs pretty much for free as far as the CPU is concerned. > > So really the doors are blown wide open now for physics again. You could > increase the complexity probably 100 or 1000 times and now you'd be back in > David Kaemmer's position a few years ago writing GPL for a Pentium 166. > > So what do you do with all the new CPU power available for the physics > department? > > Quite frankly I don't see much to add to a vehicle physics model that isn't > already being done that a player would ever notice. Give the same physics > model to five different groups of people and tell them to model a certain car > with a certain set of tires on it and you and anyone else will swear you're > driving five different physics models when you get the cars back. You're not.On Wed, 17 Sep 2003 13:33:25 +0200, "Stefan Larsson"
<[email hidden]> wrote:
Still, what extra accuracy would that give you? The problem is that
you're simulating more & more, modeling more inaccuracies with every
model. At some point you lose progress.
Even for scientific things I wonder about the purposes of too much
modeling. The more complex the model, the more off your variable
guesses will be.
As Todd indicated somewhat, for entertainment purposes we're nearly
there, CPU-wise, except perhaps for damage modeling.
Scenegraph operations are almost impossible to do efficiently for
every situation out there. On the other hand, perhaps some techniques
can be thought off that are generally useful and do ok with brute
force.
Just like the Z-buffer is actually a very inefficient brute-force way
of doing Z-sorting, but so suited for hardware that's it's now really
useful. :)
>Depends on if you're into entertainment or engineering. In engineering it
>is very important to get it right. In entertainment it would not be as
>crucial, but then why use a complicated model at all?
Right, the last few percent of the detail is lost already for
entertainment purposes.
For engineering, the data becomes more important than the model, I
think. No use in simulating cylinders to get extra 'realism' if you
give it the wrong data, and often you can't get accurate data either.
>Probably. FEM is overkill for games... But I guess it could be used with
>advantage in damage simulation?
For entertainment, just denting a model at the point of impact is
enough. Who cares if it's totally accurate; as long as it looks dented
& damaged, it's ok.
>> At this stage you could dive into better tire models, more sophisticated
>> aerodynamics simulation approaches, and so forth. Keep in mind though that if
>> you have good measured data for any of that than there may not be much point.
>
>True. The next step would maybe be to make the computer control a
>hydraulic seat where you actually feel the small details much better?
That's what we do. :) A high-end steering wheel does help you feel
details better. But for now, I just use a simple noise function to
simulate road granularity. Works ok. Who can tell the difference if I
were to input the exact road structure? The projects needing that kind
of info are getting more niche.
>> What impresses me is how guys like Sebastien Tixier deal with programming good
>> dynamics for Playstations' CPU :-)
Indeed. :) Doing things in a shortcut way is always more difficult
than taking the real-life approach (well, considering you want
realistic results).
>Fine with me... I'm discussing this issue due to my own simulations which
>are of crank angle resolution... These don't run i real time... :(
That's fortunately just a matter of time.
The problem with simulations at one point is that you need so much
knowledge to build good data for it (i.e. a car built out of
components gets harder to simulate if you model it inside & out).
Simplifications will always be used, no matter if the hardware can
deal with it.
Ruud van Gaal
Free car sim: http://www.racer.nl/
Pencil art : http://www.marketgraph.nl/gallery/
Quoted text (18 lines)
>On Wed, 17 Sep 2003 08:42:56 +0000, J. Todd Wasson wrote: > >... >> I.e., the physics runs so fast the CPU can't even tell the engine how much time >> elapsed since the last time the physics calculations RAN, much less how long it >> actually took them to run... Again, that timer's minimum duration is 0.001 >> second. > >This depends on the time resolution of the simulation. If you only intend >to simulate the vehicle dynamics... Fine. But if you are going into the >engine to simulate the combustion also (wouldn't that be necessary to be >able to adjust valve timings and such?) you would end up with a much >smaller time scale and the resulting model of the car would probably be >stiff (big difference between different time scales of the model). A >smaller time scale combined with a larger model would certainly increase >the computational load. I read somewhere that to be able to simulate a car >responding to small variations in the surface it is necessary to increase >the time resolution to around 8000 - 10000 Hz...Quoted text (13 lines)
>> Granted, if you threw in a bunch of cars running the same model and added good >> collision detection/response code that percentage will climb, but my point is >> that on today's PCs what we've considered to be really nasty car models over >> the past few years will run basically for free. Collision detection/response >> is most likely the greatest hog in a typical physics engine, but then only when >> a few objects get close to each other. > >Agree... This is a very complicated problem if it is going to be accurate. >One cannot test collision between all possible surfaces at the same time. > >Hopefully these kinds of calculations will be included in future graphics >cards... I read that in an interview with a developer at Nvidia regarding >his vision of GPU:s in the future.On Wed, 17 Sep 2003 13:39:57 +0000, Ruud van Gaal wrote:
I don't know, more "accuracy"? Is there any studies done on which level of
details that is unnecessary with e.g. a screen and a force-feedback wheel
as the only user inputs?
I guess that is something that they are aiming at.
> Right, the last few percent of the detail is lost already for
> entertainment purposes.
> For engineering, the data becomes more important than the model, I
> think. No use in simulating cylinders to get extra 'realism' if you
> give it the wrong data, and often you can't get accurate data either.
The only thing I can think of is to simulate more realistic engine sound
and capturing individual misfires and such...
> For entertainment, just denting a model at the point of impact is
> enough. Who cares if it's totally accurate; as long as it looks dented
> & damaged, it's ok.
Hardcore fans care? The average 14-year old PS/2 owner won't. I would
indeed find it quite amusing if the car actually gets affected in a
"realistic" sense.
> That's what we do. :) A high-end steering wheel does help you feel
> details better. But for now, I just use a simple noise function to
> simulate road granularity. Works ok. Who can tell the difference if I
> were to input the exact road structure? The projects needing that kind
> of info are getting more niche.
Probably. After all it's all about fooling the user to think it is real?
Maybe one should work more with psychology than modelling? :)
The upside is that you, as a constructor, learn a lot about the behaviour
of the system you are trying to simulate, even if the model does not
capture the reality precisely... And by building a complicated model
you are more aware of what particular effects a simplification would
lead to.
/Stefan
Quoted text (9 lines)
> Still, what extra accuracy would that give you? The problem is that > you're simulating more & more, modeling more inaccuracies with every > model. At some point you lose progress. > Even for scientific things I wonder about the purposes of too much > modeling. The more complex the model, the more off your variable > guesses will be. > > As Todd indicated somewhat, for entertainment purposes we're nearly > there, CPU-wise, except perhaps for damage modeling.Quoted text (7 lines)
> Scenegraph operations are almost impossible to do efficiently for > every situation out there. On the other hand, perhaps some techniques > can be thought off that are generally useful and do ok with brute > force. > Just like the Z-buffer is actually a very inefficient brute-force way > of doing Z-sorting, but so suited for hardware that's it's now really > useful. :)Quoted text (10 lines)
>>Fine with me... I'm discussing this issue due to my own simulations which >>are of crank angle resolution... These don't run i real time... :( > > That's fortunately just a matter of time. > > The problem with simulations at one point is that you need so much > knowledge to build good data for it (i.e. a car built out of > components gets harder to simulate if you model it inside & out). > Simplifications will always be used, no matter if the hardware can > deal with it.
On Thu, 18 Sep 2003, Stefan Larsson wrote:
> The upside is that you, as a constructor, learn a lot about the behaviour
> of the system you are trying to simulate, even if the model does not
> capture the reality precisely... And by building a complicated model
> you are more aware of what particular effects a simplification would
> lead to.
Haven't been getting all of this thread on my news server...
Anyway, it will be a long time before anyone can model tires completely,
off-line or especially realtime! As I understand it, current finite
element models work fairly well (non-realtime) for constant temp operation,
but add in temperature effects (material properties change) and
the problem gets much bigger.
So we continue to work with various different simplified schemes, some
of which have the advantage of also giving some physical feel for
the tire as a flexible structure that grips and/or slides over the road.
"ymenard" wrote:
> > I would hardly compare F1C to a professional-quality driving
> > simulation, which is what I was under the impression the Westies were
> > aiming for.
> I disagree. F1C is at least an existing sim. RL is just renders createdto
> attract business possibilities. F1C has tons of fanatics, you know, me,
> you, racing fans, who are creating dozens of mods. They are as muchdevoted
> as the West Bros. are with their renders.
True, but EA have put out god knows how many retail versions of this series
and it still feels unfinished. Sure the modders have done some great jobs
but at the back of at least my mind there is the ultimate deficiencies of
the game & graphics engine and the limited multiplayer (and singleplayer
come to that).
I know many will disagree with the graphics engine bit - it makes for great
screen shots but still doesn't feel right on the move IMHO.
Whilst it has provided some fun - I don't feel that F1C provides a basis for
the future of simming. That may lie with what the modders can do with the
enabled NR2003, or who knows, with the Wests
> Teddy bears, leather wallets to put in your business card CD, a garagewhere
> you roll your car out.. CMON!!!
I have to agree, that isn't the future for simming. If RL really is aimed
at the niche enthusiasts then they should know we just want something to
drive, even if it is a single track, single car with beta written on every
trackside advert. That should be priority one, not garages or working engine
internals.
The goal, however, if achieved is way beyond what the EA F1 series ever set
out to attain.
Cheers
Tony
Quoted text (14 lines)
>> True. You might have seen the bland, simple graphics in my sig. That >physics >> model does the whole suspension system except for the actual physical ARB >> installation and attachment points, tie rods, and steering rack location >(not >> much CPU time there really anyway with a good approach). Roll centers and >so >> forth are essentially moving in real time and all that fun stuff, and it >still >> runs around 200-250 fps on my P4. (That looks worse than 60, quite >frankly, >> but that's not the point.) > >I've only seen stills on your webpage.Quoted text (28 lines)
> >> Stefan, lookup tables like you describe are used to speed things up in some >> areas, but to give a sense of graphics/physics CPU usage scale: This is >> currently running so fast the physics engine has to stop and wait a few >> milliseconds every graphics frame because the clock resolution is only >0.001 >> second, and if the engine can't get a few cycles in during that 1 >millisecond >> (meaning it repeats all of the physics calculations several times, >including >> doing the entire set of tire force calculations several times for each >tire), >> it'll time warp. At 200-250 fps it time warps by a factor of around 4, so >you >> tell me what percentage of the CPU time is being used by the physics engine >> every cycle :-P 0.0...%? 0.2%? >> >> I.e., the physics runs so fast the CPU can't even tell the engine how much >time >> elapsed since the last time the physics calculations RAN, much less how >long it >> actually took them to run... Again, that timer's minimum duration is 0.001 >> second. > >This depends on the time resolution of the simulation. If you only intend >to simulate the vehicle dynamics... Fine. But if you are going into the >engine to simulate the combustion also (wouldn't that be necessary to be >able to adjust valve timings and such?)Quoted text (15 lines)
>> Granted, if you threw in a bunch of cars running the same model and added >good >> collision detection/response code that percentage will climb, but my point >is >> that on today's PCs what we've considered to be really nasty car models >over >> the past few years will run basically for free. Collision >detection/response >> is most likely the greatest hog in a typical physics engine, but then only >when >> a few objects get close to each other. > >Agree... This is a very complicated problem if it is going to be accurate. >One cannot test collision between all possible surfaces at the same time. >Quoted text (22 lines)
>> As another example, on our current project we're giving the physics engine >> 5-10% of the CPU. Really it's not using more than probably 1-2% at this >point, >> and I've got just about every bell and whistle in there I can think of >> (well....almost.... ;-)). Adding some chassis flex or tire distortion >would >> not impact anything unless you insisted on duplicating FEM engineering >software >> line for line. What if you could get within 5% of the same results but on >code >> that runs 10,000 times faster? If your tire data might be off by any more >> than 5% there's no point bothering with the FEM approach because the model >is >> no more accurate either way. It's cool to say "this has an FEM model", but >if >> it's not more accurate because of it's inclusion, isn't it just hype? (Of >> course you have to take your hat off to whoever wrote it anyway!) > >Depends on if you're into entertainment or engineering. In engineering it >is very important to get it right. In entertainment it would not be as >crucial, but then why use a complicated model at all? >Quoted text (31 lines)
>> Anyway, the other 95-98% currently is all graphics. The vehicle model does >> quite a lot of things the GPL model does not do (and skips a couple things >it >> did), but what's great is that GPL was intended to run on only a Pentium >166 if >> I remember correctly. I'm surprised it ran, quite frankly, let alone with >such >> pretty graphics for the time. That was brilliantly done, but on a 1Ghz+ >system >> that technology runs pretty much for free as far as the CPU is concerned. >> >> So really the doors are blown wide open now for physics again. You could >> increase the complexity probably 100 or 1000 times and now you'd be back in >> David Kaemmer's position a few years ago writing GPL for a Pentium 166. >> >> So what do you do with all the new CPU power available for the physics >> department? >> >> Quite frankly I don't see much to add to a vehicle physics model that isn't >> already being done that a player would ever notice. Give the same physics >> model to five different groups of people and tell them to model a certain >car >> with a certain set of tires on it and you and anyone else will swear you're >> driving five different physics models when you get the cars back. You're >not. > >The damage models... If you crash into a wall in GPL only the wheels will >fall off. A more accurate damage model where the hull is getting damaged >and probably aerodynamics is affected too and where the damage (and wear) is >inherited to the next race unless you fix it? >Quoted text (10 lines)
>> Hand them a physics model with FEM suspension versus one with straight, >> geometrically calculated suspension kinematics, don't say anything, and see >if >> anyone notices a difference. The upper rod shortened 0.0000003 mm in turn >two. >> Did you catch that? ;-) > >Probably. FEM is overkill for games... But I guess it could be used with >advantage in damage simulation? >Quoted text (8 lines)
>> The FEM suspension still won't run on today's PCs in real time. Other >> approaches do easily. Either way you can bet that the tire force and >> aerodynamic input data is not accurate enough to make up for the added >> complexity. > >Depends on the number of nodes you are using... And if the model is linear >or nonlinear. >Quoted text (9 lines)
>> At this stage you could dive into better tire models, more sophisticated >> aerodynamics simulation approaches, and so forth. Keep in mind though that >if >> you have good measured data for any of that than there may not be much >point. > >True. The next step would maybe be to make the computer control a >hydraulic seat where you actually feel the small details much better? >Quoted text (9 lines)
>> What impresses me is how guys like Sebastien Tixier deal with programming >good >> dynamics for Playstations' CPU :-) >> >> Ack... Got carried away. It's late and I'm getting delirious.. 8-) > >Fine with me... I'm discussing this issue due to my own simulations which >are of crank angle resolution... These don't run i real time... :( >Quoted text (20 lines)
>From: "Stefan Larsson" [email hidden] >Date: 9/18/2003 3:36 AM Eastern Daylight Time >Message-id: <[email hidden]> > >On Wed, 17 Sep 2003 13:39:57 +0000, Ruud van Gaal wrote: > >> Still, what extra accuracy would that give you? The problem is that >> you're simulating more & more, modeling more inaccuracies with every >> model. At some point you lose progress. >> Even for scientific things I wonder about the purposes of too much >> modeling. The more complex the model, the more off your variable >> guesses will be. >> >> As Todd indicated somewhat, for entertainment purposes we're nearly >> there, CPU-wise, except perhaps for damage modeling. > >I don't know, more "accuracy"? Is there any studies done on which level of >details that is unnecessary with e.g. a screen and a force-feedback wheel >as the only user inputs? >Quoted text (19 lines)
>> Scenegraph operations are almost impossible to do efficiently for >> every situation out there. On the other hand, perhaps some techniques >> can be thought off that are generally useful and do ok with brute >> force. >> Just like the Z-buffer is actually a very inefficient brute-force way >> of doing Z-sorting, but so suited for hardware that's it's now really >> useful. :) > >I guess that is something that they are aiming at. > >> Right, the last few percent of the detail is lost already for >> entertainment purposes. >> For engineering, the data becomes more important than the model, I >> think. No use in simulating cylinders to get extra 'realism' if you >> give it the wrong data, and often you can't get accurate data either. > >The only thing I can think of is to simulate more realistic engine sound >and capturing individual misfires and such... >Quoted text (8 lines)
>> For entertainment, just denting a model at the point of impact is >> enough. Who cares if it's totally accurate; as long as it looks dented >> & damaged, it's ok. > >Hardcore fans care? The average 14-year old PS/2 owner won't. I would >indeed find it quite amusing if the car actually gets affected in a >"realistic" sense. >Quoted text (9 lines)
>> That's what we do. :) A high-end steering wheel does help you feel >> details better. But for now, I just use a simple noise function to >> simulate road granularity. Works ok. Who can tell the difference if I >> were to input the exact road structure? The projects needing that kind >> of info are getting more niche. > >Probably. After all it's all about fooling the user to think it is real? >Maybe one should work more with psychology than modelling? :) >Quoted text (16 lines)
>>>Fine with me... I'm discussing this issue due to my own simulations which >>>are of crank angle resolution... These don't run i real time... :( >> >> That's fortunately just a matter of time. >> >> The problem with simulations at one point is that you need so much >> knowledge to build good data for it (i.e. a car built out of >> components gets harder to simulate if you model it inside & out). >> Simplifications will always be used, no matter if the hardware can >> deal with it. > >The upside is that you, as a constructor, learn a lot about the behaviour >of the system you are trying to simulate, even if the model does not >capture the reality precisely... And by building a complicated model >you are more aware of what particular effects a simplification would >lead to.Quoted text (9 lines)
> >/Stefan > > > > > > >Quoted text (21 lines)
>From: Doug Milliken [email hidden] >Date: 9/18/2003 11:12 AM Eastern Daylight Time >Message-id: <[email hidden]> > > >On Thu, 18 Sep 2003, Stefan Larsson wrote: > >> The upside is that you, as a constructor, learn a lot about the behaviour >> of the system you are trying to simulate, even if the model does not >> capture the reality precisely... And by building a complicated model >> you are more aware of what particular effects a simplification would >> lead to. > >Haven't been getting all of this thread on my news server... > >Anyway, it will be a long time before anyone can model tires completely, >off-line or especially realtime! As I understand it, current finite >element models work fairly well (non-realtime) for constant temp operation, >but add in temperature effects (material properties change) and >the problem gets much bigger. >On 25 Sep 2003 09:41:29 GMT, [email hidden] (J. Todd Wasson) wrote:
...
I recently read a very interesting article on floating point
precision. The boundary was something (when using floats) of 10^-4
(0.0001) for the stepsize. After that, floating point roundoff got
worse, so your accuracy actually suffers while you'd think a 0.00001
(10^-5) would give you more.
So with 6000Hz you're already pretty close to the limit, unless you
move to doubles, which for game purposes at least totally overboard.
>>Agree... This is a very complicated problem if it is going to be accurate.
>>One cannot test collision between all possible surfaces at the same time.
>
>Sure you can ;-) But resolving the collision impulses between multiple
>contacts is not something I've seen done even for a rigid body system, so I
>have to agree.
I'm not sure what you mean, but ODE (www.opende.org IIRC) does
multiple contacts quite nicely. Still, I manage to average out all the
generated contact points into 1 (just average the normals and contact
depth) and it still continues to do nicely (if you hit a pilar at the
edge; don't do this at home, kids).
>That would be interesting to read and I'm sure they've given that a lot more
>thought than I have, but then manufacturer's might be getting into
>standardizing something that may not be the best approach for all situations.
>You can get away with simpler collision detection/response in a car sim than
>you could in say a FPS where a dead character falls against a wall and needs to
>bend accordingly.
I agree that a software library which handles low-end things (like
generating contacts between various types of primitives) looks far
more useful. But let's not stop them from trying to think of
something; they might have a brilliant insight somehow which we're
overlooking. ;-)
After all, that's how revolutions come to pass; everybody thinks
something is impossible, then someone approaches the problem in
another way.
> If the graphics card or other add-on "collision chip" is set
>up to handle the complex scenario by design than we're losing a lot of
>optimization potential. It's a waste.
Unless the hardware is significantly faster at the brute force method.
That's what Z-buffers are doing for example; wasting a lot of power
where you could mostly just sort your polygons. But in practice, the
hardware is faster than your Z-sorter often.
>>True. The next step would maybe be to make the computer control a
>>hydraulic seat where you actually feel the small details much better?
>
>Ruud is working on this :-) Looks very cool :-)
Feels very cool too. :)
Ruud van Gaal
Free car sim: http://www.racer.nl/
Pencil art : http://www.marketgraph.nl/gallery/
Quoted text (9 lines)
>>smaller time scale combined with a larger model would certainly increase >>the computational load. I read somewhere that to be able to simulate a car >>responding to small variations in the surface it is necessary to increase >>the time resolution to around 8000 - 10000 Hz... Then add some more cars to >>the >>track... > >Maybe.. I've run my vehicle model at 6000Hz and things were still pretty fast. > It depends on how you're handling the suspension, I suppose.Quoted text (25 lines)
>And add the possibility to keep track of the wear of many of the >>parts in the car and/or engine. It seems as if it would be quite easy to >>get the simulation taking a lot of time... >> > >If wear was being handled through methods used in the wear research papers >I've >seen, then yes, this could kill you. However, then we get back into an area >where adding complexity doesn't necessarily improve accuracy. Are you sure >the >component moduli and so forth are accurate? I don't think anybody except >manufacturers have data on brake wear, for instance. Here's an area where >you >could probably do just as well with a simplified model because of so many >unknown variables. > >For instance, tire wear in my system is handled through a very basic, simple >equation and then adjusted with a coefficient. As long as the basic >relationship is right you can fine tune things. The same could easily be >done >for all the other parts on the car without a noticable hit. You could >probably >get away with modelling wear for a particular part with only 4-10 math >operations per cycle (my PII-333MHz could do about 50-60 million per second). >Quoted text (26 lines)
>From: [email hidden] (Ruud van Gaal) >Date: 9/25/2003 6:43 AM Eastern Daylight Time >Message-id: <[email hidden]> > >On 25 Sep 2003 09:41:29 GMT, [email hidden] (J. Todd Wasson) wrote: > >... >>>smaller time scale combined with a larger model would certainly increase >>>the computational load. I read somewhere that to be able to simulate a car >>>responding to small variations in the surface it is necessary to increase >>>the time resolution to around 8000 - 10000 Hz... Then add some more cars to >>>the >>>track... >> >>Maybe.. I've run my vehicle model at 6000Hz and things were still pretty >fast. >> It depends on how you're handling the suspension, I suppose. > >I recently read a very interesting article on floating point >precision. The boundary was something (when using floats) of 10^-4 >(0.0001) for the stepsize. After that, floating point roundoff got >worse, so your accuracy actually suffers while you'd think a 0.00001 >(10^-5) would give you more. >So with 6000Hz you're already pretty close to the limit, unless you >move to doubles, which for game purposes at least totally overboard. >Quoted text (13 lines)
>>>Agree... This is a very complicated problem if it is going to be accurate. >>>One cannot test collision between all possible surfaces at the same time. >> >>Sure you can ;-) But resolving the collision impulses between multiple >>contacts is not something I've seen done even for a rigid body system, so I >>have to agree. > >I'm not sure what you mean, but ODE (www.opende.org IIRC) does >multiple contacts quite nicely. Still, I manage to average out all the >generated contact points into 1 (just average the normals and contact >depth) and it still continues to do nicely (if you hit a pilar at the >edge; don't do this at home, kids). >Quoted text (18 lines)
>>That would be interesting to read and I'm sure they've given that a lot more >>thought than I have, but then manufacturer's might be getting into >>standardizing something that may not be the best approach for all >situations. >>You can get away with simpler collision detection/response in a car sim than >>you could in say a FPS where a dead character falls against a wall and needs >to >>bend accordingly. > >I agree that a software library which handles low-end things (like >generating contacts between various types of primitives) looks far >more useful. But let's not stop them from trying to think of >something; they might have a brilliant insight somehow which we're >overlooking. ;-) >After all, that's how revolutions come to pass; everybody thinks >something is impossible, then someone approaches the problem in >another way. >Quoted text (9 lines)
>> If the graphics card or other add-on "collision chip" is set >>up to handle the complex scenario by design than we're losing a lot of >>optimization potential. It's a waste. > >Unless the hardware is significantly faster at the brute force method. >That's what Z-buffers are doing for example; wasting a lot of power >where you could mostly just sort your polygons. But in practice, the >hardware is faster than your Z-sorter often. >Quoted text (7 lines)
>>>True. The next step would maybe be to make the computer control a >>>hydraulic seat where you actually feel the small details much better? >> >>Ruud is working on this :-) Looks very cool :-) > >Feels very cool too. :) >Quoted text (10 lines)
> >Ruud van Gaal >Free car sim: http://www.racer.nl/ >Pencil art : http://www.marketgraph.nl/gallery/ > > > > > >On 25 Sep 2003 11:17:35 GMT, [email hidden] (J. Todd Wasson) wrote:
...
Ah indeed, sorry about that. ODE's penalties seem ok though; it allows
for some interpenetration, which probably can save you from doing too
much collision checks to make sure every object doesn't
interpenetrate.
And there was a linear step collision solver that was talked about
some time ago (I stopped reading the maillist) which avoided big
matrices and instead could do it with little matrices step by step
(and faster too).
Yep, some time ago. :) Actually one drove it once.
Now that my cars can tumble over without flying into space though, I
might need to test again. ;-)
Ruud van Gaal
Free car sim: http://www.racer.nl/
Pencil art : http://www.marketgraph.nl/gallery/
