Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Flight Unlimited - Fluid Dynamics ?

34 views
Skip to first unread message

Mr GS Dietachmayer

unread,
Jun 26, 1995, 3:00:00 AM6/26/95
to
There seems to be general agreement in this newsgroup that
Flight Unlimited (FU) models the dynamics of flight
very well. On the back of the box, in the game manual,
and in articles in PC Gamer, the designers attribute
this realism to the use of Computational Fluid
Dynamics (CFD) to model the flow around an aircraft.
I confess I find this claim somewhat interesting, as
most of the CFD I see in this field is the sort of stuff
that takes many hours of computing time on high end
workstations and super computers. I suspect that
Looking Glass and I have different interpretations
of 'CFD'. Can anyone, particularly someone from LG,
comment on the techniques/approximations used in the
CFD implemented in FU?

Cheers,
Gary.

Net Paranoia Caveat One:
I work in CFD for a living, and whilst there are certainly
many greater experts than I, I do have some idea what I am
talking about.

Net Paranoia Caveat Two:
I am _not_ criticising FU the game at all, merely seeking
some clarification on how it it works.

Looking Glass tech support

unread,
Jun 26, 1995, 3:00:00 AM6/26/95
to
g...@momth2.maths.monash.edu.au (Mr GS Dietachmayer)
wrote:

>There seems to be general agreement in this newsgroup that
>Flight Unlimited (FU) models the dynamics of flight
>very well. On the back of the box, in the game manual,
>and in articles in PC Gamer, the designers attribute
>this realism to the use of Computational Fluid
>Dynamics (CFD) to model the flow around an aircraft.
>I confess I find this claim somewhat interesting, as
>most of the CFD I see in this field is the sort of stuff
>that takes many hours of computing time on high end
>workstations and super computers. I suspect that
>Looking Glass and I have different interpretations
>of 'CFD'. Can anyone, particularly someone from LG,
>comment on the techniques/approximations used in the
>CFD implemented in FU?
>
>Cheers,
> Gary.

Hi Gary,

Obviously there is no way of doing the type of CFD that
is done on high end work stations and Crays. However we
don't need to compute to 20 decimal places. We only need
to be accurate enought to render to a 640x480 or 1024x768
pixel screen.

Whether this is really CFD depends on how one defines
CFD. We do really compute the flow of air as a fluid
over the control surfaces of the planes we model, but not
anywhere near as precisely as do those who are designing
aircraft or any real life applications.

Anyway, regardless of whether the CFD claim is marketing
hype or not, Flight Unlimited models flight in a more
realistic manner than other programs for the pc.

Steve
LGT


Mr GS Dietachmayer

unread,
Jun 28, 1995, 3:00:00 AM6/28/95
to
In article <3sn2lj$5...@sundog.tiac.net> Looking Glass tech support

I'm afraid these statements make no sense to me.
'Industrial strength' CFD doesn't work to 20 decimal
places either. The point I was trying to make is that
to get even the crudest possible approximation (i.e., errors
of 50% or more) of airflow around an object as complex
as an aircraft using CFD techniques that I am familiar with
requires _hours_ of computing time.
Also, I would have thought that the flight model and the graphical
display were totally independent issues. A military spec
program might have a very sophisticated flight model coupled
with just a text output of coordinates, speeds, angles, etc.
In other words, inaccuracies in a flight model would show up
irrespective of the level of sophistication in the graphical
output.

>
>Whether this is really CFD depends on how one defines
>CFD. We do really compute the flow of air as a fluid
>over the control surfaces of the planes we model, but not
>anywhere near as precisely as do those who are designing
>aircraft or any real life applications.

If you 'really compute the flow of air as a fluid' in real
time then I think its fair enough to claim that you are
doing CFD. However, I am not aware of any CFD technique that will
let you do this (again, in real time) on a PC to _any_
level of accuracy. Obviously I am not privvy to the inner
workings of FU, but if I were asked how its flight model
worked I would guess that:
a) You spent hours running CFD programs to get the flow
around the aircraft over a wide range of conditions,
and that the program has this data within it. When
FU runs it looks at what the aircraft is currently
doing, and then interpolates the current airflow
from its 'library' of _pre-calculated_ flows.
b) Basically the same as a) except wind tunnel data is
used to provide the 'library'.
Again, I am guessing, and would be delighted to be proved wrong
by an explanation of _how_ you compute the flow.

>
>Anyway, regardless of whether the CFD claim is marketing
>hype or not, Flight Unlimited models flight in a more
>realistic manner than other programs for the pc.

A few points here. First, let me thank you for taking the time
and interest to respond. Second, the 'marketing hype' comment
worries me a little, but that is purely my personal view.
In any advent, a discussion on the ethics of advertising is
not relevent to this group and would lead inevitably to a flame
fest. Third, I guess the CFD claim really piqued my interest
because (according to PC Gamer) you have a physics graduate
on staff, and he might be expected to take a different approach
to modelling of the flight dynamics. Indeed, as the favourable
comments that FU has received suggests, perhaps he has made
significant advances in this field, it would just be nice to
know how he did this (as CFD does not seem the likely candidate).

>
>Steve
>LGT
>
>
>

Looking Glass tech support

unread,
Jun 28, 1995, 3:00:00 AM6/28/95
to
g...@momth2.maths.monash.edu.au (Mr GS Dietachmayer) wrote:

>I'm afraid these statements make no sense to me.
>'Industrial strength' CFD doesn't work to 20 decimal
>places either. The point I was trying to make is that
>to get even the crudest possible approximation (i.e., errors
>of 50% or more) of airflow around an object as complex
>as an aircraft using CFD techniques that I am familiar with
>requires _hours_ of computing time.
>Also, I would have thought that the flight model and the graphical
>display were totally independent issues. A military spec
>program might have a very sophisticated flight model coupled
>with just a text output of coordinates, speeds, angles, etc.
>In other words, inaccuracies in a flight model would show up
>irrespective of the level of sophistication in the graphical
>output.
>

Well a very similar discussion was held on this very topic in
this newgroup in April. Our physics person did participate in
that discussion. My response was a somewhat limited condensation
of some of the comments that were posted earlier. Unfortunately,
he is already busy working on a new project and most likely
does not have the time to repeat the previous posting.

Additionally, I hope that you can uderstand that we don't want to
undermine our competative position by revealing secrets that
our competitors don't know, but would like to use.

However, I can say that interaction of a fluid (air) moving over
the control surfaces of the planes is computed in real time. If
that does not fit your definition of cfd then I guess then to your
we do not use CFD.

I will point out that I have received email from many pilots who fly
the real versions of these aircraft commenting on how similar the
planes modeled in Flight handle compared to the real thing.

>If you 'really compute the flow of air as a fluid' in real
>time then I think its fair enough to claim that you are
>doing CFD. However, I am not aware of any CFD technique that will
>let you do this (again, in real time) on a PC to _any_
>level of accuracy. Obviously I am not privvy to the inner
>workings of FU, but if I were asked how its flight model
>worked I would guess that:
> a) You spent hours running CFD programs to get the flow
> around the aircraft over a wide range of conditions,
> and that the program has this data within it. When
> FU runs it looks at what the aircraft is currently
> doing, and then interpolates the current airflow
> from its 'library' of _pre-calculated_ flows.
> b) Basically the same as a) except wind tunnel data is
> used to provide the 'library'.
>Again, I am guessing, and would be delighted to be proved wrong
>by an explanation of _how_ you compute the flow.
>

The physics calculations are all done in real time. You can do
any maneuver in Flight that is possible in real life in any order
you choose. That's the failing of look up tables is that it is very
easy to choose to do a series of maneuvers that are not contained
in the look up table and then the reality breaks down. I can also
tell you that the most powerful computer used in the Flight development
process was a P-90.

If you choose not believe me then all I can say is sorry. It would be
commercial suicide to reveal the physics behind Flight, because if we did
Microsoft, Microprose, etc. would all stop using lookup tables and their
planes would fly as realisticly as our do.

Steve
LGT


Mr GS Dietachmayer

unread,
Jul 3, 1995, 3:00:00 AM7/3/95
to

I felt that the messages in this thread were beginning
to become somwhat long and unwieldy, hence the following
summary and additional comments. Any corrections by LGT
will be most welcome.

1) AGREEMENT
That FU (Flight Unlimited) is an excellent game in general,
and that its flight modelling has been particularly well
received.

2) AGREEMENT
That 'the real-time computation of the interaction of a fluid
(air) moving over the control surfaces of an aircraft' can be
legitimately termed as Computational Fluid Dynamics (CFD).
(The definition in quotes is that of Steve/LGT.)

3) DISAGREEMENT
That CFD as defined in (2) is possible on a computer no
more powerful than a Pentium. By 'possible' here I mean that
the CFD results are sufficiently accurate that they are useful
in determining the flight dynamics for the various aircraft
modelled in FU.
LGT contends yes, this is possible, I contend no.
(LGT's claims as represented in the March issue of PC-GAMER
are actually much stronger than this. There they argue that
they have the ability to model the flight dynamics of an
arbitrary object, and that their flight modelling is more
accurate than that based on wind-tunnel testing.)

4) ARGUMENTS FOR (3).

4a) LGT don't seem to have said this explictly, but implicitly
their argument seems to run :

i) We have very good flight models in FU.
ii) Other games without CFD have relatively poor
flight models.
iii) The combination of (i) and (ii) implies that it
is the 'CFD' in FU that leads to the accurate
flight models.

The obvious flaw in this argument is that there may be
other ways (besides 'CFD') to achieve a good flight
model. Indeed, the military-spec simulators that I am
aware of do not use CFD at all.

5) ARGUMENTS AGAINST (3).

5a) If LGT can really do the calculations described in (2)
they are wasting their time with mere PC games. They
have achieved in a few years what has not been done
in decades by thousands of people working at
Universities, Government Departments, and private
organizations (such as Boeing). A similar achievement
in medicine might have the headline 'doctor performs
open heart surgery with nothing but his bare hands
and two band-aid strips'.

6) FURTHER COMMENTS

6a) From Steve/LGT's last post: 'If you choose not to believe
me then all I can say is sorry'. I am sure that Steve is an
intelligent, hard-working and honest individual. However, _belief_
has nothing to do with the issue. I don't _choose_ to
believe or disbelieve. LGT has made some extravagant claims,
with little evidence (apart from (4)) to back them up.
Offer me some arguments based on fluid dynamics/CFD and I
will be happy to re-examine my position.

6b) According to Steve/LGT this discussion was dealt with fully
in this Newsgroup in April. If anyone has kept a copy of the
relevent posts I would really like to see them.

6c) I understand that your physics person is extremely busy
(I imagine you all are), but how long does it take to
type something like 'we use a simplified version of
Bernoulli's equation, applied at just a few points on
the aircraft'?

6d) I understand concerns about commercial confidentiality,
but I don't want the details. In any case, the game world
seems to be advancing so fast that by the time your
competitors could take advantage the technique may
be close to obsolete (witness Doom).

6e) I imagine that FU is selling very well. Do none of its
users have an interest in how it works??????

Cheers,
Gary.

stu...@edufee.fee.uva.nl

unread,
Jul 3, 1995, 3:00:00 AM7/3/95
to
Hi,

I bought FU last week and it doesn't work.
I tried it with and without the Univbe driver and made the
bootdisk according to the install guide, nothing works....
My system is:
486/50DX2 VLb, with 8 mb
S3 SVGA VLB Videocard, VBE1.2 compliant
Soundblaster Pro + Double-speed CDROM

After starting flight.exe, i hear the sound of a starting engine
through the speaker, than the system gives a divide overflow
error and hangs.
Does anyone know if this is a videocard problem or should i
perhaps upgrade to 16mb ram to use FU?

Greetings, Andre.

E-mail(internet): 922...@edufee.fee.uva.nl


Michael Shavit

unread,
Jul 3, 1995, 3:00:00 AM7/3/95
to
In article <3t7v99$f...@harbinger.cc.monash.edu.au>,
g...@momth2.maths.monash.edu.au says...


Hello LGT, please comment to the writing from Dietachmayer. I bought your
game, it's cool, it's great, it's fantastic, even if it doesn't work with
CFD...

michael


M.Sices

unread,
Jul 3, 1995, 3:00:00 AM7/3/95
to

>
>5a) If LGT can really do the calculations described in (2)
> they are wasting their time with mere PC games. They
> have achieved in a few years what has not been done
> in decades by thousands of people ...

I do not know very much about CFD (and thats why I bet a
lot of folks aren't piping up about this), but my brother
is a physicist. He knows nothing about flight sim games at
all, but when I asked him about LGTs claim, he immediately
stated his opinion that it was a marketing ploy. He stated
that unless LTG has come up with some algorithm that no one
knows of (including NASA etc), that the CFD involved would
have to be simplified to the point that it would no longer
-be- CFD (ie provide that level of realism)


>Do none of its
> users have an interest in how it works??????
>

I find that FUs level of realism in flight simulation is
by far the best/most refined I have tried to date. And I
am a pilot myself, and have flown almost every other sim
out there.

There is one other flight game, Confirmed Kill, which claims
to have a model which is based on airflow over a wing shape
(Im not sure if they used the term CFD) as opposed to a
'table model' (where the sim looks up pre-computed/ inputed
performance characteristics given certain conditions). As
such, it does seem to me that some 'new' way of simulating flight
is being implemented lately. This is good.


Still, instead of calling it Computational Fluid Dynamics - and
having all of us physics dweebs getting shot down by our
brothers, LGT should have come up with a more accurate term for
what they were providing.


--
Michael Sices, dr...@cris.com
AirWarrior (#4135), Draco -*FE*- CO: Flying Elvises!
http://www.cris.com/~draco

Greg Chamberlain

unread,
Jul 3, 1995, 3:00:00 AM7/3/95
to

>
>There is one other flight game, Confirmed Kill, which claims
>to have a model which is based on airflow over a wing shape
>(Im not sure if they used the term CFD) as opposed to a
>'table model' (where the sim looks up pre-computed/ inputed
>performance characteristics given certain conditions). As
>such, it does seem to me that some 'new' way of simulating flight
>is being implemented lately. This is good.
>
>

Confrimed Kill is not using the table model, that I know for sure.
If memory serves correctly CK is using force modeling...


James Kingdon (PAC)

unread,
Jul 4, 1995, 3:00:00 AM7/4/95
to
>>>>> "Greg" == Greg Chamberlain <gr...@basenet.net> writes:

In article <3t9sb9$3...@mail.one.net> gr...@basenet.net (Greg Chamberlain) writes:

>> There is one other flight game, Confirmed Kill, which claims
>> to have a model which is based on airflow over a wing shape (Im
>> not sure if they used the term CFD) as opposed to a 'table
>> model' (where the sim looks up pre-computed/ inputed
>> performance characteristics given certain conditions). As
>> such, it does seem to me that some 'new' way of simulating
>> flight is being implemented lately. This is good.

Greg> Confrimed Kill is not using the table model, that I know for
Greg> sure. If memory serves correctly CK is using force
Greg> modeling...

Just making a wild guess, I'd say that force modeling probably is the
term that Looking Glass should have used to describe their model, which
would have avoided all this tedious stuff about the one true meaning of
CFD.

Overall, FU feels more like flying than CK does. It's hard to pin down
why, but it might be something to do with the inertia in the yaw and
pitch of the aircraft. Then again, it might just be that I'm biased
because I don't usually fall out of the sky in FU :)

James.

0 new messages