PDA

View Full Version : Royal Aeronautical Society interview Oleg



Philipscdrw
01-03-2007, 07:06 AM
Another piece of info from the Birmingham IFC! This came from the Royal Aeronautical Society "The Aerospace Professional" magazine. Interesting, isn't it, that they'd interview Oleg and neglect the massed ranks of Microsoft Flight Simulator developers?...

Ahem. This was scanned from the magazine and then converted to text with an optical character recognition program. Scanner mistakes are left in because they're funny.


From Cold War aero engineer to flight sim king

At the International Flight Simulation Convention in Birmingham, UK, Tim Robinson interviewed Oleg Maddox, the creator of the award-winning 11-2 Sturmovik series of PC flight sims about his career, the latest addition to the series, Sturmavik- 1946 and has a preview of the new Battle of Britain: Storm of War flight sim, slated to be released in December 2007 and published in the west by Ubisoft
The 11-2 series, since it was released in 2001, has won critical acclaim from both gamers and real pilots and sparked an interest in the Eastern front air war in WW2, thanks to its sublime flight modelling (based on blade-element calculations that allow for spins and stalls which are calculated on the fly), outstanding graphics, high level of detail and historical accuracy.
Its creator, Oleg Maddox, has had an interesting career for a games designer. A former aeronautical engineer at a military technical institute in the Russian aerospace industry, his standards of perfection for aerodynamic modelling have raised the bar for flight simulation on the home PC
Since its early beginning the series has added more aircraft and maps through both free patches and paid for add-ons such as the /Aces Expansion Punk (reviewed in The Aerospace Professional by Captain Eric Brown, Past President of the Society), and Pacific Fighters to encompass almost all the theatres of WW2 and aircraft from biplanes like the Gladiator to early jets like the YP-80 and Me2S2 to become, in the end, less of a game and more an interactive aviation history experience.
The latest ? all in one edition, Sturmovik 1946, (published by Ubisoft] ? was released in December and adds a number of what-if aircraft prototypes on the German side and historical aircraft on the Russian side for a hypothetical 1946 scenario, as well as extra aircraft for the Far East, such as the Ki-27 for Flying Tiger; scenarios and extra maps for Manchuria, Kiev and Burma
This, however, will be the last hurrah for the old 11-2 'engine' ? because Oleg and co have been working on a new sim, Battle of Britain- Storm of War (BoB.SoW), This is planned to be a true sequel to 11-2, featuring even more
graphical detail, more advanced flight modelling and promises to be the most historically accurate simulation of that pivotal air battle ever
We put a few questions to Oleg about his background and the links between PC home simulation and real aviation.
Q: Oleg can you tell us about your background?
A: I was 11 years working in the military flight research in Russia.
Q: Were you working on simulation in the Russian aerospace industry?
A: No ? it was classified but involved military aircraft and
electronics.
Q: Why did you leave?
A: The end of Cold War ? many things changed and if I stayed I would have ended up making microwave ovens.
Q: Have you ever considered going back into the aerospace industry?
A: No ? don't.pay enough money now!
Q: How did your interest in flight simulation start? What was your inspiration?
A: I was really interested in computers and what they could do. I used to play F-29 Reta/iator and Wing Commander and wanted to do my own flight sim but in those days the technology wasn't there yet to do it properly. I am also a paraglider pilot as well. By about 1997-1998 the PC hardware was closer to the point where we could realise my vision.
Q: The 11-2 series has been an outstanding success ? did you ever
think that it would be so successful?
A: Of course ? it was made by us! Seriously, we couldn't expect the level of success.
Q: What input have you had from real pilots on the series?
A: A lot of our Beta testers on the team are real pilots. Feedback from pilots has been extremely good. In addition the Russian Knights
display team also use it for training and for giving virtual aerobatic demonstrations at arrshows.
Q: Now that you will be releasing BoB.SoW next year ? is there any thought to 'opening up' 11-2 to a/low people to create their own maps, aircraft?
A: We will not be making 11-2 'open source' ? but will license it for third party developers
Qr What is new in BoB;SoW?
A: It is a whole new step and whole new level of standards. In essence it will be '11-2 2'. It will be much more 'open' than 11-2 and we will give people the ability to add their own aircraft, maps and objects. But for online flying and multiplayer competitions, there will be a core 'Maddox certificated' set of aircraft to avoid cheating. After the Battle of Britain, the plan is to expand the sim with other aircraft and other theatres.
Q: Have you ever been tempted to develop a modern or Cold War jet sim ? for example to model the Su-27 or MiG-29?
A: No ? I'm only interested in going up to Korea ? WW2 air battles with guns-only dogfights are much more interesting.
Q: In Storm of War you are including a Su-26 aerobatic aircraft as a bonus aircraft ? why is that?
A: This was put in by the request of some users. Not everyone wants to shoot and fight. Some pilots (both real and virtual] want to fly aerobatics. It is also there to increase trust in our flight modelling as a comparison to the period aircraft. There are not many warbird pilots in the world but there is real data and real pilots for this aircraft. In fact some of the BoB team also worked on the real Su-26 for Sukhoi.
Q: Some aircraft in 1946 are extremely strange and never flew ? how did you conduct research for these?
A: (llya Shevchenko, lead developer at 1C: Maddox Games and head of RRG Studios.) Some of these we have lots of details for ? e.g. for the MiG-9 there is lots of
documentation. For the what-if aircraft, we had to use our own expertise in aeronautical engineering and discover solutions. For example, we found that the German Heinkel Lerche VTOL was hopelessly underpowered ? we had to double the engine power to get it to fly! The Ta-1S3 also we found had a very weak tail that needed strengthening. So these aircraft required even more research than usual.
(So in a sense by developing these virtual aircraft with a high fidelity simulation ? Maddox and co had conducted important historical research themselves ? which shows that even if the war had progressed another year ? flaws in these radical designs would have probably prolonged their development and not been the miracle weapons that the Luftwaffe hoped for).
0: Bo8:5oW will feature some impressive weather effects. Can you te/l us about these? Will they be historically accurate to the day's weather it history?
A: (Oleg) The sims dynamic campaign excludes having exact
day's weather but it will be close and will be within a range of available weather. In BoBSoW we will also be doing some extremely advanced weather modelling, modelling weather fronts, air currents, updrafts, downdrafts in a dynamic, evolving weather system Real pilots will be able to recognise cloud types and know that a certain type of turbulence is present. It will be a real time weather system across the whole map
Q: The Battle of Britain obviously featured radar for trie first time. Will this be modelled?
A: Radar will be integrated into the gameplay.
Q: 11-2 features multiplayer play ? how will you recreate the massive formations of that period both offline and online?
A: There will be scalable Al to reduce the processor load when Al
aircraft are at long range. Online play will support up to 128 players at once.
Q: The last Battle of Britain flight sins ? BoB II: Wings of Victory
featured an impressive strategic game within a game. Will Storm of War feature such a strategic layer or will it be entirely focused on the single pilot experience?
A: (Oleg] There will be advanced squadron management. A: (llya) In addition, in eoe;5otVwe will feature 'individual' Al instead of the four levels of skill in IL-2. In BoB there will be Al pilots who are good shots, have low morale or are more aggressive etc. etc. to give much more combinations and make to feel you are facing or flying with individual humans.
Q; Many people new to flight sims found 11-2 a steep learning curve ? will BoB:SoW feature better interactive training to get those unfamiliar with torque, stalls etc. up and running?
A: There will be improved training.
(Indeed there is the possibility of the inclusion of a flyable Tiger Moth in the finished product as a dual control primary trainer.]
Conclusion
I was privileged to get a brief hands-on preview of Battle of
Arado Ar234B.
Britain: Storm of War at the show. The build is an early one and there is still a year to go, but even now the sim is looking extremely impressive. Cockpit detail in the Bf109 is photorealistic and now allows for six degrees of freedom (so with the mouse or TrackIR device) you can look around struts or the cowling Terrain below now looks like south east England with full 3D trees and woods and realistic white cliffs of Dover In addition there was buffeting when I flew near the edge of the cliffs showing the superior updraft modelling built into the sim. To top it all it ran extremely smoothly ? even when flying around at low level. We look forward to this milestone in PC flight sims being released in late 2007. * <div class="ev_tpc_signature">

------------------------------------------------------------
PhilipsCDRw

PF_Tini's Simple Guide to Switching 4.04m, 4.05m, and 4.07m. (http://forums.ubi.com/eve/forums/a/tpc/f/23110283/m/7351046415)
Flying on Hyperlobby as EAF_T_Dozer

Makabi-
01-03-2007, 11:51 AM
thanks for posting that. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

F19_Ob
01-03-2007, 12:10 PM
This is extreemly bad news......for all other simulators out there.

http://forums.ubi.com/images/smilies/agreepost.gif<div class="ev_tpc_signature">

http://img.photobucket.com/albums/v382/f19_ob/ob_ver2.jpg

MOH_Hirth
01-03-2007, 12:15 PM
With SOW sucess, "Sir" Oleg will will drink tea with Queen Elisabeth.

CHDT
01-03-2007, 01:29 PM
No ? I'm only interested in going up to Korea


Ihaaaaaaa :-))))))

BadA1m
01-03-2007, 02:07 PM
Brings tears to my eyes.......<div class="ev_tpc_signature">

There is only one tactical principle which is not subject to change. It is to use the means at hand to inflict the maximum amount of wound, death, and destruction on the enemy in the minimum amount of time."
- General George Patton Jr

Tator_Totts
01-03-2007, 04:46 PM
Originally posted by CHDT:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">No ? I'm only interested in going up to Korea


Ihaaaaaaa :-)))))) </div></BLOCKQUOTE>


Dream come true.<div class="ev_tpc_signature">

http://home.carolina.rr.com/squad/AG-51/Stanger.gif

Feathered_IV
01-03-2007, 07:13 PM
Sadly, microwave ovens was not a miss-print http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif<div class="ev_tpc_signature">

***********************************************

http://server2.uploadit.org/files/Feathered-sigpic.jpg

"Intelligent, normally observant and answered all questions freely. He was arrogant and proud to be a pilot. Fellow prisoners in hospital consider him mentally unstable."

Waldo.Pepper
01-03-2007, 09:08 PM
What's he got against microwave ovens!<div class="ev_tpc_signature">

http://img.photobucket.com/albums/v516/WaldoPepper/sig/p61rev.jpg

NonWonderDog
01-04-2007, 10:22 AM
So, IL-2's engine works the same basic way as X-Plane's? That was always my guess, but it's nice to see it printed somewhere. I remember being ridiculed in some argument about stalls when I said that was probably the way it worked, anyway.

NAFP_supah
01-05-2007, 05:43 AM
Oleg is a paraglider pilot? :| You'd think he could atleast afford to fly cessna or something.<div class="ev_tpc_signature">

http://supah.chaotic.nl/profiles/profile-sig.jpg (http://supah.chaotic.nl)

mrsiCkstar
01-05-2007, 06:23 AM
KOREA!!! http://i3.photobucket.com/albums/y60/thegnrbar/emoticons/yay.gif <div class="ev_tpc_signature">

http://home.no.net/jonmarja/images/f4usig.jpg

Mysticpuma2003
01-05-2007, 06:54 AM
I preferred the video interview http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

Seriously, thanks for the info, every little helps!

Cheers, MP.<div class="ev_tpc_signature">

http://www.aqqm31.dsl.pipex.com/Mysticpuma.jpg

Ugly_Kid
01-05-2007, 09:54 AM
Originally posted by NonWonderDog:
So, IL-2's engine works the same basic way as X-Plane's? That was always my guess, but it's nice to see it printed somewhere. I remember being ridiculed in some argument about stalls when I said that was probably the way it worked, anyway.

No, it does not. If it is written somewhere it does not make it automatically right, that article has quite a few mistakes. IL-2 does not employ blade element analysis as X-Plane, Oleg even himself wrote as much in earlier discussions.<div class="ev_tpc_signature">

http://www.f19vs.se/fokker_now.jpg

Marcel_Albert
01-05-2007, 10:35 AM
Thank you for posting this interview http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

I guess we'll need a powerful multicore cpu to run it , probbaly quad core with all the calculations and effects of the weather and more complex ai .

The project is truly exciting , i can't wait for SoW:Bob , i wish all the best for 1:C and Oleg for this job , great project http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif<div class="ev_tpc_signature">

---------------------------------
"Non Nobis Domine , Non Nobis , Sed Nomini Tuo Da Gloriam."
In Memoriam Jacques de Molay .

NonWonderDog
01-05-2007, 03:59 PM
Originally posted by Ugly_Kid:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by NonWonderDog:
So, IL-2's engine works the same basic way as X-Plane's? That was always my guess, but it's nice to see it printed somewhere. I remember being ridiculed in some argument about stalls when I said that was probably the way it worked, anyway.

No, it does not. If it is written somewhere it does not make it automatically right, that article has quite a few mistakes. IL-2 does not employ blade element analysis as X-Plane, Oleg even himself wrote as much in earlier discussions. </div></BLOCKQUOTE>

I'm sure it's not really analogous to X-Plane (I don't know why I said that, because FB doesn't have to analyze arbitrary user-created planes), but it certainly calculates more of the FM in real time than CFS3 did. I'm sure you've watched your target's separated wings fall to earth after a kill?

I wouldn't be so quick to say it definitely doesn't use some part of blade-element theory as a base for some portion of the FM. It seems fairly likely to me, if it does any real-time wing calculations at all. If it does, though, it would definitely be significantly different and more optimized/simplified than X-Plane's implementation. The X-Plane guy didn't invent blade-element theory, you know, and Oleg might have just been separating himself from his competitors with that old quote (he does that a lot, if you haven't noticed http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif). Blade-element analysis has been around for a long time.

Blade-element theory is an unfortunately haughty name for something very simple, too. It's just a method where you divide a propeller or wing into several sections and apply 2D airfoil theory to each section. The 2D airfoil data is really easy to find or generate with tools like XFOIL (although post-stall data is difficult), so this is the most obvious method for real-time flight calcs. Perhaps Oleg does something more complex (like vortex-lattice or CFD) beforehand and pre-programs all the wing data into his flight models, but he's on record multiple times as saying the FMs are not based on tables.

leitmotiv
01-05-2007, 04:56 PM
Oleg has extremely impressive credentials. He must get a huge laugh out of all the make-believe aeronautical engineers who add such substance (in terms of pages) to the IL-2 forums.

Ugly_Kid
01-06-2007, 01:44 AM
Originally posted by NonWonderDog:
I'm sure it's not really analogous to X-Plane (I don't know why I said that, because FB doesn't have to analyze arbitrary user-created planes), but it certainly calculates more of the FM in real time than CFS3 did. I'm sure you've watched your target's separated wings fall to earth after a kill?


Wings falling and rolling while falling is no indication of a flight model - more likely it is an animation. Why an earth would you spend resources simulating something that is completely on the effects category. Being complexer than CFS is also no yardstick.

What we have in FM is in one part equations of motion and on the second part the aircraft itself. What tabular part or blade element theory helps you to solve is the forces resulting from the airflow at different angles and this gets noted in the equations of motion. Tabular means that for example drag polar is define as table and interpolated between the points, Oleg's "physical" means that he uses an equation for it, similar things can be said, for example, for engine power vs. altitude and speed. Oleg has gone to the records for stating that for BoB he actually considers changing some of the parts back to tabular.



I wouldn't be so quick to say it definitely doesn't use some part of blade-element theory as a base for some portion of the FM. It seems fairly likely to me, if it does any real-time wing calculations at all.


Blade-elements help x-plane to define the aerodynamic coefficients for the unknown or rather for any kind of aircraft or let's say approximate (now whether you get the real performance to the point is again another thing). You can just as well define stability derivatives and aerodynamic coefficients straight ahead and still get a working coupling between equations of motion and the local forces. Another thing is the simplications that you do to the equations i.e., apparently before 4.xx roll inertia was not considered at all.

There is only one angle of attack and yaw angle going into the "black-box" called aircraft and the resulting output that goes to the equations of motion defines how a manouver is carried out. Whether this black box is defined with tables or with physical functions/or just functions is open.



Perhaps Oleg does something more complex (like vortex-lattice or CFD) beforehand and pre-programs all the wing data into his flight models, but he's on record multiple times as saying the FMs are not based on tables.

Wishful thinking, if the aircraft FM was something "real shape" based it would not oscillate that much, eh? It is therefore more logical to conclude that the coefficients/stability derivatives are defined by hand and can be individually tuned. I am sure that a software company developing a 40EUR game is not paying annual fees of some 20000 EUR for CFD licenses to define aerodynamic characteristics of yet another aircraft, particularly as this approach is very time consuming and still not answer to it all. I don't know why people on simulation boards always bring up the CFD as an ultimate solution, not even every aircraft manufacturer can afford it in the development work.

Additionally, why would you waste so much effort, since you already know the performance that the aircraft can have. CFD is not done in the real-time (not that you said anything like that, though, but some people seem to think that the way how 3D models look like in the game has an influence on FM).

If there was something that complex in the background, don't you think we would have seen some nice pressure plots already long time ago to shut up the most persistent whiners? So far I only see sceletors of internal 3D structure of the new aircraft in BoB.

I can only repeat Oleg has himself stated that there is no such thing on the background. I believe the opposite once Oleg bothers to say as much.<div class="ev_tpc_signature">

http://www.f19vs.se/fokker_now.jpg

Aaron_GT
01-06-2007, 06:46 AM
20000 Euros for the licences, and ten times that for the computer power per plane to do the analysis!

Aaron_GT
01-06-2007, 06:46 AM
leitmotiv: check your PMs

leitmotiv
01-06-2007, 07:11 AM
Got it! Reply winging your way!

holy_one_2
01-06-2007, 09:56 AM
ofcourse about real 16:9 POV ...silece only.
I think we wil see "wide view" -fish eye.

NonWonderDog
01-06-2007, 11:15 AM
You misunderstand me, Ugly_Kid. The stalls in the game and stall recovery are dynamic, and can't possibly be scripted or table based. That doesn't mean I think it does CFD to find the lift post-stall, but it absolutely MUST keep track of angle of attack for each wing half and derive forces from that. At the very least, each wing half has lift and drag polars associated with it, and the forces are determined from speed and angle of attack. For more accuracy, the wing could be divided into two or more sections (thus allowing real-time handling of washout and MUCH more accurate control behavior post-stall) and the total force found from a simple numerical integration of each section across the wing. This method is technically a blade-element method, although it's absolutely more simplified than what's in X-Plane. Most of the calculations could be done beforehand and stored in lookup tables or curve-fits, and it would run very quickly.

And while full-viscous CFD is extremely expensive and time-consuming, there's really no need. All they'd really need to do to get ballpark figures for the flight model is run a vortex-lattice solver akin to AVL (http://web.mit.edu/drela/Public/web/avl/) (which is free under GNU). You might have enough to program in the polars for everything in just a few days of work.

It's strictly impossible not to consider roll inertia, too, if the flight model is at all based on forces. I suppose ailerons could have supplied a tabulated delta-omega instead of a moment about the roll axis in 3.00, if one more floating-point calculation would have been ruinous to performance, but that seems like more work than would have been worth it (many many data points for behavior in a spin to be at all accurate). I always thought that quote meant that ordinance now had an effect on moments of inertia or something like that. Or, and this would support my blade-element thinking, moment of inertia was once one averaged number applied to the aggregate forces from the half wing instead of being applied incrementally to the forces from each wing element.


I'm not sure what you mean when you compare tabular and physical flight models, though. I see no practical difference between an interpolated table for drag polar and a curve-fit for drag polar, as both require physical equations for interpretation into motion. I was always told (not having programmed a flight model) that a table-based flight model was one where such general things as acceleration were tabulated based on speed, altitude, and power setting. That as compared to physical flight models which compute accelerations from forces and mass.

Lodovik
01-06-2007, 12:02 PM
In BoB there will be Al pilots who are good shots, have low morale or are more aggressive etc. etc. to give much more combinations and make to feel you are facing or flying with individual humans.

Long for this waiting been Yoda has. Jubilant news this is.
Hope it'll possible to create a tracking program to follow AI pilots careers through the campaign http://forums.ubi.com/images/smilies/16x16_smiley-happy.gif<div class="ev_tpc_signature">

<<The 20 mm cannons built by the tender hands of Komsomol girls' volunteer workers corps played their deadly masurkah.
Octobriana pulled her LA-7 up and away as another of the Rodinas' enemies fell screaming to his doom.>>

Ugly_Kid
01-06-2007, 05:42 PM
Originally posted by NonWonderDog:
You misunderstand me, Ugly_Kid. The stalls in the game and stall recovery are dynamic, and can't possibly be scripted or table based. That doesn't mean I think it does CFD to find the lift post-stall, but it absolutely MUST keep track of angle of attack for each wing half and derive forces from that. At the very least, each wing half has lift and drag polars associated with it, and the forces are determined from speed and angle of attack. For more accuracy, the wing could be divided into two or more sections (thus allowing real-time handling of washout and MUCH more accurate control behavior post-stall) and the total force found from a simple numerical integration of each section across the wing. This method is technically a blade-element method, although it's absolutely more simplified than what's in X-Plane. Most of the calculations could be done beforehand and stored in lookup tables or curve-fits, and it would run very quickly.



Absolutely not and whereas you might notice that the accelerated stalls are still very much left wing dip oriented (although significantly improved after 4.xx) as they always were it all hints to a very scripted behaviour. Direction of rotation of the engine seems to be completely immaterial here.

Stalls and spins in simulators have been around for quite awhile some doing it better some doing it worse. WW II Fighters had the effect done quite well EAW less so, IL-2 with 3.xx FM was somewhere in between, now better and yet not perfect. Stalls and spins have been around for years and as such are no proof of existence of higher level of simulation behind it.

About the roll inertia you need to dig deeper into the history. It is very well possible to simulate flight without admission of it, it only affects the initial roll rate and momentum (inertia becomes first a moment with angular acceleration not with angular velocity), however from calculation costs I could not understand omitting it, it is hardly an effort. Again that bit can be read from history too.

VFS-214_Hawk
01-06-2007, 06:31 PM
Don't let Oleg fool you. He has already worked on and built microwave ovens....I have one. Every time I ope the door, it makes a bombay door open sound. Every time the timer finishes, its followed by an explosion sound.<div class="ev_tpc_signature">

---------------------------------------------------------------------------------------------------
Did I mention we need the "Slot" map for us PTO flyers?

http://f4ucorsair.com/bentwing%20Pictures/bumpersticker.jpg

Ugly_Kid
01-06-2007, 11:58 PM
Originally posted by VFS-214_Hawk:
Don't let Oleg fool you. He has already worked on and built microwave ovens....I have one. Every time I ope the door, it makes a bombay door open sound. Every time the timer finishes, its followed by an explosion sound.

After the war it was possible to buy Heinkel scooters and Messerschmitt sewing machines - so he's in a good company
http://www.cybermotorcycle.com/gallery/heinkel/images/Heinkel%20Tourist%201962%20India.jpg <div class="ev_tpc_signature">

http://www.f19vs.se/fokker_now.jpg

Ugly_Kid
01-07-2007, 12:17 AM
Originally posted by NonWonderDog:
And while full-viscous CFD is extremely expensive and time-consuming, there's really no need. All they'd really need to do to get ballpark figures for the flight model is run a vortex-lattice solver akin to AVL (http://web.mit.edu/drela/Public/web/avl/) (which is free under GNU). You might have enough to program in the polars for everything in just a few days of work.


Again predicting the linear part of the flight envelope with aerodynamic coefficients can be done quite easily with hand calculations since the performance aspects are known. Vortex-lattice is limited to thin lifting surfaces and takes you only that far - additionally, it won't give you a bit when approaching a stall or going into a violent manouver like spin. So again what's the point?

Also how you define the aerodynamic coefficients, even if you had a wind tunnel for it, has little to do with what the actual FM is using. You would not need to do such a work beforehands in preparation if the code worked directly by using blade elements, now would you. In this case the code would figure it all out for you, you would just need to type in the shape. Would be quite easy to implement a 3rd party aircraft modeler for it, too. None, which we have.

Again, there is no indication of a more complex simulation in IL-2 except that introdcution bit from the interview. And again it is in direct controverse with Oleg's own statements from the past. I think you are just making too much out of it...<div class="ev_tpc_signature">

http://www.f19vs.se/fokker_now.jpg

WWMaxGunz
01-07-2007, 12:51 AM
Originally posted by Lodovik:
Hope it'll possible to create a tracking program to follow AI pilots careers through the campaign http://forums.ubi.com/images/smilies/16x16_smiley-happy.gif

Like AI pilot stays and changes characteristics as analog to learning from experience?
If so then perhaps visits from aces sent around teaching should also improve AI pilots,
I have read that in Russia that did happen from I-16 onward. Perhaps also the presence of ace
pilots in your own group would make quicker good changes of the surviving AI to simulate the
passing of information and training of pilots done between game campaign missions? It would
be a gas!

msalama
01-07-2007, 12:54 AM
And again it is in direct controverse with Oleg's own statements from the past.

What kinda past we talking 'bout? The v3.x -> v4.x upgrade might've changed that, too...<div class="ev_tpc_signature">

-----------------------------------------------------------------------------------------------------

Hippies FTW!

starfighter1
01-07-2007, 02:29 AM
1)'scalable AI' 2) 'more combination'
3) 'real time weather system across the whole map' (real time is defined very clear !? http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif
4) Korea Scenery http://forums.ubi.com/images/smilies/clap.gif

1-2 things which are running in some combatsims
like updatet Falcon4 AF
or 3) in Zivisims like X-Plane 8.5/8.6 in best manner http://forums.ubi.com/images/smilies/10.gif

But..what about an intelligent campagne warfield !?

The different between X-Plane physics engine and other sims like FS or mixed FM basics ..have a look here:
interview with Austin Meyer..

http://macologist.org/showthread.php?t=1676

I'll put it another way:

In Microsoft Flight Simulator, you've gone and told the program how you think your plane should fly, and in X-Plane, you tell the program what shape you are flying, and it tells you how it really flies.' (Austin Meyer)

I guess the same to IL-2/FB and future of SOW BoB basic FM

more here about blade theory and new features in sim technology basics:

http://www.ae.su.oz.au/aero/propeller/prop1.html

http://aerodyn.org/Rotors/models.html

http://www.ilr.tu-berlin.de/WKA/technik/free.wake.html

http://uplink.space.com/printthread.php?Cat=&Board=busi...ain=454270&type=post (http://uplink.space.com/printthread.php?Cat=&Board=businesstech&main=454270&type=post)


by the way:
Lead Pursuit(F4AF/news at E3) and XSI 'Fighter OPS'(based on parts of X-Plane engine) are in developing new combatsims and even a high advanced pilot trainee compare to USAF procedere ..by first steps in T-6 Texan... (OK some more like Eagle Dynamics/LockOn/Black Shark)

2007/08 will become a interest scenery of combatsims on high level http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Ugly_Kid
01-07-2007, 03:34 AM
Originally posted by msalama:
What kinda past we talking 'bout? The v3.x -> v4.x upgrade might've changed that, too...

Good question, Oleg's statement comes from earlier days.

However, even when change to 4 did a lot, I am still sure that they did not break all the existing odd 100 planes into blade elements and reprogram the whole aircraft part of the FM from scratch for a free patch. I am pretty sure of this.

And what for? The normal flight envelope part with oscillations and dynamic simulation can be done adequtely with stability derivatives (one for whole aircraft). Blade element theory gets first interesting once you don't know anything from aircraft except the shape. Here we have exact known performance to match. It is easier to tune changing one or two coefficient, with blade element approach you'd need to change the aircraft shape. I doubt it would bring essentially more.

If there had been such a substantiatial change from 3.x to 4.x we would have noticed quite a few aircraft not matching climb or top speeds or that kind. When FB 1.0 came out there were quite a few aircraft where performance was seriously off. This was also in some extent case with succesive add-ons AEP or PF. What was changed was some simplifications for equations of motion part were removed, one of which was accounting for turning prop disk. BoB might bring more of this kind of changes/perfections to FM.

It is also questionable whether blade element theory does a better prediction in error situations like spin or whether you need to "help" it a bit. In many aircraft spin can result in rudder flow being shadowed by separated flow and remain ineffective. Similarly , a t-tail can be shadowed by main wing in a stall. Does one element know how and what kind of flow comes from the other elements to it. I believe that in these situations even x-plane comes to its limitations, but this is speculation from my part.

We have too similar behaviour of all planes all featuring similar spin, similar speed of rotation similar attitude. There are dramatic differences in RL here. All planes AFAIK also change from a "normal" spin to flatspin if you botch it and recovery is also similar. Defenately not what is common for all the real aircraft. This has a serious scripted odour for me.

I stick to what Oleg has said - it would be, however, nice if he or actually the guy doing the FM would put more light into the matter.<div class="ev_tpc_signature">

http://www.f19vs.se/fokker_now.jpg

starfighter1
01-07-2007, 04:27 AM
UglyKid
as I got a response from A.M.last year..
the only main limitation is the OS problem in 'real time' at this PC-Sim X-Plane

but by running a real time OS Linux System(workstation or Cluster by Red Had for example)within delaytime of under 20 ms you get fine results. But in this case X-Plane must and can be updatet or recoded http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif ..100TSD bucks and you have it http://forums.ubi.com/images/smilies/10.gif (example: running a high dynamic movement simulator platform aproaved by FAA)
at normal or highend PC and with gamers equipment nowadays http://forums.ubi.com/groupee_common/emoticons/icon_rolleyes.gif ...low progress(very few)to feel better as a combat pilot http://forums.ubi.com/groupee_common/emoticons/icon_frown.gif

I'm interest how XSI will mix the FM (part based on X-plane engine) in new Fighter Ops even at the prop trainers

and BoB...many questions and few answers from the developer http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif




Originally posted by Ugly_Kid:
.....in these situations even x-plane comes to its limitations....

It is also questionable whether blade element theory does a better prediction in error situations like spin or whether you need to "help" it a bit. In many aircraft spin can result in rudder flow being shadowed by separated flow and remain ineffective. Similarly , a t-tail can be shadowed by main wing in a stall. Does one element know how and what kind of flow comes from the other elements. I believe that in these situations even x-plane comes to its limitations, but this is speculation from my part.

We have too similar behaviour of all planes all featuring similar spin, similar speed of rotation similar attitude. There are dramatic differences in RL here. All planes AFAIK also change from a "normal" spin to flatspin if you botch it and recovery is also similar. Defenately not what is common for all the real aircraft. This has a serious scipted odour for me.

I stick to what Oleg has said - it would be, however, nice if he or actually the guy doing the FM would put more light into the matter.

NonWonderDog
01-07-2007, 02:46 PM
Hmmm...

I really don't think Oleg's flight model just uses a single set of constant stability derivatives with scripted stall at a pre-set stall angle of attack. That's been around for a long long time, and you can find exactly that kind of model for free with Flightgear. The fidelity of it is perfectly potty, if you want my opinion, and I'm almost certain Oleg has said stalls were unscripted (since FB, at least).

Oleg takes great pains to separate himself from the competition, and will readily boast about the many new technologies he introduced with his simulator. Perhaps these new technologies are just methods of plotting non-linear stability derivatives from models, and the real-time flight model really is bog standard. Maybe what sets his flight model apart is the use of non-linear stability and control derivatives for post-stall modeling. I don't know, though, that doesn't seem very revolutionary.

I'd expect that he models the wing, tail, and fuselage as separate bodies with separate parameters. It's really too much to believe that the flight model for when you lose a wing or tailplane or control surface is completely scripted. I don't know if this could be considered revolutionary, but it could possibly be mistaken for a blade-element method by someone who didn't know any better when it's explained in Olegish. (I guess you could actually call it a blade-element method with one element per wing, but then you're just being silly.)

Similarities in spins can't be pinned on the engine, though. It's perfectly possible the engine is capable of making planes that behave differently in spins (I think the Macchi spins quite well, for instance), but the data fed into it isn't accurate enough to show it. Where's he going to get detailed stability data from, anyway? I don't know how developed control theory was in the 1930s and '40s, but I've never seen a NACA chart with lists of stability and control coefficients. Either Oleg pulls these things out of thin air or he's got some kind of computer model he runs beforehand. Perhaps the computer model, for lack of data, generates parameters that lead to similar stalls for all the planes.

Ugly_Kid
01-07-2007, 11:53 PM
Originally posted by NonWonderDog:
Oleg takes great pains to separate himself from the competition, and will readily boast about the many new technologies he introduced with his simulator.

What competition? CFS is the only one that comes into my mind and that AFAIK relies to tabular approach. His "new technologies" what he has concretely stated is using functions instead of tables.



Where's he going to get detailed stability data from, anyway? I don't know how developed control theory was in the 1930s and '40s, but I've never seen a NACA chart with lists of stability and control coefficients. Either Oleg pulls these things out of thin air or he's got some kind of computer model he runs beforehand.

The maths even behind CFD have been around for some hundred years. Quite a lot of shoulder work was done in 30s, and reading from biographies such as Lerche the stability and control were systematically tested in Rechlin and others. I don't understand why people think that world did not exist before CFD and supercomputers. Books for preliminary design, i.e. Roskam or Torenbeek contain some helpful methods for definition, I am positive Russians have their own ones to add. Etkin's dynamics of flight deals with control issue and is pretty helpful in definition of the control effectivity with some of its charts. ESDUs brings a great deal of help. And yes some values they defenately pull out of sleeve, the way how one aircraft is a subject to torque more than another and completely opposite to the known behaviour is one that comes into my mind. Control efficiency too, some most famous aircraft get tuned (the eternal roll rate discussions) but rest are simply ignored. Particularly this bit indicates that the real shapes do not play role here.<div class="ev_tpc_signature">

http://www.f19vs.se/fokker_now.jpg

starfighter1
01-10-2007, 06:01 AM
@Ugly_Kid

this is http://www.ccur.com/default.aspb http://news.ccur.com/isd_default.asp?h=1
http://www.ccur.com/isd_aboutus.asp?a=1
http://news.ccur.com/corp_news_pressrelease.asp?pressreleaseid=481
the leading company running special software in 'real time' Linux(special Concurrent Red Hat or Suse)
and even rendering by this company http://www.multigen.com/ or partners http://www.ccur.com/isd_partners.asp?p=1
on industrial standard (Open GL) http://forums.ubi.com/images/smilies/10.gif

gamers O.M. world is gamer world 'simulation world is simulation world ' http://forums.ubi.com/images/smilies/touche.gif

basic FM aerophysics in this 'Videosimulation-Game ' is a simplified calculation either in function or tables or both .... I can't see any real time response in aerpdynamic procedere
at this basic game engine ...

next point: aerodynamic floating in real time weather http://forums.ubi.com/groupee_common/emoticons/icon_cool.gif

Austin Meyer's and Randy Witt's speaking about MS FS is same as in IL-2/FB-BoB
http://macologist.org/showthread.php?t=1676
......
2) Tell us about the physics engine at the heart of X-plane. Specifically:

a) We notice that what is special about X-Plane is that blade element theory is at the heart of the physics engine. What are the strengths of blade element theory as implemented in X-Plane?

Austin: Well the main benefit is that it's designed for use by engineers. Blade element theory breaks the aircraft down to its various elements: stabilizers, propeller, etc., and finds the forces on each piece, then adds them all up, then divides by the mass and moment of the aircraft to get the acceleration.

The realism provided by blade element theory is what makes it good as a tool for engineers. For the pilots, it's beneficial because it flies like a real airplane would do. For non pilots, well, presumably they went and bought a flight sim to see what it's really like to fly an airplane, and not to fly spaceship cameras in a shooting game, so if you want to know what it's like to fly an airplane, that's what X-Plane delivers for you. It's an incredibly accurate representation of how an airplane will respond to the pilot.

Now the hobbyist can design an airplane, and simulate it in X-plane to see how it flies. You don't need to be an engineer to try this! You can think what kind of shape you'd like the plane to be, drop it into X-Plane, and it'll show you how it flies. Now if you're not an engineer, you might come out with a design that does not fly, or does not fly well, but X-Plane isn't going to give you some preconceived notion of how your plane is supposed to fly. It dutifully finds the forces on each piece of your plane, and will tell the hobbyist how his plane will really fly.

b) Can you contrast this with MS Flight Simulator's Table-based method?

Randy: Well MS Flight Simulator is programmed with the use of these huge tables where it looks up all the variables of what the pilot is doing, and what the plane is doing, and it finds the movement of the plane in the next frame. In other words, someone manually put in all these numbers, and if there's an error, then it's right there in the table. X-Plane doesn't have any tables. It looks at your airplane, and actually calculates how it responds.

I'll put it another way:

In Microsoft Flight Simulator, you've gone and told the program how you think your plane should fly, and in X-Plane, you tell the program what shape you are flying, and it tells you how it really flies.

c) There are supercomputing clusters that model fluid dynamics. What approximations are involved in X-Plane's blade element approach to run on a desktop PC?

Austin: That's a very good question. Computational fluid dynamics (CFD) is a science whereby we find how air flows around complex shapes. In CFD, computing one moment in time might take a computer, a supercomputer in fact, running for a week. Now the desktop PC needs to be running 40 frames per second, not 1 frame per week!

Blade element theory is different from CFD. In CFD, every bit of the airplane affects airflow, and every bit of the airflow affects every other bit of the airflow, and every bit of the airplane. Now we're talking say modeling say a million parcels of air, and a million pieces of the plane. Now say you consider a single packet of air, just consider how long it would take to compute its effect on all the other 999,999 pieces.

It's absolutely ridiculous, it might take a million cycles of time to arrive at the final solution, just to compute the propagation effects of the pressure of one cell moving from packet to packet to make it to one on the far end-- think of throwing a rock into a lake and the ripples it generates as each point of water transfers the pressure to the next piece.

And those pieces can affect the airplane, so we're talking now a million cubed packets worth of mass. It just gets wild, wild, just wild. When you do CFD you don't even want to think about doing a total output dump, because if you were thinking of dumping that amount of data, you're talking about gigabytes of data no human could every possibly look through. So here, we're talking about diagnostics, which is a science in itself. We have to design a diagnostic to find to find that one little byte of mass that went wrong. It's incredibly difficult.

So blade element theory is different. BET says: let's break down the aircraft into 200 pieces, instead of a million. And rather than finding a million parcels of air around the airplane, it finds 200. And instead of looking at every element's pressure field distribution into all the other elements, we have each element finding its own force. But BET can also have elements affecting each other, for example in downwash and propwash the elements have forces affecting one another. But it's not like this circular iterating logic of computational fluid dynamics, where you might need 100,000 iterations to get one answer, but blade element theory needs just one iteration.

Now these are both old, old theories. Blade element theory is really old, and computational fluid dynamics, well you can look it up, but from what I remember it was designed by the royal engineer for the King of England back in the days of Christopher Columbus. The King wanted to have faster ships for his fleet, now this is back in the 1300s or 1400s! And that engineer starting thinking about the hull's movement through water. Even though the theory could exist back then, it couldn't actually be applied to come up with a solution that you could use. But the theory is old, and blade element theory is also old. You can read more about BET here.

d) For the lay-reader, give a specific example (or examples) of how X-Plane reflects the way a real plane would respond, and where MSFS would not.

Austin: Well the instrument in Microsoft flight simulator, I see that they move in notchy steps. I think the instrument jumps as the sim looks up elements in the table. So whereas Microsoft is more like watching stop motion animation, X-Plane gives you completely fluid motion. It's like a real instrument in an airplane.

The other thing is ground handling. I recently flew a tail-dragger in Microsoft Flight Simulator... now the tail-dragger is an airplane where the 3rd wheel is in the back of the plane, and not the front of it... and when that airplane intersected with the ground, it did not behave like a real airplane!

It would spin around in these arbitrary angles, but not with the sensation that it was a real plane with mass, momentum, and real landing gear touching real ground. In X-Plane you can actually feel the instability of the airplane kick in as it tries to spin around. We have the landing gear actually simulated as a damped spring, and as it intersects the ground, it's actually computed.

You can totally tell the difference, and I mean it's just the tip of the iceberg in X-Plane! The same goes for the aerodynamic handling. It's more subtle, but you'll definitely notice it, even if you're not a pilot. It'll fly like a real airplane.

...



iv) How good is the response of the instrument panel in comparison to real life, as well as Microsoft Flight Simulator?

Randy: The instruments portray the way the aircraft is responding in the simulator's flight model. I don't know that there is any inaccuracy in what the instruments are showing - or any real difference between MSFS and us in this department. I think what you're really asking is "how realistic is it, as compared w/ Microsoft Flight Simulator?". The answer to this is that we have a much more realistic and fluid flight model than any other flight simulator designed for the personal computer. This is due to our method (element blade theory) and provides, we estimate, a 30% improvement in the fluidity and realism of the simulator.

The US Army (I think it was the Army, check me on this Austin) recently conducted an analysis of X-Plane and "certified" that our sim was accurate to a 5% degree of accuracy in the realism of the flight model and to a 10% degree of accuracy in the overall feel and response of the sim. (Don't ask me how you determine a percentage accuracy in the "feel" of a sim. I don't know.)

Austin: Randy, I am NOT sure it was the Army, no... but independent users have done papers that they have presented at aviation conferences that show the sim is accurate to within 10% in handling. You measure this by measuring the pitch and roll and heading rates of rotation of the craft, and how fast the planes come back to their original heading when displaced, and how many times they overshoot before coming to rest at their trimmed attitude.

...

an update on 64 Bit real time X-Plane to run on FAA certified X-Plane to run on CC The iHawk?
http://www.ccur.com/isd_solutions_redhawklinux.asp?o9=1
is possible http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

so long about 'real time' on PC-based Games and real time Simulation on industrial standards

I guess O.M. and most of the users mixed up a lot in this field but even O.M. should now it http://forums.ubi.com/groupee_common/emoticons/icon_frown.gif
but often he talks a lot of 'nonsens' at gamers interviews as a nice markenting and promotion gag(tall)story http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif


[QUOTE]Originally posted by Ugly_Kid:
His "new technologies" what he has concretely stated is using functions instead of tables.

[QUOTE]
Where's he going to get detailed stability data from, anyway?

starfighter1
01-11-2007, 03:19 AM
this is free...

http://www.flightgear.org/

based on 1996 established http://jsbsim.sourceforge.net/
(see documents )
http://sourceforge.net/projects/jsbsim

I'm not shure the O.M. dev team use several basics of this open source C++
selling as an encodet Warbirdsim http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

anyway ..many things even at game combatsim software were copied in the past,
updatet or extended http://forums.ubi.com/images/smilies/10.gif








Originally posted by Aaron_GT:
20000 Euros for the licences, and ten times that for the computer power per plane to do the analysis!