Support of different wavelengthes in ArmA2

[S!fkaIaC 10]

I am wondering if the ArmA 2 engine will support the total independent simulation of different kinds of radiation.

Which means Source1 could represent visible light, Source2 a specific IR band below the light, Source3 a Radar Transmitter of a specific band.

Each Sensor must have a kind of sensivity / damping value to define how sensitive a certain sensor is (some sensors covering the radiation of more than one source).

Especially for Radar it would lead into a much more realistic simulation.

[PhilippRauch 0]

Somehow nice but i think thats reserved for serious games, aka VBS...

[S!fkaIaC 10]

I guess it even does not exist like I described in VBS2 atm.

[M.Andersson(SWE) 4]

Somehow nice but i think thats reserved for serious games, aka VBS...

Why??!

[De_little_Bubi 1]

cause we would get flir if there are simulated ir-rays

[TeRp 1]

;1272469']Why??!

That's a good question.

It's absolutely easy to port something from ArmA to VBS2, yet it seems to be almost impossible to do it the other way..

[sparks50 0]

Yes. I guess the government pays a bit better than us.

Edited April 20, 2009 by sparks50

[PhilippRauch 0]

Yes. I guess the government pays a bit better than us.

Thats what i meant... but i sureley wouldnt mind if its possible/included in ArmA2 or ArmAx ... :D

Proper simulation of wavelenghts is a different beast than creating a nice battlefield game/sim... in a game it would be anyway some kind of workaround/emulation and hence many hardcore simmers wouldnt be happy either, so i would rather have more units, worlds and missions in my game than that, but again if would be easy to implement than im all for it!

[S!fkaIaC 10]

The reason why I am asking for this feature is that we use in ArmA more and more hight-tech-weapons, but by the nature of the ArmA 1 engine they are very odd simulated (well, god for the time when it was coded, bad if you compare with reality). From that POV you can simulate with the ArmA 1 engine another WWII environment.

Since many modern weapon characteristics are classified it is difficult to make a simple radar guided rocket anyway. Since sensor charackteristics, tracking capabilities etc are to complex and as I said, unknown, one should be very careful creating "Ãœberwaffen". Earliest HARM rockets had a hit ratio of 25% only in reality, most Flight-SIMs simulated much higher rates.

But if one is serious with the attempt to simulate modern weapons, it starts with the proper simulation of waves and their reflection, deflection etc.

If the propagation is correctly simulated, it is already a big deal, no radar signals anymore if the target is sharp behind a mountain e.t.c. Or your IR scanner is blinded if your target is in the middle of a burning town with lots of burning houses and burning vehicles.

[PhilippRauch 0]

Yeah, i can second that, but do we need a 'Simulation' or is just haveing the proper 'feel' of it ingame? Thats what i mean with workarounds/emulation...

Many Values are simply not available to public, which doesnt mean they are classified, just subscribe to Janes if you got the xxxx$ and you will have most non-classified data right away. But do we need proper 'scientific' simulation inside a gameengine??

I dont think so, the difference in faking and feeling it and 'proper' scientific representation wont be really noticable anyway, if you look at the values/numbers... sure you will see a difference but we wanna play not write nice articles with big statistics and other numbergames, or do we?

I really, REALLY would love to have a more 'proper' inbuild IR for sure, NVS i like already, but for example my Javelin addon is just using whats inside the engine and fakes 'proper' behaviour, like the missile is using some scripts that guide it to that target which position is always known by the engine, which the real javelin missile simply cannot, the RL Javelin uses a intricate visual comparing 'thingsystem' which will keep it locked on selected target... i just fake that behaviour inside my scripts to make it handle/feel/look like the RL would behave/do/fly ... if we would go for 'proper' scientific simulation of something soo complicated as rays and behaviour of wavelenghts then we would need supercomputerclusters for simply doing a CTF game!?

I would rather have some kind of 'layer' system for textures, where with some globalvariable 'IRON=true' will switch all non_ir textures simply off or set their alpha to 0 so only the IR textures will be shown, additionally some changes to the lighting engine/shaders (like the green for NVS) and we should have nice 'representation' of 'proper' IR systems...

or something like that... :)

[max power 21]

Light and other simulated effects are not the same thing in 3d. Light in realtime 3d is not rays or waves. Therefore, the light simulation could not be used to simulate radar. In VBS 2, I'm guessing that their flir implementation is a lot like the NVG effects in ArmA. I'd imagine that the FLIR effects modify the properties of the camera and use custom textures and materials to provide the glowy effect. If you wanted different strobes and stuff to be visible by different cameras, I'm sure that's doable, perhaps with something as simple as flags. Radar does not need a camera or anything at all so there would be little point in rendering it using the graphics engine. It would be better served I think using a totally different simulation altogether.

[PhilippRauch 0]

for Radar check SPON Radar or some of the stuff in mandos missile suite... nice implementation of radar, and since the engine ALWAYS knows where any object/vehicle/unit is at, there is no problem 'simulating' different kinds of radar systems (2d or 3d radar systems are already possible)...

[Spokesperson 0]

You already have a good shader based simulation of light. Different light conditions, NV is the IR simulation and so on. Radar is already in game since OFP, but there are no rays or anything like that, just a slimmed down result based version of it. Real simulation would probably require ray tracing at huge distances and not render much different results from what there is already, but at a much higher performance cost.

But ray tracing would be great for things you can see with your eyes. New kind of reflections, mirrors, shadows etc. But that's DirectX 11.

Edited April 20, 2009 by Spokesperson

[PhilippRauch 0]

Well, according to some interviews BIS said they dropped dx11 because it had no significant advantage over dx10 and a smaller userbase (which was a good decision, now i ONLY ;) need to upgrade MB and CPU *g).

And reflections are ingame (ArmA 2 !) , just check the developers blog and the entry with/about the new armoury, there you can see the unit being reflected in a pond of water while doing a crosscountry course or so...

[max power 21]

dx11? ArmA and ArmA 2 are in dx 9.

[Spokesperson 0]

DX11 is not ready yet I think, and there's no hardware for it. DX10 is "used" by some games. But the reflections available with DX9-10 tech is in no way comparable to that of real time ray tracing.

[S!fkaIaC 10]

I would rather have some kind of 'layer' system for textures, where with some globalvariable 'IRON=true' will switch all non_ir textures simply off or set their alpha to 0 so only the IR textures will be shown, additionally some changes to the lighting engine/shaders (like the green for NVS) and we should have nice 'representation' of 'proper' IR systems...

This is NOT the way. To many limitations if you do it this way.

It might work in a very odd way for IR partly, but having a airport radar that should not be sensitive for waves used by target radar and visa versa, IR, light, ...in the same environment - I have doubts that it works.

The sensitivity-behavior of sensors might be simplified in the solution I prefer, but for the propagation you can use more or less the same model as for visible light but with different parameters.

And you need to need to create new parameters like the presence time (e.g. absolute start time + perid e.t.c) of the transmitter synced with mission time to simulate a modulated radar in sync for all clients. Also signal strength and horizontal and vertical focus should be changeable.

[PhilippRauch 0]

Where is the limitation exactly? No 'REAL' IR beams being used? For what? It just needs to handle/look and more importantely 'feel' like its viewed through some IR system.. sure it would be nice to have some influence of enginestate (heat generated) on the textures and such but then we will also need ammbient temperature, which we just dont have in a/this game right now, hell who cares, if i can use it and have the same advantage and shortcomings of actual fielded IR equipment then its all right for me.

There is just no point in using up resources for raytracing from IR all the way up to UV or Microwaves or ... hell lets do some gamma ray raytracing or quarks... :p !

You can already program all kinds of different behaviours/systems like sensors and such inside ArmA1 it all depends on your abilities and phantasy! And funky Radar systems are already done and can even be improved, so lets start the work!

But they just dont use rays or anything like that since its NOT NECESSARY! The gameengine has all the data we need, no need for tracing gazillions of rays of IR and microwaves...

What needs to be done are interfaces/scripts that interpret the data to some 'edible' or usable format like a radarscreen or some warhead... all has been done in ArmA and even in OFP (Like RWR, tracking missiles)! And i am not talking about the 'radar' included in the vanilla arma...

To clarify my point further:

There is simply NO use or advantage in using raytracing for such things since they can be had at a lesser CPU cost already and with some clever use of engine abilities even more so.

[S!fkaIaC 10]

Use case:

Mission+addons that should support:

- 2 radar wavelenghtes

- IR

- normal light

- multiple radar sources to illuminate the target(s)

Now my questions:

- how you make sure with your approach that a virtual sensor for Radar wave type 2 (which should be visualised on a radar screen in an AA control vehicle) does not react on virtually transmitted waves of IR sources (fire, heat, ..) ; radar waves Type 1 and normal virtual light sources (Sun, Moon, lamp pole...)?

- how you sum up the signal at sensor if the illumination of a plane received by a target tracking radar by 2 ore more radar transmitters should be simulated?

- How you simulate that the Sun is transmitting in Light band and IR band, but not (at significant level above noise) at radar band(s)?

- how you simulate short pulses of radar so that it can not be received by spinning dishes if they do not point to that direction?

- how do you simulate reflection/deflection of all those waves by water/rocks causing noise that disturbes/blinds some sensors?

Now tell me that all that is not needed, then I say without that it is not serious to model any weapon system after 1955 with that engine.

[vektorboson 8]

@S!fkaIaC

I think you are thinking too much in physical terms, whereas in games, even simulation games, it's all about faking stuff to look or handle like 'reality'.

In games there is no such thing as wavelength simulated; for 'light', games use RGB-triplets instead of a wavelength spectrum.

As for IR/thermal imaging, maybe it will be possible to implement them, if it will indeed be possible to write custom shaders.

As for Radar: Well, since it's not a 'visual' thing, it's possible to script it. This indeed requires to raycast into the world; for vehicles it's easy to do. You check if the vehicle is in the radar cone, you use an 'abstract' RCS-value (radar cross section) to calculate the reflection 'power' and display the radar return accordingly. If the vehicle in the cone has a ESM-sensor, well, you notify that vehicle/pilot of the radar spike.

Also, it's easy to simulate two different radar wavelengths, since it would be just querying if the sensor supports wavelength 1 or wavelength 2.

This is mostly very simple vector math and using the right programming structures.

As for ground clutter: Well, it isn't much different than the above, though I see performance problems. Also, how to draw that (ground) radar screen?

[S!fkaIaC 10]

Thanks for the clarification vectorboson. My problem is not that I think to much in physical terms, it is rather that I think only in physical terms :D

It is a kind of stomach feeling that solutions like you and PhilippRauch desctibed might do the job for a "raw" simulation, but if it comes to some tiny differences which result in the end into 0% or 100%, no grey between (missed or hit), it matters how accurate a certain simulation is.

Of course you do not virtually transmit waves like endless vectors of endless amount of bullets with their source in the virtual transmitter.

But working with that much simplified models you described makes it almost impossible do model the sometimes huge differences between all the radar- and rocket guidance systems in the world. They would work all the same way.

[vektorboson 8]

But working with that much simplified models you described makes it almost impossible do model the sometimes huge differences between all the radar- and rocket guidance systems in the world. They would work all the same way.

Of course you need to simplify the models, since you wouldn't be able to have a realtime simulation.

I'll take radar reflectance of airplanes as an example: You could indeed calculate how the surface of the airplane reflects radar waves (of specific wavelength) and you could also take material reflectance/absorption coefficients into account. But you could also abstract those values into a RCS-matrix which is calculated/set for let's say six directions and thus speed up the calculations.

Now the big questions are:

- Are there extreme differences between the complex simulation and the simple simulation?

- Are those extreme differences probable?

- How much computing time does the complex simulation take away? Is the game still playable if the complex simulation is used under 'normal' playing conditions?

What I want to say is, that what a physician considers a raw simulation might still be a pretty complex and realistic simulation, but with bigger error bars, and with the possibility to run realtime as opposed to physically correct simulation which runs rather offline.

[gonk 0]

looking too deep.. this can be done pretty easily but requires more time from the modellers.. All they would do is add another LOD to each model ... eg an Thermal LOD.

[vektorboson 8]

looking too deep.. this can be done pretty easily but requires more time from the modellers.. All they would do is add another LOD to each model ... eg an Thermal LOD.

Thermal imaging is easier than an extra LOD; it'd be only an extra material setting and an extra texture. That texture would take only little memory, since it doesn't need to be high-res and it only needs one component (intensity), or two (intensity and alpha).

[PhilippRauch 0]

Thanks vektorboson for going into more technical details, im just an humble artist with more taste for prettieness than braincells:803: for physics or coding.

You clarified exactly as i meant it ... :D