NAVFLIR hotspot detector - a proposal

[amalahama]

After watching that video I can see that the sensor is anything but trustworthy.

 

Exactly. Even the random-placed chevrons would not be a proper solution if always the real vehicles are spotted by the system, which watching the video seem quite unlikely in RL. I would go for leaving the feature out until DCS engine provides the technology to look for "hot pixels" in the screen with low CPU impact. Otherwise, it's like flying with labels...

 

Regards!

[shagrat]

Exactly. Even the random-placed chevrons would not be a proper solution if always the real vehicles are spotted by the system, which watching the video seem quite unlikely in RL. I would go for leaving the feature out until DCS engine provides the technology to look for "hot pixels" in the screen with low CPU impact. Otherwise, it's like flying with labels...

 

Regards!

I would prefer a simple randomization, that omits to show ALL vehicles and creates a couple false positives. The final result can be pretty close to the real thing, without requiring loads of CPU time.

 

Keep in mind, in real life a stone heated up by the sun or a small body of water hidden between bushes, dark trees or light sandstones may provide temperature contrasts to a FLIR to trigger false positives, so a "random" hit plus not showing every vehicle in range of the FLIR would be a good balance between NOT having a vital assistance system and having a 100% failure proof "enemy pointer".

 

In real life the false positives are corrected after some 5 sec so it is a very important help on putting the ARBS sensors (DMT) on the target location, refine the aim and get a one pass bombing as part of a strike or BAI mission.

Without the NavFLIR helping to identify the targets these missions will get very difficult .

 

We need to simulate a balance between unrealistic perfection and unrealistic absense of the system...

[amalahama]

I would prefer a simple randomization, that omits to show ALL vehicles and creates a couple false positives. The final result can be pretty close to the real thing, without requiring loads of CPU time.

 

Keep in mind, in real life a stone heated up by the sun or a small body of water hidden between bushes, dark trees or light sandstones may provide temperature contrasts to a FLIR to trigger false positives, so a "random" hit plus not showing every vehicle in range of the FLIR would be a good balance between NOT having a vital assistance system and having a 100% failure proof "enemy pointer".

 

In real life the false positives are corrected after some 5 sec so it is a very important help on putting the ARBS sensors (DMT) on the target location, refine the aim and get a one pass bombing as part of a strike or BAI mission.

Without the NavFLIR helping to identify the targets these missions will get very difficult .

 

We need to simulate a balance between unrealistic perfection and unrealistic absense of the system...

 

I'm not a fan of "randomness", but I'm totally onboard for clever simplifications. If at least the position of static structures like buildings, roads or power towers were stored in the scenery list and available through the SDK, then a satisfying solution could be done, unfortunately it's not the case. Plus I'm not that sure of the real usefulness of the system, taking into consideration the wide FOV it covers with the low res of the sensor.

 

ED will fix the scenery list and complete it with all the static objects at some point. But I'm pretty sure that when the random script is done, it won't be redone again.

 

Regards!

[shagrat]

I'm not a fan of "randomness", but I'm totally onboard for clever simplifications. If at least the position of static structures like buildings, roads or power towers were stored in the scenery list and available through the SDK, then a satisfying solution could be done, unfortunately it's not the case. Plus I'm not that sure of the real usefulness of the system, taking into consideration the wide FOV it covers with the low res of the sensor.

 

ED will fix the scenery list and complete it with all the static objects at some point. But I'm pretty sure that when the random script is done, it won't be redone again.

 

Regards!

The list of objects is no problem, at all. You can iterate through all static objects, including buildings.

The problem is your CPU computing power per cycle... As with a lot of other things in DCS you need to balance "realism" vs. resources available.

 

Thus it is simply a matter of "computing" real life behaviour to 100% and accept a frame rate of 1-2 fps with a hi-end PC, or "simulating" how the system works and appears on the HUD?

 

The false positives the real life system shows are corrected after a couple seconds, usually faster than you would slew a sensor in them and verify them. (Look at the videos).

 

So if we get a little randomized "clutter" that behaves like the false positives in real life, it would at least prevent us from getting an unrealistic 100% accurate magic FLIR that only recognizes active vehicles.

 

And THAT is what is the current state!

 

So before we get NO false positives and only 100% accurate targets, I prefer ANY randomized, roll-the-dice or whatever solution to at least "simulate" false positives!

[GazAce]

I'm kinda happy with what I've got

 

I also play ARMA III in s/p and do use the hot spot detector sometimes but I could prob get away with not using it at all. I think the IR targeting system in 2.1 and 1.5 is pretty good, realistic or not it works nicely and at the risk of performance or code rewriting issues, why bother? I'm really happy with what's available atm. Just saying :) And I wonder how that system if were implemented, would react in conditions of snow or extreme cold?

Peace, out :thumbup::pilotfly:

Edited October 31, 2017 by GazAce

[shagrat]

I also play ARMA III in s/p and do use the hot spot detector sometimes but I could prob get away with not using it at all. I think the IR targeting system in 2.1 and 1.5 is pretty good, realistic or not it works nicely and at the risk of performance or code writing issues, why fiddle? I'm really happy with what's available atm. Just saying :)

 

The NAVFLIR has a "Autodetection" mode that points out targets with "v" shaped markers, unlike IR-video where you have to find the targets yourself.

 

In real life this systems help to point at "potential" targets so you can point the sensor to it, but in real life it also is not perfect giving you false targets as well as valid ones. Ultimately the pilot needs to determine the target from a false positive.

 

It still is an important help to identify the exact spot of say a pre-briefed artillery postion when running in low level at 450kts where the waypoint is near the suspected position and the "v" markers help to find the individual targets.

 

This video shows the system at work

 

[GazAce]

The NAVFLIR has a "Autodetection" mode that points out targets with "v" shaped markers, unlike IR-video where you have to find the targets yourself.

 

In real life this systems help to point at "potential" targets so you can point the sensor to it, but in real life it also is not perfect giving you false targets as well as valid ones. Ultimately the pilot needs to determine the target from a false positive.

 

It still is an important help to identify the exact spot of say a pre-briefed artillery postion when running in low level at 450kts where the waypoint is near the suspected position and the "v" markers help to find the individual targets.

 

This video shows the system at work

 

 

 

 

Thanks kindly Shaggers, appreciate that mate

[DragonShadow]

One interesting thing to do while simplifying the simulation of this and providing false positives, might be to just consider the color of the objects and terrain being drawn. One could assume that darker/certain color spots get warmer from sunlight while over terrain. Eg. a hot road for example or buildings. This would provide some consistency to false positives instead of being at random.

[Zeus67]

One interesting thing to do while simplifying the simulation of this and providing false positives, might be to just consider the color of the objects and terrain being drawn. One could assume that darker/certain color spots get warmer from sunlight while over terrain. Eg. a hot road for example or buildings. This would provide some consistency to false positives instead of being at random.

 

That will require terrain interrigation, which is something we don't want to do.

[MonnieRock]

Unfortunately though the best option means the most amount of work - it is a system that works well in at least Arma 3 but will result in more work. That is a remodel of how infrared works in DCS.

 

In Arma 3 there is an actual dynamic temperature system - object surfaces and in some cases internals actually have a temperature given to them, this temperature is dynamic, so as vehicle engines start-up they get hotter giving better contrast to IR sensors, same goes for barrels when firing or aircraft and vehicles shutting down. Again this takes current weather conditions into consideration. This means we get a more accurate simulation of the Vs

 

Unfortunately this means applying a system to not only IR sensors in DCS but also all of our maps and every single object. Then making it dynamic for weather conditions and states.

 

For me it's a dream FLIR system and in Arma 3 it works beautifully especially having it dynamic.

 

Oh well, just an idea - though a pretty costly one.

 

As much as many if us would love such a solution, the current limitation is likely the map size.

 

While the tiny ARMA map land mass is a bit smaller than the modeled City of Las Vegas on the NTTR map, this system would still require to query ALL objects on the map like buildings, trees, static cars, signs, traffic lights, bushes, etc. (we exclude the ground and water here) to determine their temperature. So if the view expands over Las Vegas, that would mean a LOT of object temperatures to query...

 

As Zeus said, this would require more CPU load per cycle in the Sim World, which is pretty much a guarantee for one-digit fps...

 

And you would need to store this information in the memory for all objects.

 

What we need here is a balance between fidelity and feasibility to give us realistic results without bogging down performance.

 

Another factor would be how IR missiles interpret any hotspots. I presume the system would be part of the core software, or are we talking about something exclusive to the AV-8B?

 

It's a toughie, and whatever solution is arrived at, it's bound to have limitations. For me, something that renders an image that gives us a typical set of returns, but doesn't necessarily react to what's on the ground may have to suffice. I dread to think what the hit on processing resources might be if there's an algorithm included to calculate what might or might not throw up false returns.

 

I guess it is just a question of what compromise RAZBAM feel is appropriate.

 

^this... A complete overhaul of the IR detection system can only be part of DCS World and thus require a feasible compromise as CPU cycles is a limited resource.

 

When part of the avionics/sensor suite of a "Combat" aircraft is neutered in a simulation, it becomes nothing more than pretty switches and dials. Not the tool which the real pilots , ie the NAVFLIR, use to complete "Combat" mission objectives. If someone wanted neutered avionics/senors suites, get Flaming Cliffs.

 

I could fully understand if a part of avionics/ senors suite was still classified that some limitations would be present.

 

When a 3rd Party Developer is limited by the platform which they are developing for, time for the platform to get updated.

 

Where does this "compromise" end? Few things said over the years.

AI can not have higher fidelity flight models (AFM): CPU cycles/poor frame rates

Advanced damage modelling for all air frames: CPU cycles/poor frame rates

IR/FLIR problems(which many air frames use including the new F\A-18C) since beta A-10C: CPU cycles/poor frame rates/total rework of maps/objects

 

Until we use the processing power now available(all new processors are multicore) the single threaded platform/with one thread for sound is going to be limited in CPU cycles.

 

Where is the "Can Do" mentality versus the "You have to settle for less"?

 

I congratulate as well as respect RAZBAM for making a wonderful AV-8B N/A but sorry, will not purchase any new moduels/terrains until the platform matures to what is needed for Combat operations.

 

Happy Simming

[shagrat]

When part of the avionics/sensor suite of a "Combat" aircraft is neutered in a simulation, it becomes nothing more than pretty switches and dials. Not the tool which the real pilots , ie the NAVFLIR, use to complete "Combat" mission objectives. If someone wanted neutered avionics/senors suites, get Flaming Cliffs.

 

I could fully understand if a part of avionics/ senors suite was still classified that some limitations would be present.

 

When a 3rd Party Developer is limited by the platform which they are developing for, time for the platform to get updated.

 

Where does this "compromise" end? Few things said over the years.

AI can not have higher fidelity flight models (AFM): CPU cycles/poor frame rates

Advanced damage modelling for all air frames: CPU cycles/poor frame rates

IR/FLIR problems(which many air frames use including the new F\A-18C) since beta A-10C: CPU cycles/poor frame rates/total rework of maps/objects

 

Until we use the processing power now available(all new processors are multicore) the single threaded platform/with one thread for sound is going to be limited in CPU cycles.

 

Where is the "Can Do" mentality versus the "You have to settle for less"?

 

I congratulate as well as respect RAZBAM for making a wonderful AV-8B N/A but sorry, will not purchase any new moduels/terrains until the platform matures to what is needed for Combat operations.

 

Happy Simming

Well it ends where you are required to setup a separate water cooled server for simulating the environment etc. and a second one to handle the plane, with hardware alone exceeding US$ 5,000 to US$ 6,000 excluding the know how to set it up and configure the network...

 

The current PC hardware for home use, simply can't handle that, even the setup above may struggle.

 

It is either doing it like the RWR, the IFF, and ECM in all other planes (simplified/not 100%, but believable and "feeling" realistic), or leave the system out completely.

 

The current options we discuss to implement it are:

 

A) we get a system that only shows valid contacts, unlike the real system. So we will have a "neutered" system that is flawless

 

B) a more believable randomization of "false hotspots" to simulate the real life behaviour as much as possible, not making it unrealistically easy to find targets

 

C) not simulate the NAVFLIR's capability to detect hotspots/potential targets at all, degrading the planes capabilities

 

D) wait for a 8GHz CPU or multi-core CPU magic that can handle the load, then develop a new core DCS system to handle heat mapping on the maps and objects, maybe in the 2030ies (this is my personal and rough estimate with no intricate background of the complexity or future development in the PC area, or EDs plans).

 

So what is your preferred choice?

[dimitrischal]

Well it ends where you are required to setup a separate water cooled server for simulating the environment etc. and a second one to handle the plane, with hardware alone exceeding US$ 5,000 to US$ 6,000 excluding the know how to set it up and configure the network...

 

The current PC hardware for home use, simply can't handle that, even the setup above may struggle.

 

It is either doing it like the RWR, the IFF, and ECM in all other planes (simplified/not 100%, but believable and "feeling" realistic), or leave the system out completely.

 

The current options we discuss to implement it are:

 

A) we get a system that only shows valid contacts, unlike the real system. So we will have a "neutered" system that is flawless

 

B) a more believable randomization of "false hotspots" to simulate the real life behaviour as much as possible, not making it unrealistically easy to find targets

 

C) not simulate the NAVFLIR's capability to detect hotspots/potential targets at all, degrading the planes capabilities

 

D) wait for a 8GHz CPU or multi-core CPU magic that can handle the load, then develop a new core DCS system to handle heat mapping on the maps and objects, maybe in the 2030ies (this is my personal and rough estimate with no intricate background of the complexity or future development in the PC area, or EDs plans).

 

So what is your preferred choice?

 

Proper thermal mapping and multi core compatibility are doable within current specs easily. So is semi believable ecm and radar operation for that matter.

As for terrain hotspot simulation for the av8 certainly some compromises can be made but claiming a proper thermal map is resource hungry to a forbidding extend I think is inaccurate.

The a10c is out for years and it main target acquisition system is CRAP. It’s also the same target acquisition system for the harrier and the f18 at least for ground targets. This is unacceptable...

[shagrat]

Proper thermal mapping and multi core compatibility are doable within current specs easily. So is semi believable ecm and radar operation for that matter.

As for terrain hotspot simulation for the av8 certainly some compromises can be made but claiming a proper thermal map is resource hungry to a forbidding extend I think is inaccurate.

The map is requiring memory, the OBJECTS (trees, buildings, traffic signs) need to be iterated through a list.

That means the list needs to be checked regularly...

If it was that easy and not requiring ressources I am sure we would have it already.

The a10c is out for years and it main target acquisition system is CRAP. It’s also the same target acquisition system for the harrier and the f18 at least for ground targets. This is unacceptable...

So go and play any alternative that is better in simulating modern aircraft... Oh wait a second. :dunno:

[dimitrischal]

The map is requiring memory, the OBJECTS (trees, buildings, traffic signs) need to be iterated through a list.

That means the list needs to be checked regularly...

If it was that easy and not requiring ressources I am sure we would have it already.

 

So go and play any alternative that is better in simulating modern aircraft... Oh wait a second. :dunno:

 

If it needs more resources let it be. Others have made it happen so should ED. Especially if this is the main targeting aid for 3 of their modules.

 

You’re telling me to quit playing the a10 because the flir is broken and I’m asking for a proper(ish) iteration. Sure buddy you’re on to it. Keep up the good work.

On the meanwhile settle for trees with higher thermal contrast than tank engines on the best modern combat sim available.

[shagrat]

If it needs more resources let it be. Others have made it happen so should ED. Especially if this is the main targeting aid for 3 of their modules.

 

You’re telling me to quit playing the a10 because the flir is broken and I’m asking for a proper(ish) iteration. Sure buddy you’re on to it. Keep up the good work.

On the meanwhile settle for trees with higher thermal contrast than tank engines on the best modern combat sim available.

May be I can't explain it good enough, "others" didn't have to deal with areas the size of DCS maps and "others" don't require CPU power for querying a quazillion objects in your FOV.

 

"Others" do not require lots of CPU power to calculate the fuel systems, electrical systems, hydraulic systems, AI for hundreds of combat vehicles, the ballistics for every round fired, bomb dropped, rocket launched.

They usually have an animated Soldier-Doll with 7 damage areas or a vehicle / helicopter / plane with an arcade mouse control and very simplified flight dynamics to handle...

 

It isn't "broken" it is like this for a reason.

 

I have not seen any other flightsim(!) on the market that simulate modern IR equipment 100% realistic in the last two decades.

 

There are professional Tank simulations that do it nicely in a 40x30km box, but that is about it. The infantry simulations with added arcade planes can't handle the large environment and object density with the requiered visibility for a good flight sim.

 

We are back to decide between A, B, C and D... as I mentioned above.

 

And I am still very sure, if there was any simple solution that just requires a high end PC ED would have it implemented a long time ago. Maybe with an option to switch it on/off.

[dimitrischal]

May be I can't explain it good enough, "others" didn't have to deal with areas the size of DCS maps and "others" don't require CPU power for querying a quazillion objects in your FOV.

 

"Others" do not require lots of CPU power to calculate the fuel systems, electrical systems, hydraulic systems, AI for hundreds of combat vehicles, the ballistics for every round fired, bomb dropped, rocket launched.

They usually have an animated Soldier-Doll with 7 damage areas or a vehicle / helicopter / plane with an arcade mouse control and very simplified flight dynamics to handle...

 

It isn't "broken" it is like this for a reason.

 

I have not seen any other flightsim(!) on the market that simulate modern IR equipment 100% realistic in the last two decades.

 

There are professional Tank simulations that do it nicely in a 40x30km box, but that is about it. The infantry simulations with added arcade planes can't handle the large environment and object density with the requiered visibility for a good flight sim.

 

We are back to decide between A, B, C and D... as I mentioned above.

 

And I am still very sure, if there was any simple solution that just requires a high end PC ED would have it implemented a long time ago. Maybe with an option to switch it on/off.

 

It’s a black and white screen with a negative color filter in the back... Anything would be better end of story, even a half assed iteration with a query and draw distance limit.

Even the way it is now with lower terrain contrast and increased object brightness would suffice.

If I do it manual via luas it fails integrity check... It’s too bad to defend to any degree on its current iteration so please give me a break.

 

I don’t care if they go full house on the thermal department nor do I want it. I want something usable instead of being unable to discern operating tanks on a winter field.

I’m sure 2,5 will bring improvements I’m just waiting.

[DragonShadow]

That will require terrain interrigation, which is something we don't want to do.

 

Ah yes, I’m not sure what your limitations of access are as module makers but the thought was if you have have access to graphics buffer/shaders, one possibility might be to handle this directly there. That is instead of querying the terrain, to consider the pixels of the hud area being drawn and draw the hotspot markers based on the buffer or even directly as a shader in GPU. Admittedly I’m not sure how well that would scale in performance in DCS comparatively, but an alternative approach at any rate.

[Northstar98]

As much as many if us would love such a solution, the current limitation is likely the map size.

I'm not having a go or anything but surely the entire map doesn't need to be interrogated at once? The sensor has a limited FOV and from the video not a colossal range. The same concept of rendering distances and FOV applies here, it still applies to Arma 3 as that's how it overcomes rendering tonnes of objects.

 

While the tiny ARMA map land mass is a bit smaller than the modeled City of Las Vegas on the NTTR map, this system would still require to query ALL objects on the map like buildings, trees, static cars, signs, traffic lights, bushes, etc. (we exclude the ground and water here) to determine their temperature. So if the view expands over Las Vegas, that would mean a LOT of object temperatures to query...

 

As Zeus said, this would require more CPU load per cycle in the Sim World, which is pretty much a guarantee for one-digit fps...

 

In Las Vegas then yes I can agree however for me the solution that I can see is the same concept to level of detail, that is from afar only buildings are taken into account (but again, I say afar - what's the approximate effective range of the sensor?) the sensor's FOV should also surely thin out most objects, I mean maybe I don't have a single clue what I'm talking about and maybe there might be still too many objects (unless you want to kill the detail in Las Vegas). Thing is though if there's one place I can think of where I'd not only expect to see a Harrier screaming through at low level, but also where realistically targets are going to be it's Las Vegas. Attack helicopters maybe, Harriers no.

 

What we need here is a balance between fidelity and feasibility to give us realistic results without bogging down performance.

Could not agree more I just thought that because Arma 3 has an ideal system, that still has things like LOD and rendering distances so that frames don't drop. And all the Arma 3 system really is, is a dynamic texture - our FLIR system is AFAIK a static texture

 

May be I can't explain it good enough, "others" didn't have to deal with areas the size of DCS maps and "others" don't require CPU power for querying a quazillion objects in your FOV.

Again, I can't think of a situation where I would have the whole map in the FOV of the sensor, and I would be so high up that the resolution and effective range of the sensor I'm guessing would have difficulty picking targets up.

 

"Others" do not require lots of CPU power to calculate the fuel systems, electrical systems, hydraulic systems, AI for hundreds of combat vehicles, the ballistics for every round fired, bomb dropped, rocket launched.

Going to have to disagree here, DCS is superior in flight dynamics and aircraft system modelling but that applies to the player aircraft only, and for very good reason. Arma 3 has accurate simulation for every single weapon, including ballistics, penetration and fragmentation mechanics that applies to everything including ground vehicles, there are things that are missing from said fictional vehicles, but it's ahead in DCS in that anything that isn't an aircraft in DCS gets a HP based damage system, AFAIK 1 damage zone which is the entire vehicle - which is why I can defeat tanks with ERA on top of composite armour with HE shells. DCS on the other hand doesn't have to deal with round penetration in walls, through buildings, through armoured vehicles and the damage system for most weapons which have anything to do with cluster bombs isn't quite there yet.

 

They usually have an animated Soldier-Doll with 7 damage areas or a vehicle / helicopter / plane with an arcade mouse control and very simplified flight dynamics to handle...

Well lets take a step back, anything that isn't an aircraft in DCS gets what 1 damage area? It's entirely HP based. At least with Arma 3 apart from maybe buildings you get realistic infantry simulation with individual damage zones for limbs head etc and then get a damage system for vehicles that is only slightly dumbed down (but still using multiple damage zones, down to individual wheels). Yes it's true Arma 3 has simplistic flight simulation and limited damage modelling (still better than HP based though) and more arcade style flight modelling and arcade style systems but it's still miles better for other things, considering where it's heart really is which is an infantry simulator, it has realistic ballistics and damage for every single weapon, including penetration and fragmentation, in DCS we have realistic ballistics but no penetration and fragmentation that I've seen being described as broken (particularly noticeable on cluster bombs such as the BK90). DCS also doesn't have to deal with underwater at all which means no underwater weapons and no amphibious vehicles - Arma 3 also has this). What's also apparent in DCS is what is quite possibly the worst ground vehicle suspension system and driving system for tracked vehicles in that the whole tracks and wheels stay with the terrain perfectly keeping their form and the rest of the hull bobs randomly up and down like a small boat in the waves, so that in some cases the tops of the track clips through the top of the vehicle, not to mention it looking pretty questionable. I mean that's the opposite to what suspension should be as suspension should dampen movement of the hull not make it bob around. Plus not a single vehicle can neutral steer, to be honest tank track physics is about the same as it is in FSX - like any other wheeled vehicle with a better turning radius. On top of this not a single vehicle has any form of gun stabilisation, the optic ports also don't look like they're actually being looked through they just look like the F4 camera with some symbols and text photoshopped over it (as expected because no vehicle has it's own internals) and the paid asset pack has this problem also, okay yes I'm pretty certain WW2 tanks didn't have 2-plane stabilisers but the issues with suspension persist. Also lots of vehicles have functionality that is either totally absent or somewhat lacking: vehicle lights, weapon reloading for tube launched/rail launched missiles on vehicles like the Roland ADS, BMP-1, BMP-2 etc; shell casing ejection on the T-72, T-80U and T-90A, smoke dischargers, folding of RADARs, hydraulic jacks on the SpGH DANA, the list goes on.

 

Now Arma 3 despite being an infantry simulator at heart, still does ground vehicles amazingly, they can neutral steer, they have proper suspension etc, DCS has none of that. So the same argument applies to DCS, only areas that aren't Arma 3's principle highlight are modelled better that the things that aren't DCS's highlight. I can accept that a lot of these things aren't a problem for most people, it's low priority and that's absolutely fine by me, but we simply can't dismiss other things as being totally rubbish in other aspects that aren't their focus when DCS is no better and in some cases worse. (Please don't take any of the above the wrong way, I'm just making a point, I realise that DCS's focus is on aircraft which it does far better than anything else out there, but it still has even more crude things than arcade flight dynamics, however for a flight simulator it's still the king in my books).

 

There are professional Tank simulations that do it nicely in a 40x30km box, but that is about it. The infantry simulations with added arcade planes can't handle the large environment and object density with the required visibility for a good flight sim.

Again, the visibility range and LOD comes into play, they get over the fact that they can't handle huge maps all at once by having visibility ranges, DCS also has visibility ranges, LODs and FOVs - it's the same exact system just with IR as well as visible (but with IR detail can be taken out the window as the system in at least the harrier only needs to choose 10 hot spots at any one time, but yeah again I'm fairly sure I have no idea what I'm talking about and so this would be a good point to shut myself up).

 

The other solution that I can see is keep the current system (different texture) but employ a system that detects the white (hot) in a darker (colder background) and just pick out the 10 most intense (which should be vehicles regardless), or randomly select 10 spots, that makes it randomised, but still gives some realism. But then again

That will require terrain interrogation, which is something we don't want to do.

So out the window that goes.

 

Once again, just to be clear, it's just an idea, probably a bad one - I don't know much of anything, I'm not a developer, I don't know everything.

 

Kind Regards,

 

Ollie

Edited November 4, 2017 by Northstar98

[shagrat]

I'm not having a go or anything but surely the entire map doesn't need to be interrogated at once? The sensor has a limited FOV and from the video not a colossal range. The same concept of rendering distances and FOV applies here, it still applies to Arma 3 as that's how it overcomes rendering tonnes of objects.

 

In Las Vegas then yes I can agree however for me the solution that I can see is the same concept to level of detail, that is from afar only buildings are taken into account (but again, I say afar - what's the approximate effective range of the sensor?) the sensor's FOV should also surely thin out most objects, I mean maybe I don't have a single clue what I'm talking about and maybe there might be still too many objects (unless you want to kill the detail in Las Vegas). Thing is though if there's one place I can think of where I'd not only expect to see a Harrier screaming through at low level, but also where realistically targets are going to be it's Las Vegas. Attack helicopters maybe, Harriers no.

 

Could not agree more I just thought that because Arma 3 has an ideal system, that still has things like LOD and rendering distances so that frames don't drop. And all the Arma 3 system really is, is a dynamic texture - our FLIR system is AFAIK a static texture

 

Again, I can't think of a situation where I would have the whole map in the FOV of the sensor, and I would be so high up that the resolution and effective range of the sensor I'm guessing would have difficulty picking targets up.

 

Going to have to disagree here, DCS is superior in flight dynamics and aircraft system modelling but that applies to the player aircraft only, and for very good reason. Arma 3 has accurate simulation for every single weapon, including penetration and fragmentation mechanics -> for every vehicle - anything that isn't an aircraft in DCS gets a HP based damage system -> which is why I can defeat tanks with ERA on top of composite armour with HE shells. DCS on the other hand doesn't have to deal with round penetration in walls, through buildings, through armoured vehicles and the damage system for most weapons which have anything to do with cluster bombs isn't quite there yet.

 

Well lets take a step back, anything that isn't an aircraft in DCS gets what 1 damage area? It's entirely HP based. At least with Arma 3 apart from maybe buildings you get a realistic infantrymen and then get a damage system for vehicles that is only slightly dumbed down. Yes it's true Arma 3 has simplistic flight simulation and limited damage modelling and arcade flight systems but it's still miles better for other things considered for where it's heart really is which is an infantry simulator, it has realistic ballistics and damage for every single weapon, including penetration and fragmentation, in DCS we have realistic ballistics but no penetration and fragmentation that I've seen being described as broken (particularly noticeable on cluster bombs such as the BK90). DCS also doesn't have to deal with underwater which means no underwater weapons and no amphibious vehicles - Arma 3 also has this). What's also apparent in DCS is what is quite possibly the worst ground vehicle suspension system and driving system for tracked vehicles in that the whole tracks and wheels stay with the terrain perfectly and the rest of the hull bobs randomly up and down, so that in some cases the tops of the track clips through the top of the vehicle, not a single vehicle can neutral steer, not a single vehicle has any form of gun stabilisation, the optic ports also don't look like they're actually being looked through they just look like the F4 camera with some information drawn over it (as expected because no vehicle has it's own internals) but - even the paid asset pack has this problem.

 

Now Arma 3 despite being an infantry simulator at heart, still does ground vehicles amazingly, they can neutral steer, they have proper suspension etc, DCS has none of that. So the same argument applies to DCS, only areas that aren't Arma 3's principle highlight are modelled better that the things that aren't DCS's highlight. I can accept that a lot of these things aren't a problem for most people, it's low priority and that's absolutely fine by me, but we simply can't dismiss other things as being totally rubbish in other aspects that aren't their focus when DCS is no better and in some cases worse. (Please don't take any of the above the wrong way, I'm just making a point, I realise that DCS's focus is on aircraft which it does far better than anything else out there, but it still has even more crude things than arcade flight dynamics, however for a flight simulator it's still the king in my books).

 

Again, the visibility range and LOD comes into play, they get over the fact that they can't handle huge maps all at once by having visibility ranges, DCS also has visibility ranges, LODs and FOVs - it's the same exact system just with IR as well as visible (but with IR detail can be taken out the window as the system in at least the harrier only needs to choose 10 hot spots at any one time, but yeah again I'm fairly sure I have no idea what I'm talking about and so this would be a good point to shut myself up).

 

The other solution that I can see is keep the current system (different texture) but employ a system that detects the white (hot) in a darker (colder background) and just pick out the 10 most intense (which should be vehicles regardless), or randomly select 10 spots, that makes it randomised, but still gives some realism. But then again So out the window that goes.

 

Once again, just to be clear, it's just an idea, probably a bad one - I don't know much of anything, I'm not a developer, I don't know everything.

 

Kind Regards,

 

Ollie

A simple question: How do you know which of the billion map objects is in range of the sensor, if you are a computer?

[Northstar98]

A simple question: How do you know which of the billion map objects is in range of the sensor, if you are a computer?

 

The answer is how the current visibility range system works. The answer to how DCS knows to do the things demonstrated here. How does it know not to render trees and buildings and vehicles and missiles and ships and smoke and explosions etc etc etc far into the distance from where my point of view currently is. That's why I see things popping up visually as I get closer and disappearing visually as I move further away. I mean AFAIK and please correct me if I'm wrong DCS already does what your question is asking. The evidence is here: https://www.youtube.com/watch?v=2oL4Ci4wQYk

 

Object far into distance, DCS decides either not render it or render it in low detail as per the logic behind LODs because - player viewpoint too far away, so anything to do with the object visually is affected.

 

I mean I see, with my own eyes, not being able to see trees and buildings far into the distance, only when I get closer and they all come into view does my performance become impacted at present on my entry-medium level PC running at high graphics settings, this is because (to my untrained eyes, again, I don't claim to know anything, I'm not a developer) they aren't being rendered from further afar. Also what about level of detail? I mean FSX does this, so does Arma 3. What's the point of rendering something in high detail when its so far away that I can just about make out the shape and general colour. I mean to me if DCS is rendering things in high detail that I can't even see from my viewpoint then there's a potential performance problem just staring at you. (Again FSX does this).

 

If you feel that I'm spouting nonsense, as I might well be doing please tell me and I will shut myself up.

Edited November 4, 2017 by Northstar98

[shagrat]

How does the current visibility range system work? How does it know not to render trees and buildings and vehicles and missiles and ships and smoke and explosions far into the distance from where my point of view currently is. That's why I see things popping up as I get closer. I mean AFAIK and please correct me if I'm wrong DCS already does what your question is asking.

 

I mean I see, with my own eyes, not being able to see trees and buildings far into the distance, only when I get closer and they all come into view does my performance become impacted at present on my entry-medium level PC running at high graphics settings, this is because (to my untrained eyes, again, I don't claim to know anything, I'm not a developer) they aren't being rendered from further afar. Also what about level of detail? I mean FSX does this, so does Arma 3. What's the point of rendering something in high detail when its so far away that I can just about make out the shape and general colour.

Please answer the question.

[Northstar98]

Please answer the question.

 

Shagrat it's irrelevant it's already in DCS. I can see sliders for them in the settings. Oooh look at that, a separate one for trees!

 

I've tried to conduct myself peacefully here with as little personal confrontational and as little implicit personal attacks as my level of English and sentence structure will permit. But if you're attempting to make me look stupid as I suspect you are by trying to pry an answer to a "simple question" (your words not mine) out of me that you know full well that I am unable to provide to a level of satisfaction that I assume you want from me after I've explained above that I'm not a programmer, I don't know what I'm talking about, I don't know how to implement it, all of these are things I've stated above), then I'm afraid I have no business in making suggestions ever again, I will delete every single one I've ever made as clearly to you at least I have no right to do so because I'm apparently one of inferior intelligence. I've only tried to make a point and try and support that point with what little insight I have.

 

If you feel I've been rude, dishonest, disrespectful even implicitly so in replying to your replies about my suggestion then please tell me, I won't bite! It was never my intention to do so and if you feel that way and you tell me then I will be ashamed if I don't apologise and at the very least consider editing my past responses in any more than a heartbeat if that's the tone I've conveyed. Don't be afraid to tell me that my suggestion simply won't work, if I'm wrong, if my justification is wrong then please just say so, pick it apart, maybe it's impossible whatever the weather please enlighten and correct me! I won't bite I promise! I've already made it clear that I'm not a programmer.

 

The thing is you're asking me this question, which is for a feature that has been in DCS and many other games for ages while I spawn at Batumi facing south inside the cockpit, looking ahead, getting relatively nice FPS for what I'm used to on my entry-medium level outdated PC, when I look to the north or north-east however, they drop. Now what is to the north/north east I hear you ask? The answer are hills/mountains most of them totally infested with trees, as well as Batumi itself making up in total thousands of objects. But when I look away and the hill and Batumi are behind be and not in view the FPS increases and I have smooth performance. What does my seemingly apparent puny, inferior mind interpret from this? The answer seems as though when I'm not looking at something it isn't rendered, explaining the better performance when the hill isn't physically in view and outside of the FOV. In other words accomplishing the same thing as what I've suggested with using visibility ranges/FOVs and LOD as a workaround for lots of objects, saving performance. It's already in DCS, it's been here for a while, how many times need I say this?

 

I told you I'm not a programmer - I don't know what I'm talking about in regards to more technical aspects of how I myself would accomplish this task as if I were a developer, that's why I'm not one. I don't know the specific answer, to how DCS, FSX, Arma 3, KSP, War Thunder, SF2, SH3 blah blah blah do this. I can just about write "code" that tells my computer to display a warning message and shut down after a time limit after clicking a shortcut (which can be entertaining making viruses as a practical joke when somebody tries to go on the internet using a shortcut only to be greeted with an error message and a computer that shuts down for April fools day) but I am not (I repeat not a developer), otherwise I'd be trying to actively accomplish it, instead of making mere suggestions.

 

The question you should be asking is how can this system be applied to work for this sensor as well, again I don't know the exact precise answer, it's like asking a receptionist how SARH works - balance of probability suggests they don't know, only SARH can be googled and the key concepts and answer known within seconds, the finer details about programming are not.

 

I mean like I said I don't know if you're trying to humiliate me by making me look stupid, I've already told you what to do if you want me to silence myself. Heck I'll even make a thread telling the community just how little I know if that satisfies you.

 

Ollie

Edited November 4, 2017 by Northstar98

[shagrat]

Shagrat it's irrelevant it's already in DCS. I can see sliders for them in the settings. Oooh look at that, a separate one for trees!

 

I've tried to conduct myself peacefully here with as little personal confrontational and as little implicit personal attacks as my level of English and sentence structure will permit. But if you're attempting to make me look stupid as I suspect you are by trying to pry an answer to a "simple question" (your words not mine) out of me that you know full well that I am unable to provide to a level of satisfaction that I assume you want from me after I've explained above that I'm not a programmer, I don't know what I'm talking about, I don't know how to implement it, all of these are things I've stated above), then I'm afraid I have no business in making suggestions ever again, I will delete every single one I've ever made as clearly to you at least I have no right to do so because I'm apparently one of inferior intelligence. I've only tried to make a point and try and support that point with what little insight I have.

 

If you feel I've been rude, dishonest, disrespectful even implicitly so in replying to your replies about my suggestion then please tell me, I won't bite! It was never my intention to do so and if you feel that way and you tell me then I will be ashamed if I don't apologise and at the very least consider editing my past responses in any more than a heartbeat if that's the tone I've conveyed. Don't be afraid to tell me that my suggestion simply won't work, if I'm wrong, if my justification is wrong then please just say so, pick it apart, maybe it's impossible whatever the weather please enlighten and correct me! I won't bite I promise! I've already made it clear that I'm not a programmer.

 

The thing is you're asking me this question, which is for a feature that has been in DCS and many other games for ages while I spawn at Batumi facing south inside the cockpit, looking ahead, getting relatively nice FPS for what I'm used to on my entry-medium level outdated PC, when I look to the north or north-east however, they drop. Now what is to the north/north east I hear you ask? The answer are hills/mountains most of them totally infested with trees, as well as Batumi itself making up in total thousands of objects. But when I look away and the hill and Batumi are behind be and not in view the FPS increases and I have smooth performance. What does my seemingly apparent puny, inferior mind interpret from this? The answer seems as though when I'm not looking at something it isn't rendered, explaining the better performance when the hill isn't physically in view and outside of the FOV. In other words accomplishing the same thing as what I've suggested with using visibility ranges/FOVs and LOD as a workaround for lots of objects, saving performance. It's already in DCS, it's been here for a while, how many times need I say this?

 

I told you I'm not a programmer - I don't know what I'm talking about in regards to more technical aspects of how I myself would accomplish this task as if I were a developer, that's why I'm not one. I don't know the specific answer, to how DCS, FSX, Arma 3, KSP, War Thunder, SF2, SH3 blah blah blah do this. I can just about write "code" that tells my computer to display a warning message and shut down after a time limit after clicking a shortcut (which can be entertaining making viruses as a practical joke when somebody tries to go on the internet using a shortcut only to be greeted with an error message and a computer that shuts down for April fools day) but I am not (I repeat not a developer), otherwise I'd be trying to actively accomplish it, instead of making mere suggestions.

 

The question you should be asking is how can this system be applied to work for this sensor as well, again I don't know the exact precise answer, it's like asking a receptionist how SARH works - balance of probability suggests they don't know, only SARH can be googled and the key concepts and answer known within seconds, the finer details about programming are not.

 

I mean like I said I don't know if you're trying to humiliate me by making me look stupid, I've already told you what to do if you want me to silence myself. Heck I'll even make a thread telling the community just how little I know if that satisfies you.

 

Ollie

Ok. I'll try to explain a bit. The current World (what you "see" is one Iteration through all objects, area layers, trees, AI objects etc. for visibility) now you need to do this again for the "IR value" up to doubling the required CPU cycles.

 

So two object tables rather than one.

 

That needs to be done each frame or at least every second frame.

[Northstar98]

Ok. I'll try to explain a bit. The current World (what you "see" is one Iteration through all objects, area layers, trees, AI objects etc. for visibility) now you need to do this again for the "IR value" up to doubling the required CPU cycles.

 

So two object tables rather than one.

 

That needs to be done each frame or at least every second frame.

 

Okay I understand that, much appreciated - that's all I needed.

 

But instead of just copying over this system and applying the 'IR value' could it not be made so that the FOV and visibility range is smaller (but done somehow without nerfing the sensor) so that on balance less objects are visible to the sensor compared to the user's eyes if you see what I mean? I fear this system is too much for the reasons you have stated, which I guess oh well.

 

The only other thing that I can think of for a system is using the current IR system, which essentially (if I'm not mistaken) is what we as the user is able to see but is a different texture to the sensor. Objects likely to be hot are white, objects likely to be colder are darker. Would it be possible some way to get the sensor, for it's FOV to take a look at this image, and in a similar way to camera face detection find say the areas of white, lets say it points out the 10 most intense spots or 10 random spots and then gives you the V over that area, the hot-spot could be your target, but it could be also be a false contact, giving you a better representation of the sensor as opposed to a god-mode sensor that will only pick out objects or a randomised sensor. So instead of interrogating objects individually, it interrogates the in game texture that normally our eyes and brains have to interpret to find hot-spots. Maybe this suggestion is even worse performance wise but I don't know.

Edited November 4, 2017 by Northstar98

[shagrat]

Thank you that's much appreciated - that's all I needed.

 

But instead of just copying over this system and applying the 'IR value' could it not be made so that the FOV and visibility range is smaller (but somehow without nerfing the sensor) so that on balance less objects are visible to the sensor if you see what I mean? I fear this system is too much for the reasons you have stated. Which oh well.

 

The only other thing that I can think of for a system is using the current system, which essentially (if I'm not mistaken) is what we as the user is able to see but is a different texture to the sensor. Objects likely to be hot are white, objects likely to be colder are darker. Would it be possible some way to get the sensor, for it's FOV to take a look at this image, and in a similar way to camera face detection find areas of white, lets say it points out the 10 most intense spots and then gives you the V over that area, the hot-spot could be your target, but it could be also be a false contact, giving you a realistic representation of the sensor. So instead of interrogating objects individually, it interrogates the asset that normally our eyes and brains interpret. Maybe this suggestion is even worse but I don't know.

If I understand the current System correct, it uses a way of contrast and B/W on the "normal" textures.

 

In one of the first posts Zeus67 from Razbam explained that he discussed this with EDs devs. There was a clear statement it isn't possible to do an object based iteration at the moment.

 

Edit: https://forums.eagle.ru/showthread.php?p=3274852

Edited November 4, 2017 by shagrat