Foliage cover and drawing distance?

[wipman 1]

No way you can see a camouflaged person 300m away, through scope or not...

Look at this guy 2m away

Hi, that photo is not very good, it haven't much quality i mean; don't have depth as the image would have if you were in that room IRL, so is not a good example. But the camouflage works IRL... that's it's pourpouse... you wouldn't spot an inmovile man laying on the ground with a few bushes arround and behind if he've the face and visible skin painted. Unless... you look through a scope... and even then, if you move the scope, binocs or whatever too fast... you wouldn't see him/her; IRL what you use to spot 1St is the weapon silouette before the man/woman's silouette. Is not blured... is blended, but that's hard to get properly on a flat screen, overall when IRL a man/woman would use the proyected shadows to cover his/her body in a better way together with the BDU's camo and maybe some of the enviroment, grass... branches... etc; but without shadows, is easyer to spot the man/woman silouette on a screen. Let's C ya

[Bouben 3]

but without shadows, is easyer to spot the man/woman silouette on a screen. Let's C ya

That leads me to a conclusion that there should be grass and shadows too in Arma while zooming through a scope. Good point.

[liquidpinky 11]

Hi, in my opinion it can't be very improved for the general public because it depends on your machine's power; draw shadows over the distance will improve the game not only on this field that we're talking about, it'll also improve the general shape of the enviroment making it more realistic far away than the 50m of the grass & shadows drawing distance that we've now on the ArmA2, but as i've said... i think that this depends only of the hardware that you've. Let's C ya

Currently with ToH you can adjust shadow draw distance, but it is a major load on the system.

The slight transparency (alpha blending) answer is the simplest and will not bog down peoples PCs either.

It would also be the same across the board so no one could get an advantage over others no matter what their hardware or settings.

Making it the obvious choice for multiplayer.

[dragon01 902]

Indeed, for some reason even snipers in Ghilie suits stick out like a sore thumb most of the time. I hope BI finds some workaround for that issue.

[Fornax 0]

Hi, that photo is not very good, it haven't much quality i mean; don't have depth as the image would have if you were in that room IRL, so is not a good example. But the camouflage works IRL...

Ya the photo was a bit for fun, but it shows that if they include distant textures (through scopes) it will be much more difficult to see far away enemies than what we have with the barren view now...

That would be much better, as I believe they are not capable of including distant (through scopes etc) foliage cover...unfortunately...

[dragon01 902]

In fact, it's about the only place outside of Afghanistan where this camo actually does it's job (or so I heard from a soldier who used it). Guess that "Universal" in UCP means "Useful when hiding from an angry wife or the Taliban." :)

Jokes aside, I found that when scoped in, the amount of detail is not adjusted, despite the fact you see a much smaller area. I think that somehow optimizing it could really help, and the foilage draw distance (among other things) should increase as you scope in.

[metalcraze 290]

When you are scoped in the camera does not move forward. The stuff you see without the scope just gets scaled up.

If the camera was to move forward you'd keep seeing your VD from that position which will be a great cheat since that isn't how magnifying stuff works. Neither it will look like magnification.

At best the calculation will have to be done that calculates what exactly you will see by moving the camera forward to render the stuff there based on your current VD. But balancing that will not be easy.

It will require proportional VD adjustment + FOV adjustment. And with different resolutions/view proportions (4:3/16:9/16:10/5:4/etc) it can be a real PITA as I imagine.

Not only that but due to camera movement (since when camera moves stuff around it moves too according to its 3D position which isn't how it is when magnifying) you may easily reveal a hiding enemy that you shouldn't be seeing when just magnifying stuff IRL. Think about it as moving camera to a corner of the building - suddenly with your scope you are able to see what's behind the corner.

I guess BIS can somehow extend grass draw distance when scoped in but it may easily kill performance since much much more grass will be rendered.

Edited March 1, 2012 by metalcraze

[fraczek 4]

When you are scoped in the camera does not move forward. The stuff you see without the scope just gets scaled up.

If the camera was to move forward you'd keep seeing your VD from that position which will be a great cheat since that isn't how magnifying stuff works. Neither it will look like magnification.

At best the calculation will have to be done that calculates what exactly you will see by moving the camera forward to render the stuff there based on your current VD. But balancing that will not be easy.

It will require proportional VD adjustment + FOV adjustment. And with different resolutions/view proportions (4:3/16:9/16:10/5:4/etc) it can be a real PITA as I imagine.

Not only that but due to camera movement (since when camera moves stuff around it moves too according to its 3D position which isn't how it is when magnifying) you may easily reveal a hiding enemy that you shouldn't be seeing when just magnifying stuff IRL. Think about it as moving camera to a corner of the building - suddenly with your scope you are able to see what's behind the corner.

I guess BIS can somehow extend grass draw distance when scoped in but it may easily kill performance since much much more grass will be rendered.

Nobody I guess wants the 3D position of cam changed when in scoped view. Just magnify and increase object/grass viewdistance. No seeing around corners of course, as the 3D position wouldn't change. One problem I see with it would be balancing the AI though. How do you simulate a scoped weapon wielding AI who should see the grass but have same difficulty of seeing the camo'ed opfor units in the grass through the scope? Mind you, I am not sure how current scoped AI mechanics work...if the AI gets an advantage to spotting or accuracy if using a scoped weapon or not...

[-Coulum- 35]

It will require proportional VD adjustment + FOV adjustment. And with different resolutions/view proportions (4:3/16:9/16:10/5:4/etc) it can be a real PITA as I imagine.

Not only that but due to camera movement (since when camera moves stuff around it moves too according to its 3D position which isn't how it is when magnifying) you may easily reveal a hiding enemy that you shouldn't be seeing when just magnifying stuff IRL. Think about it as moving camera to a corner of the building - suddenly with your scope you are able to see what's behind the corner.

Sounds complex. Overly. Could it not be possible to simply make the clutter/shadow draw distance dynamic? If looking through optics of a certain magnifcation you can then simply multiply the magnification multiplyer by the base draw distance to get the new draw distance.

Ie.

Base drawing distance is 50 m approx.

When zooming by "squinting" you see things twice as big. At least that's what my rough calculations suggest

Therefore the new drawdistance for clutter would be 50 x 2 - 100m

Looking through an acog It would give you a 600m draw distance approx.

I don't see any need to move the camera forward or whatever.

The details of trees and other objects are able to increase in detail in proportion to zoom so I a figure that this kind of thing should also be possible to apply to clutter as well. but I can't say I truly understand what really going on under the hood in game. maybe I am totally off.

The one problem would be performance however. When you have a 600 draw distance would the computer only be trying to render whats directly in your view or would it also be trying to render all the clutter off the screen? if the former, I think this idea would work, especially considering the newer computers most people will have by the time arma 3 comes. If the latter... well then I hope that Bi can find a way to make it the former.

This would not only add more realistic gameplay, but it would also make the game look much much more attractive. I tell people that arma has really good graphics but they have trouble taking me seriously when they see units half sunk into a flat featureless terrain.

One problem I see with it would be balancing the AI though. How do you simulate a scoped weapon wielding AI who should see the grass but have same difficulty of seeing the camo'ed opfor units in the grass through the scope?

I honestly don't really trhink the ai's use of scopes has been balanced yet. From the testing I have done so far the scopes literally make zero dfference on the ai's spotting ability. I just did a simple test giving a teamleader a basic m16, and the other an m16 with the optics. they both spot enemies at the same time. Right now I think Bi simply needs to improve the ai spotting ability through clutter. They have made some significant improvements but the ai's spotting through grass still seems robotic and too clear cut. Ie. they are totally obliviuos because you are in grass or they instantly know exactly where you are. I wouldn't Worry about there use of scopes until after their general spotting through grass/clutter etc. abilities are fixed.

[dmarkwick 261]

I'm sure I've seen a VBS2 video demonstrating a dynamic clutter system for scoped views. Perhaps it renders clutter around a spot centered on the scoped view center position and where it hits the ground. But that sounds to me like a fast-moving scope might have some clutter-lag :)

[Minoza 11]

I'm sure I've seen a VBS2 video demonstrating a dynamic clutter system for scoped views. Perhaps it renders clutter around a spot centered on the scoped view center position and where it hits the ground. But that sounds to me like a fast-moving scope might have some clutter-lag :)

Fine with me.

[Whirly 1]

I'm sure I've seen a VBS2 video demonstrating a dynamic clutter system for scoped views. Perhaps it renders clutter around a spot centered on the scoped view center position and where it hits the ground.

If you ever find that demonstration again I would very much appreciate it if you could post a link. If we can verify with absolute certainty that this has already been done with the RV engine then we can focus on the technicalities.

[dmarkwick 261]

If you ever find that demonstration again I would very much appreciate it if you could post a link. If we can verify with absolute certainty that this has already been done with the RV engine then we can focus on the technicalities.

Found what I was looking for, pertinent footage is at 1:39.

iEJDA7ACi1I

Although I didn't see evidence of dynamic clutter as I mentioned, I still think it's of interest :)

[Minoza 11]

Well, dynamic view distance is not what I hoped for, because it doesn't mean it will render grass at that distance. Even if you only render what's within scope's view, it still doesn't mean everywhere within scope grass would appear, since you can have situations where a vast area would be covered... Means, you need another rendering radius for grass, but withing scope's area. I think it's doable that way.

[-Coulum- 35]

Means, you need another rendering radius for grass, but withing scope's area. I think it's doable that way.

Not sure I follow. Do you mean the centre of the draw distance would have to be placed where the scope Is pointing? Not sure I see the logic behind that if it is what your saying.

The vbs vid still looks promising I think. If it is possible to have distant trees and objects rendered as zoom increases it shouldn't be much more difficult to make it so distant clutter is rendered as zoom increases as well.

Edited March 2, 2012 by -Coulum-

[Minoza 11]

Not sure I follow. Do you mean the centre of the draw distance would have to be placed where the scope Is pointing? Not sure I see the logic behind that if it is what your saying.

The vbs vid still looks promising I think. If it is possible to have distant trees and objects rendered as zoom increases it shouldn't be much more difficult to make it so distant clutter is rendered as zoom increases as well.

Lets put it this way - how would you approach a situation where you have a large area within your scope? Like 2-3 miles of land in front of you, meaning you have to render both close ground cover and far ground cover as well... How would you solve this situation? Scope does narrow your FOV, but you can still have huge amount of grass in your view depending on the angle you're looking at.

Look at this image for example:

http://widescreengamingforum.com/f/u/imagecache/node-gallery-display/contrib/dr/456/extras/arma2007090517044781uh1.jpg

Edited March 2, 2012 by Minoza

[fraczek 4]

I do not see such a big problem. Normal ViewDistance would still apply, no? The distance for rendered clutter (grass,...) could vary with the FOV, for high-powered scopes, it could be much more than say a 2x sight. And if you really were looking with that 20x spotting scope from the highest mountain on Chernarus all the way south to Elektro (meaning you would have few square miles in your scope), it wouldn't make a difference - you would sooner hit your ViewDistance limit. Obviously there would need to be some careful tweaking of ClutterDistance versus FOV, but it could be done to be feasible in terms of FPS.

And for that example with looking under an angle so you would see grass in front plus grass 500m behind at once, it really depends on the engine. The overall amount of clutter might still be less than without a scope because of the narrower FOV. Meaning less or same amount of grass/clutter to render. Also some of the near clutter would still obscure the further clutter, thus it not getting rendered and not having any FPS hit.

Edited March 2, 2012 by fraczek

[Minoza 11]

Well, that actually was my initial thought, dunno why did I even mention 2nd rendering radius... Yes, normal clutter render distance would apply. And good point on the FOV vs. rendering distance, since FOV is smaller, we might even get bigger coverage (well, theoretically...).

[neokika 62]

I think the current problem is that grass is not rendered only where it is needed.

From what it seems, Arma 2 detail grass works something like this:

While it would certaintly be more efficient if it would work like this:

[metalcraze 290]

Objects that are not on screen are not rendered.

Your pic 2 is exactly how it happens. Grass is just very expensive. In fact ArmA2 is the only game on the (at least shooter) market that renders so much of it at once. Even Crysis is much more economical with it.

Nobody I guess wants the 3D position of cam changed when in scoped view. Just magnify and increase object/grass viewdistance.

And it will kill your FPS regardless. Since the game will still have to render the grass close to you. Of course since some objects won't be rendered due to smaller FOV the grass draw distance can be extended while looking through the scope BUT it won't be by much and it certainly won't cover anything 200m+ away.

With the current graphics tech we have it still comes down to the raw polygon rendering/texturing power to render something as detailed as nature. That's why games still use LODs.

Still on ArmA3 screenshots you can see that grass rendering distance is extended even beyond ArmA2's one... I'm afraid of what it will do to my videocard :D

Edited March 2, 2012 by metalcraze

[Minoza 11]

metalcraze has a point there. That's why I suggested that we could maybe use second "interest point" from which grass would be rendered. Atm it's only rendered from players view, so when you zoom in, no grass renders. I have no idea what would that second point of interest be though and if it would work as intended...

[metalcraze 290]

BIS's solution of "drowning" soldiers into the ground is fine as it is. If you will fire up Chernarus and compare how prone soldier looks in a grass 50 and 200m away you will see that it's a pretty fair representation (grass does not cover much on Chernarus even when close). They just need to add some transparency to the soldier model 400+ m away or so when prone in the grass. Such workaround won't be noticed at that distance but it will make the soldier barely visible (if spotted at all) which is how it should be.

[Minoza 11]

From the usability point, that's perfectly fine. I'm just thinking if there could be even better solution, where you'd have same usability but better looks.

[metalcraze 290]

If you will start to render grass around the projected point of your sniper optics this may look quite awkward since you may easily have a huge empty space between that point and you.

The real problem is that on Chernarus there's not much detail but grass so when engine stops rendering grass you end up with a lot of empty space. Dwarden once suggested to simply fill that empty space with random stuff like stones/bushes to make soldier barely distinguishable from the background which may work too. In fact VBS2 already has this. If you will download JCOVE-lite you will see this in action. Think empty takistani hills except filled with many many little stones visible far away. Really hard to spot a prone soldier even in such desert environment. A helmet/head looks like a yet another stone at that distance.

[Minoza 11]

True. Far textures could be improved as well, with more variance and detail if possible. That would also help.