Wind Simulation

[Uziyahu--IDF 0]

I really don't understand the decision not to include wind simulation.

There was wind simulation in Delta Force 2(so it shouldn't require a lot of processing power). It made a difference when sniping.

Wind simulation would also make helicopter flight look A LOT more realistic. How will B.I. simulate attempting to land in rough weather if they don't have wind simulation in TKOH?

[codarl 1]

There is Wind in ArmA2. It suprised me as well! but a "bug" was adressed to AI pilots "don't land into the wind if it is less then 3 m/s". It sure as hell explained why landing an MI-8 in a Firefight outpost (just nearly enough space to park a normal car) was so darn difficult.

[columdrum 11]

There is wind in arma2 but it only affects to smoke and some models( like some bushes). And it has a major flaw, its local to each client, and diferent clients can have diferent wind velocity and direction while been on the same position.

So fixing the wind locality and as extension the whole weather would be first.

But all this is already discused in 2 or 3 post arround here, use search!

[Tonci87 163]

Sniping without wind is a joke. Wind brings challenge

[maddogx 13]

Sniping without wind is a joke. Wind brings challenge

For sniping it can still be scripted, like ACE does. :)

[Przemek_kondor 13]

For sniping it can still be scripted, like ACE does. :)

So every single mission which wants so basic feature like wind affecting sniping should use script for that?

Personaly I don't care (I play this game for other features than sticking to realism), but many of my friends were really surprised there is no such factor (next to bullet type and distance) that affects bullet's ballistics.

[Tonci87 163]

I would be totally happy if it would somehow affect long range sniping for the player

[maddogx 13]

So every single mission which wants so basic feature like wind affecting sniping should use script for that?

Well, yeah. If BIS don't feel the need to implement it in the engine, what other choice do you have?

[rye1 21]

Well, yeah. If BIS don't feel the need to implement it in the engine, what other choice do you have?

Cry and put on a sadface with puppy dog eyes.

[maddogx 13]

Cry and put on a sadface with puppy dog eyes.

Awwwww! :butbut:

[metalcraze 290]

BIS must add a wind simulation. At least do it for bullets and missiles

If ACE can do it - then why not BIS?

Don't be lazy - do it.

[maddogx 13]

Ivan Buchta said in the recent interview that they intentionally aren't including bullet windage because it would overcomplicate things. It seems there is a level of realism BIS aren't willing to breach, instead preferring to keep things simple enough for new guys to learn.

I don't like it much either, but it does make a certain amount of sense. And they aren't in any way obliged to cater exclusively to the hardcore realism crowd.

Edited July 1, 2011 by MadDogX

[rye1 21]

BIS have the correct idea. Keep it simple, get in the new guys. Allow them to chose from the variety of mods that cater all crowds in every way imaginable. Arma is win.

[CarlGustaffa 4]

Why do we need sniping windage? We already have warping to cope with :p

[maddogx 13]

Why do we need sniping windage? We already have warping to cope with :p

At the moment, perhaps, but that may soon be a thing of the past. ;)

[CarlGustaffa 4]

And it has a major flaw, its local to each client, and diferent clients can have diferent wind velocity and direction while been on the same position.

It's not that much of a problem. You can set wind to be same for all clients, even if the variation thing will still happen. Solution to this problem is, well, problematic:

1) Keep the wind super uniform without as much of a hint on variability. Smoke etc would look awful without this added "turbulence", i.e. a pipe stack with smoke going at a constant angle all the time? Yuck.

2) Synchronized variability I assume would quickly cause too much net traffic. A sudden but gradual wind shift over all clients? Lag.

The only realistic option to this problem would be that the server defined some kind of seed that caused the randomization, and this seed was sent to the client at connect. Using synchronized time alone, clients could then do the whole randomization thing using seed lookup, that would become equal over all clients, without causing additional net traffic.

So maybe a switch in the setWind command that allows us to completely turn off randomness, and let the ACE guys figure out a good way to randomize it equally over clients when needed, as the need for this synchronized randomness in the vanilla game is definitely not needed. Using this switch for normal users would just be weird, but maybe just what ACE needs to do their thing?

At the moment, perhaps, but that may soon be a thing of the past. ;)

Yeah hopefully that will help. Hopefully it may also allow multiple bullet shots fired per frame? I don't know what possibilities this adds, but I would love to see fast shooting ZSU's in the future rather than the increased bullet damage.

Edited July 1, 2011 by CarlGustaffa

[maddogx 13]

I believe a hybrid solution would be best.

For starters, use a seed value to create pseudo-random variations in the wind strength at any given moment. These would be the same on all clients in MP, similar to the pseudo-randomly generated infinite terrain.

Secondly, have the overall wind strength and value be generated like it is already, except that it would always be server side, and wind direction changes would always be sent from the server to the clients.

Changes in wind direction could happen as little as once every twenty seconds I believe, since the wind doesn't change instantly, but gradually over time. So if the wind is blowing west to east at 5km/h now, then the server would generate a wind change, say south-west to north-east at 7km/h, to be changed over thirty seconds. This instruction would then be broadcasted to the clients.

The server could also change the scale of the pseudo-random wind variation. Say it's usually at 20%, so wind speeds of 10km/h could vary between 8-12km/h at any given time, the server could also globally amp up the scale to, say, 50%, giving everyone variations from 5-15km/h.

This shouldn't cause much traffic. JIP clients would need to recieve the variation seed and current scale, aswell as the past and future wind times and values, so that they can interpolate them to the current value.

[dmarkwick 261]

I believe a hybrid solution would be best.

For starters, use a seed value to create pseudo-random variations in the wind strength at any given moment. These would be the same on all clients in MP, similar to the pseudo-randomly generated infinite terrain.

Secondly, have the overall wind strength and value be generated like it is already, except that it would always be server side, and wind direction changes would always be sent from the server to the clients.

Changes in wind direction could happen as little as once every twenty seconds I believe, since the wind doesn't change instantly, but gradually over time. So if the wind is blowing west to east at 5km/h now, then the server would generate a wind change, say south-west to north-east at 7km/h, to be changed over thirty seconds. This instruction would then be broadcasted to the clients.

The server could also change the scale of the pseudo-random wind variation. Say it's usually at 20%, so wind speeds of 10km/h could vary between 8-12km/h at any given time, the server could also globally amp up the scale to, say, 50%, giving everyone variations from 5-15km/h.

This shouldn't cause much traffic. JIP clients would need to recieve the variation seed and current scale, aswell as the past and future wind times and values, so that they can interpolate them to the current value.

Yeah wind simulation across MP doesn't have to be traffic heavy. If I were to do it (and I nearly did once) I would:

Server:

1. Set basic weather parameters, wind variance across the mission, rate of change, scale of change etc.

2. Initiate a wind update timescale (example 120 seconds).

3. Initiate a first & second wind vector for the beginning & end of the timescale (direction + speed).

4. Broadcast to clients.

5. Wait for the timescale.

6. Go back to 2.

Client:

1. Set the wind to the recieved setting.

2. Ramp the wind, with small variance, to the second setting over the timescale.

3. Go back to 1.

The above basic idea would be very low on MP traffic, and synchronise wind across clients accurately enough for each to see the same effect, with regard to smoke, ballistics, weather etc.

Edited July 1, 2011 by DMarkwick

[wolfbite 8]

It would also be nice to have wind so that I can land my aircraft with some side wind to add some challenge... Hovering in helicopters would be more fun too.....

[maddogx 13]

The above basic idea would be very low on MP traffic, and synchronise wind across clients accurately enough for each to see the same effect, with regard to smoke, ballistics, weather etc.

Yup, my thoughts exactly. :)

Perhaps such an idea should be documented properly in a short but easily understood way and pitched to BIS via the Arma3 CIT as a feature request.

[dmarkwick 261]

Yup, my thoughts exactly. :)

Perhaps such an idea should be documented properly in a short but easily understood way and pitched to BIS via the Arma3 CIT as a feature request.

Might do that. I had a whole weather system planned out awhile back, but a personal tragedy halted it. It would have set a basic weather movement over a gross time, say several hours, and have small variance within that base movement. Lattitude, time of day, time of year would have been optional constraints, to give missions the possibility of different weather each time, or it could be more tightly controlled. I had written a very nice little curve-remapping function to aid with lattitude-controlled weather, so for example tropical areas at some times of the year acted differently to other times of the year with regard to probability & scale of weather conditions etc.

Maybe I should reboot it :)

Edited July 1, 2011 by DMarkwick

[CarlGustaffa 4]

Client:

1. Set the wind to the recieved setting.

2. Ramp the wind, with small variance, to the second setting over the timescale.

3. Go back to 1.

That's pretty much how I do it in Domino, but of course I can't control how fast it changes or the client side variance/turbulence.

Btw, wind near ground is rarely invariant in direction or smooth in force, due to local conditions and obstructions causing a shitload of small scale turbulence that changes the wind. It is these variations that isn't synchronized. No problem adding a globally synchronized wind change, even over time (except you can't do short times), we've done this for a long time.

But I don't want "no/small scale variance" only because I'm changing global wind direction, I still want the variations or anything particle related might look quite off.

As for wind direction changes being slow. In the air at altitude, generally true (not getting into wind shears and drafts), but near the ground?

X3uanOS2Bq0

[maddogx 13]

That's what pseudo-random variance would be for. :)

Btw. I don't believe the RV engine simulates wind turbulances caused by objects. Wind is currently just a global 2D vector and a force factor, as far as I know.

[dmarkwick 261]

That's pretty much how I do it in Domino, but of course I can't control how fast it changes or the client side variance/turbulence.

Btw, wind near ground is rarely invariant in direction or smooth in force, due to local conditions and obstructions causing a shitload of small scale turbulence that changes the wind. It is these variations that isn't synchronized. No problem adding a globally synchronized wind change, even over time (except you can't do short times), we've done this for a long time.

One thing that would help us out is if the wind were treated in-engine the same as the fog, rain & overcast, with a change to new settings over a specified timescale. And what would REALLY help me is the possibility of these things being concurrent with each other, currently we can only change the fog over time, or the rain, or the overcast, not all 3 at the same time. So it means artificially chopping up any weather changes to step each one at a time.

But I don't want "no/small scale variance" only because I'm changing global wind direction, I still want the variations or anything particle related might look quite off.

As for wind direction changes being slow. In the air at altitude, generally true (not getting into wind shears and drafts), but near the ground?

Local variance could be one of the server settings passed down to clients I guess. How tight you would wish to control this would be a mission design thing.

Edited July 1, 2011 by DMarkwick

[metalcraze 290]

Ivan Buchta said in the recent interview that they intentionally aren't including bullet windage because it would overcomplicate things.

Windage in ACE makes the game incredibly complicated.