Loading [MathJax]/extensions/Safe.js
Special Aircraft Service
the SAS Hangar => The Lounge => Stainless' new Flight Simulator => Topic started by: Stainless on December 18, 2015, 04:59:11 AM
-
I am now off work until next year, lot's of holiday to use. So I have decided to create one whole module of the game.
This is the high level weather system.
I want you guys to get involved with the design, yell at me if you hate it, let me know if you like it.
The high level design is this.
1. Add local maxima : minima
Either randomly, or manually at the mission design stage, add some high pressure zones and low pressure zones.
(http://s22.postimg.cc/irfa7o8xd/add_local_maxima_minima.jpg) (http://postimage.org/)
2. Join local minima
Join the low pressure zones with lines, these will represent the warm/cold weather fronts.
(http://s1.postimg.cc/i6k5i2l0f/link_low_pressure.jpg) (http://postimage.org/)
3. Iterate along the lines
Walk along the lines using the local high pressure zones as control points. Take the influence of all high pressure zones as a force to move the point.
(http://s12.postimg.cc/4nb54n659/apply_high_pressure.jpg) (http://postimage.org/)
4. Add occluded front
Draw a line at 90 degrees from the start segment of the above generated line
(http://s3.postimg.cc/5ac9wo0k3/add_warm_front.jpg) (http://postimage.org/)
5. Apply Coriolis effect
Adjust the line based on the distance from the start. Modify slightly based on high pressure systems
(http://s15.postimg.cc/ugq7m9jor/apply_coriolis.jpg) (http://postimage.org/)
We now have a fairly good approximation of a real world weather system which we can use to generate local weather conditions.
6. Move local maxima : minima
Weather doesn't stay constant and in a normal mission the weather has plenty of time to change. So let's enable that. Apply velocities to local maxima and minima and regenerate.
(http://s4.postimg.cc/4b0rkv7j1/apply_velocity.jpg) (http://postimage.org/)
So that's my design.
Feedback please
-
from a layman's perspective, it all looks pretty fancy... :D
looks like you're on a good path, for sure....
-
I've tried doing physics based systems before, but they just take tooooooooo loooonnnggg to run.
Even doing the whole thing in shaders, 15 - 20 seconds was common.
So this is an attempt to get weather that FEELS correct for the player without chunking game performance.
-
Stainless:
Just some thoughts:
Are we going to be able to micro-manage local weather? For instance, a series of open fields may give hot spots that make thermals that lead to turbulence. Or, here in the Niagara Peninsula, we have two Great Lakes - Ontario and Erie - that affect the weather quite drastically especially around this time of the year and the spring.
Will we be able to have drastic cold fronts that lead to wind shear? That would make flying a challenge!
I particularly like the moving pressure areas; this would add so much to the immersion and would really require us, as pilots, to be aware of the changes in wind speed and direction as we navigate across long distances.
Off the weather: Are we going to be able have maps as large as the one you show in your example???
Bob
-
I intend to allow you to define local weather variations.
I haven't quite tied down the details yet, but in short.
- Wind shear : Yes
- Terrain induced air movements : Yes
- Icing : Yes
I intend the maps to be as large as you want. I want to be able to have around the world trips.
-
Stainless:
Ahh, man. If you can pull this off watch your back - MS, or rather Steam now, will put a contract out on you 'cause you'll make Flight Simulator X obsolete. And all those mods for FSX that you have to pay for now will be worthless.
Bob.
-
Microsoft have asked me to come to the dark side a couple of times, I have resisted and stayed in the light. ;D
I did go to the Reading (England) office once, but didn't make it past the reception area. They had a massive video wall showing Gates talking rubbish. Couldn't take it. Walked out.
-
Jeez, Stainless ...! o_O
This is a hell of a project ...! :P
Dynamic weather has been missing since the very start, but of course this is logical regarding the CPU load (IL2 was designed in 1999/2000) ...
Maybe you've heard about Checkyersix's C&C effects mod that includes this:
9- Weather
• RandomWeather
• DynamicWeather
• StormFront
https://www.sas1946.com/main/index.php/topic,35826.msg391645.html#msg391645 <=== 4.09 version
https://www.sas1946.com/main/index.php/topic,26654.0.html <=== original version
Have a nice time during your holidays ... 8)
-
What about multilayer clouds? For example stormclouds can make two layers - one at bottom, one at top with "pillars" of original clouds. (But that is highly complex, connected with thermals and such... then stratospheric clouds, fog...
-
Reading about this project I feel myself like an inch -tall dwarf ... ;) :( :P :)
-
What about multilayer clouds? For example stormclouds can make two layers - one at bottom, one at top with "pillars" of original clouds. (But that is highly complex, connected with thermals and such... then stratospheric clouds, fog...
Yes to all .... evntually ;D
-
Remember Icarus, be carefull not to fly too high, those winds may blow you too close to the sun! :)
-
First stage working.
You can place local maxima and minima (high and low pressure zones) and the app computes a fairly realistic pressure map.
From the isobars in this map, flying conditions would be ...... challenging.
(http://s10.postimg.cc/pd09q6qy1/pressure.jpg) (http://postimg.cc/image/wszjbzen9/full/)
-
500 to 1100? Yep, challanging indeed. (It does induce wind, right?)
Now try working on clouds and thermals - if you have speciffied the zones for them to happen - with lower pressure they are stronger and vice versa with high pressure (but you might already know). Above those zones low clouds (150 - 600 m) of small size are created. ;)
Keep up the good work :)
-
Very interesting indeed!
I am a fan of realistic meteorological situations, in particular seasonal variations.
Funny enough I am working on a mission/campaign conversion script which has:
# From the .properties file, returns "Moderate", "Light" or "High" for setting meteorological conditions
WindDescription=`grep -i 'nWeather: ' "Temp_${prop}" | cut -d ' ' -f 7`
# Generates a random windspeed in m/s, within the range defined by the $windstrength. Ranges are bases on modern data
case $WindDescription in
[Ll]ight) WindSpeed=`shuf -i 10-29 -n 1 | sed 's/^\(.\{1\}\)/\1./'`;Gust=`shuf -i 1-5 -n 1`;Turbulence=`shuf -i 1-5 -n 1`;;
[Mm]oderate) WindSpeed=`shuf -i 30-49 -n 1 | sed 's/^\(.\{1\}\)/\1./'`;Gust=`shuf -i 1-8 -n 1`;Turbulence=`shuf -i 1-8 -n 1`;;
[Hh]igh) WindSpeed=`shuf -i 50-80 -n 1 | sed 's/^\(.\{1\}\)/\1./'`;Gust=`shuf -i 1-10 -n 1`;Turbulence=`shuf -i 1-10 -n 1`;;
esac
# Generates a random wind direction based on meteorological data.
# Angles in the array reflect the distribution pattern of wind direction in the Leningrad area (modern data used)
WindDirectionData=("255.0 Westerly" "270.0 Westerly" "275.0 Westerly" "260.0 Westerly" "265.0 Westerly" "280.0 Westerly" "282.0 Westerly" "285.0 Westerly" "90.0 Easterly" "92.0 Easterly" "95.0 Easterly" "100.0 Easterly" "85.0 Easterly" "80.0 Easterly" "180.0 Southerly" "182.0 Southerly" "185.0 Southerly" "190.0 Southerly" "175.0 Southerly" "170.0 Southerly" "225.0 South Westerly" "227.0 South Westerly" "230.0 South Westerly" "220.0 South Westerly" "235.0 South Westerly" "140.0 South Easterly" "137.0 South Easterly" "135.0 South Easterly" "130.0 South Easterly" "310.0 North Westerly" "317.0 North Westerly" "315.0 North Westerly" "320.0 North Westerly" "360.0 Northerly" "358.0 Northerly" "45.0 North Easterly" "50.0 North Easterly")
# Picks a value in the array
WindDirectionDataSelected=${WindDirectionData[$RANDOM % ${#WindDirectionData[@]} ]}
# Assigns angular value to the variable $WindDirection
WindDirection=${WindDirectionDataSelected%% *}
# Assigns string value to the variable $WindOrientation for use in mission briefing
WindOrientation=${WindDirectionDataSelected#* }
# Appends the generated meteorological data to the final outfile
printf "%s\n" [WEATHER] " WindDirection ${WindDirection}" " WindSpeed ${WindSpeed}" " Gust ${Gust}" " Turbulence ${Turbulence}" >> $targ
Lots of fun as you can see!
All the best for your project :)
-
You can also combine this data with the location based atmospheric data to give winter / summer / day / night / latitude / longitude based variations.
The next stage is to generate a wind map.
It's actually counter-intuitive. You would think that the wind goes from high pressure to low pressure, it doesn't. It goes along the isobars.
With a wind map in place you can do things like detect up-drafts caused by low level air packets striking steep terrain.
I need to do some more research into micro-bursts and down drafts before I can add them,(any helpers?) but we are getting there.
-
Okay I have pressure gradient encoded now, a precursor for the wind calculation.
Encoded as a 2D vector. Red is x , green is y.
(http://s18.postimg.cc/4smklhjt5/pressure_gradient.jpg) (http://postimg.cc/image/6x6xmklfp/full/)
-
Mighty impressive Sir!
-
Ha ha..Pardner I think you have really bitten off a big bite there...
I can remember flying the bush in Alaska where I could be at 1000 ft holding a 15 degree right crab, and look down at the snow blowing
from the opposite direction and changing directions whereby it looked like a large snake slithering across the tundra.
Flying across the hills the altimeter would read different altitudes from one side of a hill to the next side..
Weather can be a big factor & i'm wondering if we are ready to throw something like that in the il2 frey...ha ha
could be interesting though...Better send everyone out for an instrument rating though me thinks...ha ha
Have fun though...
-
An excellent (almost) freeware sim with a superb weather engine would be the latest Falcon BMS4.33...
In theory, you can built your own weather there by checking reallife weather and building it into a mission.
Also even the default mission section is mighty powerful with regard to weather, you can fly from sunny to fair to inclement with all sorts of settings within one single mission. You have also multiple cloud layers and diverse turbulence settings. ..weather setup allows more settings than I could comprehend so far.
https://www.youtube.com/watch?v=-tayg9RUlbI
Maybe worth a check.
-
The idea of this is that it's a simple approximation to a very , very complex problem, but it has real effect on the flying experience.
So my existing atmospheric code will be modified to include the new information this code provides.
So the atmospheric variables used in the simulation will be modified by the local weather conditions. This will then automatically effect things like propeller and engine efficiency, pressure based altimeters, jet engine efficiency, fuel use, damn anywhere pressure and temperature are used in an equation.
Then I can easily use the terrain to modify the velocity of the air, producing updrafts and downdrafts.
So now I have a simple approximation to the atmosphere in three dimensions.
AND it actually effects the pilot experience.
Then I can start simulating other more ephemeral effects. Icing, misting, contrails, etc.
As long as there are 36 hours per day and I don't have a life. :D
-
Wind speed computed. Assuming northern hemisphere for the moment, that will change.
Suprised me that the wind direction isn't at 90 degrees to the isobars, but turns out that is correct. Funny how detail comes out of simple code.
(http://s21.postimg.cc/oiwoiaqxz/wind_speed.jpg) (http://postimg.cc/image/5dtf8jc9v/full/)
-
Awesome!!
...just...
AWESOME!!
keep it up :)
-
Stainless:
Unless I'm missing something aren't the winds blowing from the wrong direction for the northern hemisphere? Wind blows in a counterclockwise, or anticlockwise, direction in to a low and clockwise out from a high. Your graphics show what it would look like in the southern hemisphere. I think.
Bob
-
Sorry, you are right.
Easy to fix.
-
Very promising, Stainless ...! :P
Her is a site that shows winds blowing in real time on earth ... ;)
http://earth.nullschool.net/#current/wind/isobaric/1000hPa/orthographic=5.54,46.70,3000
-
Yes I saw that site.
It shows some artifacts I want to re-create. Like the braking effect of the land. Winds tend to be faster over water.
You can also see points where the winds end. They dive into a low pressure system and just die. Really cool.
-
;)
-
First pass at cold fronts, not happy with it, but a good start
(http://s9.postimg.cc/xuxs186e7/test1.jpg) (http://postimg.cc/image/e0bqf3r6j/full/)
-
That looks promising!
:P
Cheers
Thorsten
-
All looks incredible. Thanks for all your hard work on this Stainless.
-
Pure genius.
-
Warm fronts added.
Also added the height map though it's not displaying very well at the moment.
(http://s8.postimg.cc/w51pw6zz9/fronts.jpg) (http://postimg.cc/image/kfxq8890h/full/)
-
Some more shots
(http://s23.postimg.cc/fcmqonprv/heights.jpg) (http://postimg.cc/image/tvtvq2iwn/full/)
(http://s28.postimg.cc/hwnwsfop9/combined2.jpg) (http://postimg.cc/image/lg9ui8rex/full/)
-
AMAZING !! Fantastic !! 8)
-
Did you bother to read the twenty posts that preceded yours in this thread?
-
errr yes. It is still going.
New gui
(http://s28.postimg.cc/8r7kd4eb1/new_gui.jpg) (http://postimg.cc/image/dpv2rni3t/full/)
New equations for the warm fronts
(http://s21.postimg.cc/t7rfvvr53/new_warm_fronts.jpg) (http://postimg.cc/image/ikxmqgizn/full/)
-
Is this project still going.
hi - now, i seriously hope you don't intend to make posts in this forum, which are essentially the same as your username.
please read threads before asking idiot questions.
if the poster has given an update the previous day, it's safe to assume the project is still going.
please, this is your first warning - i don't want to to see any more of this sort of 'nonsense' posts from you.
-
This is a truly fascinating project.
Makes flying more challenging in combination with dynamic weather. :)
One question though... Will there be a weather forecast girl implemented? :D
I mean, nowadays we are so used to see them doing the weather forecast.
(please...)
(http://s2.dmcdn.net/N3KDT/x240-xnj.jpg)
-
;D
I might be persuaded.
-
One question though... Will there be a weather forecast girl implemented? :D
... sure, here she is, ready to announce a cold front and getting Stainless a beer, 'cause he really deserves it ... :P
(http://i55.photobucket.com/albums/g157/micksaf/pa_beer_fridge_zpsdf408f88.jpg) (http://s55.photobucket.com/user/micksaf/media/pa_beer_fridge_zpsdf408f88.jpg.html)
-
Damn I was just about to pose a pertinent question but Micks' last post has taken it out of my head. Anyway Stainless, fantastic work!
-
Excuse me if I'm asking some dumb question, but I was reading these posts about the weather in the new game, and I don't understand "how" it will affect the playability...
I mean for example, the "Weather" will be automatically generated throughout a campaign according to the season, and no more a matter of choice?
-
Excuse me if I'm asking some dumb question, but I was reading these posts about the weather in the new game, and I don't understand "how" it will affect the playability...
I mean for example, the "Weather" will be automatically generated throughout a campaign according to the season, and no more a matter of choice?
Well the tool was designed to test my algorithms. However now I am thinking it can be used to design the weather on a per mission basis.
How will it affect game play?
In the gross case, you could design the weather so a bomber only has a small time window to attack the target before the weather closes in and you lose visibility.
You could design the weather so that a clever player can avoid detection by using the clouds. There are lot's of ways weather can affect a mission.
On the minor scale, this allows me to detect the conditions required for atmospheric events that affect flight. Downdrafts, updrafts, icing, turbulence, fogging, all these features affect you in cockpit.
Just like in real life.
Then there is navigation.
If you fly a straight line from a to b, you will miss b by a long way. The wind will push your aircraft around. I read this book when I was a child.
http://www.amazon.co.uk/Teach-Yourself-Air-Navigation-Kaspar/dp/B000YHCXRW (http://www.amazon.co.uk/Teach-Yourself-Air-Navigation-Kaspar/dp/B000YHCXRW)
It fascinated me.
I suppose it makes me a bit weird, but hey, who wants to be normal anyway.
I am now thinking about how I can get it into IL2 as well.
-
Started adding clouds.
So far I am just doing a simple pass along the weather fronts, no randomness, no terrain effects.
It's not good enough, but it's a start.
(http://s12.postimg.cc/uu8ah8cnh/clouds_start.jpg) (http://postimg.cc/image/z3d0jefwp/full/)
(http://s14.postimg.cc/nzjm4ei1t/clouds_start_2.jpg) (http://postimg.cc/image/5wqjd6m71/full/)
-
Totally reworked clouds.
Now I generate a list of "cloud masses" instead of trying to create a bitmap.
This is good because I can use this directly in a voxel cloud renderer and also I can have cloud type.
Currently supported Cumulus and Stratos, but I just need to work out the conditions required for creation of the other cloud types and I can add them.
Still not using the heightmap to generate localised clouds either, but that's on the way.
(http://s22.postimg.cc/6ncs6bmox/new_clouds.jpg) (http://postimg.cc/image/ajq42b7od/full/)
-
This one modifies the number of cumulus clouds by the local pressure.
It also adds stratos clouds in advance of the warm front.
(http://s2.postimg.cc/jpl32mhix/new_clouds2.jpg) (http://postimg.cc/image/e1esbqd6d/full/)
(http://s8.postimg.cc/lq9yxy9jp/new_clouds3.jpg) (http://postimg.cc/image/vatlktyvl/full/)
-
I've been working on localised clouds, and it's sent me into a world of simulation.
I have added location and time variables to the system and am currently adding the code to get an accurate simulation of the incident radiation from the sun for all points on the map.
With this in place and the pressure values I have already calculated I can get the air temperature at any location.
This allows me to calculate the dew point, and I can estimate the humidity from latitude, map data, and distance from a weather front.
Finally this will allow me to add localised clouds, fog, and icing conditions.
I think I may be going to far, maybe I should have done a weather girl instead....
-
Hi, Stainless,
with your knowledges of meteorology we can wait for using your program for RL local forecast based on the situation copied from TV news screens, huh ?
I admire your work.
-
Four weather features which would be interesting to see in a sim, although probably a bitch to code.
1) Icing. Based on air temp, humidity, local precipitation, etc. The physics of it might be easier to figure out than the graphics for icing on wings, prop, carb, etc.
2) Accurate precipitation. Both in terms of type and volume. Fly into a thunderhead and have your plane get beaten up by melon-sized hailstones. Gain altitude and suddenly the rain you're flying through turns into freezing rain or snow.
3) Cloud and front turbulence. Fly into an 30 km tall anvil cloud and get your plane flipped over and torn apart by updrafts. Fly along the point where two fronts collide and experience clear air turbulence.
4) Wind interaction with terrain. Updrafts and downdrafts near mountains, terrain which blocks crosswinds on landing or take-off, daily onshore and offshore breezes in coastal areas. This would allow realistic air effects when flying through/over mountains or along coasts. Recreate Steve Fosset's last flight over the Sierra Nevadas, supply flights over "The Hump," or Alaskan bush pilot flights.
-
1) Icing
Planned. I know how to do it, how to apply the effects to an aircraft, just need the weather code to give me the correct variables.
2) Precipitation
In progress. The work I am doing at the moment allows me to calculate the dew point for any given place on the planet at any given date. From the dew point and the data I already generate in the weather system we can get precipitation strength. I then need to add the temperature from the atmospheric model and we have precipitation type.
3) Cloud and front turbulence.
In progress. This was the real reason I decided to do a weather simulation.
4) Terrain effects
In progress. Follows on from the same code above.
-
Haven't posted in a while, so thought I had better update y'all.
I now have code in place that takes latitude,longitude,time and date and returns the incident radiation from the sun.
I also have code in place that converts every pixel on the map to a latitude longitude pair.
When you create the package you define the latitude and longitude of the top left of the map, and the start date and time.
So finally after a lot of coding I can generate the heating effect of the sun for each pixel in the map.
I can use this along with the atmospheric data to give a fair estimate of the ambient temperature at every point on the map.
Lot of work, but getting there.
-
Very impressive.
It looks like you understood the client brief very well (or that the brief itself was quite clear!!!).
Oooops ... same person!
-
I have been quietly following your posts. You have a lot on your hands and it appears to be a great sim. I was wondering how far along are you at this time and do you have a demo, as in sneak peaks via you tube or a demo game yet??
Thanks Joe
-
Not at that stage, it's a bunch of systems plugged together missing key ingredients.
When I have the weather system working (which is close now) I will plug a lot of it together into a demo.
-
Finally some progress to show you.
I am calculating a relative temperature map based on the terrain type and the amount of solar radiation striking the land. The display is just a relative temperature. Red is warmer than green.
The green value is calculated by taking the sin of the scaled temperature. So it peeks in the middle of the temperature range. The red value is just the scaled temperature.
(http://s28.postimg.cc/ku5csakl9/temperature.jpg) (http://postimg.cc/image/rklu1q7qx/full/)
Now I can combine this with pressure values to generate location based clouds.
-
Excellent work Stainless :). I am excited :D
Klaus109
-
Stainless,
While everything you are doing is very impressive and complicated, I am concerned what type of computer we will need to run all the math. Some of us have updated systems, but I am sticking my neck out here, I believe and IMHO, most of us do not have state of the art systems. Do you know what the specs will be based on what you have now for running your sim. Video card, processor, ram memory?
Regards,
Joe
-
The whole code base is designed to be configurable. You can turn on and off parts of the simulation until you achieve the desired frame rate.
-
Seems like you have an incredibly solid grip on what needs to be done to get weather effects right. I'm massively impressed! :)
Dumbass question: Would it be possible to plug historical weather variables into the weather engine and get something that's halfway accurate?
Believe it or not, there's actually a lot of historical weather data from WW2, at least for Western Europe, although it's incomplete by today's standards. It would be amazing if, for example, you could grab a partial weather report from the UK Met office, for August 17, 1940 for Southwest England, plug it into a game engine, and get weather output that matches meteorological conditions that pilots faced on that day.
It might seem like trivia, but the weather was different back then, and the WW2 years were climatically strange, with some unusually cold and wet winters.
-
You can stick in real world pressure readings and it will make a fair approximation of what really happened.
I ran a test the other day with data from a UK weather satellite.
It wasn't by any means 100% accurate, but it was generally correct.
The rain storm in my simulation arrived at Portsmouth at 13.10, in real life it arrived at 12.42
I will live with that.
The alternative is to ship the game with a Cray II super computer per user, not a good idea I think.
-
You can stick in real world pressure readings and it will make a fair approximation of what really happened.
That sounds very good.
I'm thinking that as long as you can input an air pressure, air temperature, humidity, and 10ths of cloud cover reading and get output that's halfway realistic, that's should be fine.
Perhaps historical accuracy could be improved a bit if you have multiple pressure, temperature and humidity readings, wind speed/direction, cloud base and precipitation, but it's obviously a flight sim, not a weather simulator.
Weather conditions for a particular time and place could be set up as preset files by 3rd party contributors.
No supercomputer needed, although desktop computers are getting there, at least by historical standards.
-
Was putting together a demo and found a massive bug.
Video quality sucks though
https://www.youtube.com/watch?v=1UMJU0oS8_4&feature=youtu.be
-
Nice! :)