Live internet weather
-
Topic authorCaptain_Avatar
- Posts: 6
- Joined: 29.10.2004
- With us: 20 years
Live internet weather
A nice feature for Earth would be a cloud layer derived from GOES satellites. Celestia would download the latest graphics from the internet and draw them onto the globe. This could also be done on a less frequent basis for the other planets where we have some cloud data.
- PlutonianEmpire
- Posts: 1374
- Joined: 09.09.2004
- Age: 40
- With us: 20 years 2 months
- Location: MinneSNOWta
- Contact:
-
Topic authorCaptain_Avatar
- Posts: 6
- Joined: 29.10.2004
- With us: 20 years
RE: Live internet weather
PlutonianEmpire wrote:Great, but what happens when we increase the time flow?
Well, since there is no such thing as a reliable weather time machine several things could happen:
1) The last set of clouds could rotate around the globe as they currently do.
2) A lower res set of default clouds could pop up.
3) A cache of cloud pics could cycle through.
-
- Posts: 835
- Joined: 27.09.2004
- With us: 20 years 1 month
- Location: Massachusetts, USA
just need today's clouds
http://www.die.net/earth/ and http://www.mapmaker.com/shadowfacts/ourworld.asp
have current cloud cover info, but I do not know where they get it.
all you would need is a new cloudmap each day, right?
have current cloud cover info, but I do not know where they get it.
all you would need is a new cloudmap each day, right?
Homebrew:
WinXP Pro SP2
Asus A7N8X-E Deluxe
AMD Athlon XP 3000/333 2.16 GHz
1 GB Crucial RAM
80 GB WD SATA drive
ATI AIW 9600XT 128M
WinXP Pro SP2
Asus A7N8X-E Deluxe
AMD Athlon XP 3000/333 2.16 GHz
1 GB Crucial RAM
80 GB WD SATA drive
ATI AIW 9600XT 128M
-
Topic authorCaptain_Avatar
- Posts: 6
- Joined: 29.10.2004
- With us: 20 years
Would be a good start....
Once,twice or 4 times a day would be acceptable. Just so the clouds would change on a regular basis.
-
- Posts: 19
- Joined: 29.10.2004
- With us: 20 years
- Location: Maarssen, Netherlands
-
- Posts: 19
- Joined: 29.10.2004
- With us: 20 years
- Location: Maarssen, Netherlands
You should also take a look at:
http://www.satsignal.net
This site is offering as set of satellite imaging tools to plot, animate and rotate geostationary weather satellite images and polar orbiters as well!
Cheers,
Ton Lindemann
http://www.satsignal.net
This site is offering as set of satellite imaging tools to plot, animate and rotate geostationary weather satellite images and polar orbiters as well!
Cheers,
Ton Lindemann
Internet Weather
You can download the latest world images and import them directly into Celestia.
http://www.meteo-maarssen.8m.com/data/celestia/
Visit the above page (Note there are two types of images listed daily.)
Download the lastest (Non IR) image. Rename your earth.jpg to earth-original.jpg and put the downloaded one in its place, and start Celestia'ing..!!
http://www.meteo-maarssen.8m.com/data/celestia/
Visit the above page (Note there are two types of images listed daily.)
Download the lastest (Non IR) image. Rename your earth.jpg to earth-original.jpg and put the downloaded one in its place, and start Celestia'ing..!!
-
- Posts: 19
- Joined: 29.10.2004
- With us: 20 years
- Location: Maarssen, Netherlands
- t00fri
- Developer
- Posts: 8772
- Joined: 29.03.2002
- Age: 22
- With us: 22 years 7 months
- Location: Hamburg, Germany
Hi all,
I have spent quite a few thoughts whether "live clouds" are within reach. Unfortunately, I am now rather pessimistic. Yes you can wrap some of the above images around the globe, but I gather, most of you are meanwhile quite spoiled as to the great possibilities of cloud display in Celestia!
The people advocating "live clouds" must take into consideration that usually there is a substantial and notably nontrivial sequence of image manipulation steps involved, before a "raw" cloud cover texture from a satellite looks as /natural/ in Celestia as most of you are used to. For example, many people have used or still use my "2k clouds". The DXT3 version is 2.7 MB (!!). Compare the size of the 1k clouds above, for example...
The steps involved to reach that nice "fluffy" appearance are not suited for automatization, unfortunately.
One has to handle/incorporate transparency, i.e. appropriately fill the alpha channel and so on.
Then one usually has to convert the clouds into DXT3 format, since various other texture layers are to be displayed as well ...(night lights, specular light, main texture, normal maps) and so one wants to be as fast as possible.
So no doubt, a decision for "live weather" would be a great compromise in quality.
Bye Fridger
I have spent quite a few thoughts whether "live clouds" are within reach. Unfortunately, I am now rather pessimistic. Yes you can wrap some of the above images around the globe, but I gather, most of you are meanwhile quite spoiled as to the great possibilities of cloud display in Celestia!
The people advocating "live clouds" must take into consideration that usually there is a substantial and notably nontrivial sequence of image manipulation steps involved, before a "raw" cloud cover texture from a satellite looks as /natural/ in Celestia as most of you are used to. For example, many people have used or still use my "2k clouds". The DXT3 version is 2.7 MB (!!). Compare the size of the 1k clouds above, for example...
The steps involved to reach that nice "fluffy" appearance are not suited for automatization, unfortunately.
One has to handle/incorporate transparency, i.e. appropriately fill the alpha channel and so on.
Then one usually has to convert the clouds into DXT3 format, since various other texture layers are to be displayed as well ...(night lights, specular light, main texture, normal maps) and so one wants to be as fast as possible.
So no doubt, a decision for "live weather" would be a great compromise in quality.
Bye Fridger
-
- Posts: 19
- Joined: 29.10.2004
- With us: 20 years
- Location: Maarssen, Netherlands
To all,
I'm sorry, but life weather in 3D is impossible, it is a dream that would be forever. It is not available. Not in any forecast model and not from satellite.
Satellite data is always 2D; you can't make 3D from it. Why: IR is only mapping surface temperatures it can see, not mapping the temperature for each level below that surface (=cloudtops). Any satellite channel (IR, visual or water vapour) can't look trough the clouds. If you see a bright spot the satellite then it could be thin high level cirrus, but also the cirrus of a thunderstorm. How would you see a 3D cloudtexture where thin cirrus appears as a thick cloud from a thunderstorm? Far from realistic.
Other problems: IR can't discreminate clouds from snow and ice or cold surfaces. This means that it is tricky to make realistic cloud textures during wintertime or over te polar regions. It seems that it is always overcast there, even with clear skies. Also for the tropics; warm low level clouds disappears in the radiation heat of the background. This can be seen om my versions too. A have a treshold for clouds if the temperatures drops below -10°C.
Clouddata is even not even presented in forecastmodels, but are calculated from humidity on the atmospheric layers. Public worldwide nummerical forecast models as AVN/GFS/MRF have humidity data for 925, 850, 700, 600, 500, 400, 300 hPa levels, but only for a 0.5° grid. Clouds can be calculated from that data, but only in a course resolution. It should be possible to calculate 3D cloudtextures from this data, but I don't have the time, experience, knowledge and tools for it. If anybody would like to create realistic looking textures, but based on course grid and with poor details then repley and I will make a dataset for each level to play with. If this works, then it should be possible make cloudtextures back to 1 January 1948 in steps of 6 hours a day!
Again this kind of data is not available, but you may ask NOAA to implementate Celestia output it into the modelruns. I don't believe they do so.
Cheers,
Ton Lindemann
weather amateur in the Netherlands
I'm sorry, but life weather in 3D is impossible, it is a dream that would be forever. It is not available. Not in any forecast model and not from satellite.
Satellite data is always 2D; you can't make 3D from it. Why: IR is only mapping surface temperatures it can see, not mapping the temperature for each level below that surface (=cloudtops). Any satellite channel (IR, visual or water vapour) can't look trough the clouds. If you see a bright spot the satellite then it could be thin high level cirrus, but also the cirrus of a thunderstorm. How would you see a 3D cloudtexture where thin cirrus appears as a thick cloud from a thunderstorm? Far from realistic.
Other problems: IR can't discreminate clouds from snow and ice or cold surfaces. This means that it is tricky to make realistic cloud textures during wintertime or over te polar regions. It seems that it is always overcast there, even with clear skies. Also for the tropics; warm low level clouds disappears in the radiation heat of the background. This can be seen om my versions too. A have a treshold for clouds if the temperatures drops below -10°C.
Clouddata is even not even presented in forecastmodels, but are calculated from humidity on the atmospheric layers. Public worldwide nummerical forecast models as AVN/GFS/MRF have humidity data for 925, 850, 700, 600, 500, 400, 300 hPa levels, but only for a 0.5° grid. Clouds can be calculated from that data, but only in a course resolution. It should be possible to calculate 3D cloudtextures from this data, but I don't have the time, experience, knowledge and tools for it. If anybody would like to create realistic looking textures, but based on course grid and with poor details then repley and I will make a dataset for each level to play with. If this works, then it should be possible make cloudtextures back to 1 January 1948 in steps of 6 hours a day!
Again this kind of data is not available, but you may ask NOAA to implementate Celestia output it into the modelruns. I don't believe they do so.
Cheers,
Ton Lindemann
weather amateur in the Netherlands
-
- Posts: 128
- Joined: 26.07.2004
- With us: 20 years 4 months
- Location: Earth 7.593358long / 47.582393lat
- Contact:
I'm sorry, but life weather in 3D is impossible, it is a dream that would be forever
3D? As for now 2D would be top of the line.
It should be possible to calculate 3D cloudtextures from this data,
IMHO it is a tricky job and a lot of peoples would profit from 2D images and may be lores 3D cloudmaps. I think this needs a new OpenSource project for it self. May be I start this if I finish the shadowremoval pythonscript.
http://shatters.net/forum/viewtopic.php?p=45903#45903
I would love to see "live" cloudtextures on Earth! But to be real cool it has to be animated for the past (old cloudtextures).
Thanks for the Idee, Dominik
ASUS A7N8X Deluxe
AMD Athlon XP 2800+ (2.08GHz)
1GB DDR RAM 333MHz
NVIDIA GeForce FX 5600 AGP 8X
AMD Athlon XP 2800+ (2.08GHz)
1GB DDR RAM 333MHz
NVIDIA GeForce FX 5600 AGP 8X