This Thread gives me a sinusoidal headache.
Why is an accurate global map hard to make?
- Thread starter Trekker4747
- Start date
What's worse, the Earth deviates quite a bit from being truly spherical.
You can have maps that correctly show the area of land masses.
Lambert Cylindrical Projection
The problem with map projections is because of the earth's curvature.
Imagine if you draw a circle on the earth with radius 1000 km centered about your house. Because the earth is curved, that circle isn't going to be 2×π×1000 km in circumference; it's going to be a bit less, call it length L. So how do you represent that circle on a flat map? You have only four options :
(i) draw that circle 1000 km from your house, such that it's circumference is 2×π×1000 km.
(ii) draw that circle closer to your house, such that it has correct circumference L, so shrinking the area inside the circle a bit.
(iii) draw a circle somewhere between these two, that doesn't represent either measurement (the radius nor the circumference) correctly.
(iv) draw something that isn't a circle, and introduce shape distortion to your map.
Repeat for the 2000km circle, the 3000km circle, etc.
That's a good one, but I prefer the Winkel Tripel. That's the one used by the National Geographic Society.
It does make Antarctica look pretty huge, but that's okay--you've got to hide all those lost cities and shoggoths somewhere. Plus, it's more fun to say "Winkel Tripel." These are important considerations.
The Gall-Peters projection is too weird for me. And I didn't realize, until I opened this thread, just how long it had been since I'd seen an actual Mercator world map. Those really are bad.
I've thought of one way of making a perfect map, but it involves fractals, and the map will have an infinite number of cuts in it. 
See how the dodekahedron can be flattened out:
Apply this idea to a polyhedron whose facets are being iteratively subdivided ad infinitum, tending to a sphere.
.
See how the dodekahedron can be flattened out:
Apply this idea to a polyhedron whose facets are being iteratively subdivided ad infinitum, tending to a sphere.
.
Last edited:
The first example they gave was if you wanted to flatten out an orange rind on a table, it would split itself into wedges, looking like this:
This has always been my favorite solution.
Mostly because it's somewhat similar to how globes are made and I think that's a good way to mentally think about it.
See 1:40: http://www.youtube.com/watch?v=qgErv6M19yY
I guess you have never tried stitching a panaroma photograph together in Photoshop? Same thing. You can't go from spherical to a flat plane without distortions.
I've thought of one way of making a perfect map, but it involves fractals, and the map will have an infinite number of cuts in it.
See how the dodekahedron can be flattened out:
Apply this idea to a polyhedron whose facets are being iteratively subdivided ad infinitum, tending to a sphere.
Also see
http://www.progonos.com/furuti/MapProj/Normal/ProjPoly/projPoly3.html
and that site generally (someone might have already linked it)
http://www.progonos.com/furuti/MapProj/Normal/TOC/cartTOC.html
I like straight spherical projection. Because then I can use it as a texture map
The point of a map isn't to be realistic, it's to be useful.
Since different people have different needs, it's not surprising there will be different projections, with different types of detail overlaid. Provided people pick the correct one for their needs, who actually cares if they're realistic?
Since different people have different needs, it's not surprising there will be different projections, with different types of detail overlaid. Provided people pick the correct one for their needs, who actually cares if they're realistic?
I like straight spherical projection. Because then I can use it as a texture map
That's not a good way of texturing a sphere, even though it's commonly done that way : Too much texture and geometry is squashed into the poles, while it's stretched thin along the equator.
Earlier this year, I helped someone who was having problems with their project because of how spheres are normally textured. If you want to add topography to the sphere, and then look at it close up, you'll see what those problems are.
The best al-round method I came up with was to use something similar to a skybox: six square textures radially projected onto a sphere that is defined with similar geometry.
Good, then, problem solved.
What I want to know is.
If our mighty human brains can take the image that gets turned upside down by our lens in our eyes and make everything appear upright and normal, then my question is, what way are we really facing?
I am blinded by my human physiology. Hehe.
If our mighty human brains can take the image that gets turned upside down by our lens in our eyes and make everything appear upright and normal, then my question is, what way are we really facing?
I am blinded by my human physiology. Hehe.
I like straight spherical projection. Because then I can use it as a texture map
That's not a good way of texturing a sphere, even though it's commonly done that way : Too much texture and geometry is squashed into the poles, while it's stretched thin along the equator.
Earlier this year, I helped someone who was having problems with their project because of how spheres are normally textured. If you want to add topography to the sphere, and then look at it close up, you'll see what those problems are.
The best al-round method I came up with was to use something similar to a skybox: six square textures radially projected onto a sphere that is defined with similar geometry.
Interesting idea. I can't say I've ever seen an Earth texture using that projection, although it would give the most undistorted mapping. All I ever see is the spherical textures. Your method does leave more seams, so it would take more care to make sure those aren't visible, whereas they're fairly tucked away with spherical projection. But done right it would eliminate the problems, and the geometry is easy enough to create (at least, in my 3D software). It would also give a better sense of relative size, because it remains fairly flat/planar.
In terms of general modeling and texturing, I still find that the basic spherical mapping and latitude/longitude geometry sphere is the easiest to work with all round, even taking into account the problems at the poles. It's really just lucky that the north and south poles of Earth have very little detail, because that type of sphere doesn't map neatly at the poles because it's triangles, and not quads.
To make a compatible texture, you could use a higher resolution spherical texture and resample it for the new geometry.
An alternative source of textures would be those from a parallel projection:
(x,y,z) --> (x,z) produces an image like this:
^ this can easily be projected out onto the skybox with very little distortion. You'd do that to create you skybox textures, and then lose the parallel textures.
The seams are very tidy, because they're along the edges of the skybox, which has nice coordinates. eg, the seam between the +x and +y planes will be along the line from (+1, +1, -1) to (+1, +1, +1).
You can sew those quads together seamlessly.
Well, you've sold me. I'll definitely keep this in mind next time I'm doing planetary projections. It's a little bit more setup work, but definitely worth it. I say all future world maps should be projected onto cubes!
I skimmed most of this thread so forgive me if it got any mention.
Anyway, why can't we use the 100 or so satellites, (you know, the ones that Big Brother owns) that we have sitting in orbit looking down on our planet to take pictures in various areas and then create a composite picture of the end result to make a truly accurate map? Isn't that what most of them are for anyway?
Anyway, why can't we use the 100 or so satellites, (you know, the ones that Big Brother owns) that we have sitting in orbit looking down on our planet to take pictures in various areas and then create a composite picture of the end result to make a truly accurate map? Isn't that what most of them are for anyway?
GoogleEarth lied to me?
I'm joking, but, I'm sure, with all the space travel (I've seen An Inconvenient Truth (2006)). It's 2011 A.D - I'm pretty sure that someone, somewhere knows what the planet really looks like, even if our weather maps off of the Equator do get distorted, I guess, look at the U.N Flag, for a flat version of it.
I'm joking, but, I'm sure, with all the space travel (I've seen An Inconvenient Truth (2006)). It's 2011 A.D - I'm pretty sure that someone, somewhere knows what the planet really looks like, even if our weather maps off of the Equator do get distorted, I guess, look at the U.N Flag, for a flat version of it.