Page 1 of 2 12 LastLast
Results 1 to 10 of 16

Thread: Aerth

  1. #1

    Wip Aerth

    First off, prior to starting work on this map I had no experience in cartography whatsoever.

    As I stated in my introduction, I want to make a world for an upcoming original story.

    Following my search for an automated method, I found this planet generator.
    The setup is fairly easy: you pick a seed, a color scheme, and a projection; the program then renders the planet the same size you chose.
    The colors are height dependent, having two gradients (one for sea, one for land).
    There are a few more parameters, but that's not relevant for this thread.
    The offline (command line) version has even more parameters. I've experimented with this program for well over 3 weeks before getting something I could work with.

    At this point, I should probably note what size I've been working with.
    The idea was to have a printable A0 sized map. According to the math that's:
    The paper: 118,9 x 84,1 cm @ 300 DPI
    The image: 14043 x 9933 pixels


    The original plan was to use the program to generate something that could be used as a guideline for a finalized version. As time passed and hours were spent drawing the basic coastline by hand, I realized that the amount of work was simply too much for me to enjoy it. Largely because progress was extremely slow and I wasn't making anything I could be happy with.

    So I started a search for cartography tutorials and found this tutorial by Tear.
    Now at least I had something to help me make what I could be happy with.
    I've been using Photoshop for roughly 15 years now and can say I could easily follow the tutorial (even though it has some inconsistencies).

    Those who have tried to create something using this tutorial know there's quite a bit of manual work involved.
    Here's where my method slightly deviates from the tutorial:
    Every alpha mask basically is a grayscale. The map that I had generated can be split into a sea-map and a land-map, then converted to a grayscale map, which can then be used as an alpha mask.
    I've written the details in a tutorial of my own (of course, heavily based on Tear's).
    The process of converting and realistically using the generated data for the map took another month or two.
    I've rebuilt the map at least thrice by this point; the final time to proof-test my own tutorial to check that I didn't miss any steps, methods or other information.
    The PSD had to be split in two (sea and land) in order to be relatively fast to process (the current versions are 2,5GB and 3,2GB) and eat away at least 6GB of RAM.
    You can imagine that even running a fast function like 'expand selection by 150px' on the full land-mass can take a good 40 seconds.

    So with the map now finally created, there was still something that had to be created: rivers... the one thing I could not generate automatically.
    At this point, there were two choices:
    - hand-draw the rivers like the tutorial says
    - create a program that uses the height map to flow a river downhill (spawn points are manual)
    I decided the latter.
    Using the same program I used to generate my world map, I created a integer based height matrix. It's basically a giant matrix where each number indicates the height of that pixel.
    The resulting program unfortunately enjoys the fairly fluctuating height map around the higher hills and mountains, drawing nearly single pixel rivers, but severely suffers from the flat plains that eventually follow. This results in more lakes than rivers.

    A little bit of insight on the general scale of things:
    As I used a program to generate a map, I know the exact size in degrees.
    The large chunk of the world I use, is 4600px for 30 degrees. To keep things mathematically simple, I decided to make 1 degree equal to 100km. This makes the whole planet (360*100km) 36000km in circumference, just slightly more than 10% smaller than our Earth (roughly 40000km in circumference).
    Going back to 30deg = 4600px, 1 degree would be (4600/30) 153,33px. This in turn makes each pixel (30*100km / 4600px ) a rough 652 meters.

    Having traveled enough around Europe and Scandinavia, I know no rivers that actually span that wide, so having a river wider than 1px would be undesirable.

    TLR
    Without further ado, I present you the result:
    Aerth

    Do not worry about having to load the full 14043x9933 image. As a programmer and website builder, I've run a program that has split the giant image (275MB) into nearly 3475 pieces of 194x194 pixels each.

    As for why the 'WIP' tag:
    - Not a single region, country, river, city (etc), ocean, sea label
    - No country borders
    - No rivers
    - Only rough estimates of biomes

    The two attached images are what the original render was (generated) and what I made of it using the Saderan tutorial by Tear. They are only a part of the full scale, but show what I started with and what the result is.

    I'd love to hear your comments on what I've created, including the things that don't make any sense (such as the size, location and border of the large desert).
    Attached Images Attached Images

  2. #2

    Default

    Maybe because I haven't included the full map, other than through a link to my own webspace (shikotei.com/Map/index.html), that you guys can't or won't respond.
    So I've added a mini-map (10% of full scale) to this post.
    Each square on the grid on this map is 500km x 500km. Each 500km also represents 5 degrees latitude/longitude.

    My main concern is the location of the desert which is positioned slightly above the equator (5 to -2.5 degrees).
    The cold region in the north starts around 30 degrees. This would place roughly 2500km between the north of the desert and the south of the cold region. This should be roughly the same as the (pure single longitude) distance between the northern deserts in Africa and the southern regions of Siberia, making this plausible... right?
    Attached Images Attached Images
    Last edited by Shikotei; 10-27-2015 at 02:24 PM.

  3. #3
    Banned User
    Join Date
    Jul 2014
    Location
    Traverse City, Michigan, USA
    Posts
    2,547

    Default

    Hello Shikotei. Don't be worried if people don't respond quickly, it can be for a lot of reasons, such as busy working, focusing on other active threads, or even sleeping

    On Earth, deserts occur primarily between 20 and 30 degrees N and S latitudes, so 5 degrees above the equator is a little close. Also, whatever latitude you use, I'd suggest the desert extend to all land areas at that latitude. It's especially noticeable how the desert is only on the center part of the map... I'd expect it to occur also on the western portion.

    Nice start on your map! Have some rep for your first map post!

  4. #4
    Guild Artisan Pixie's Avatar
    Join Date
    Jan 2014
    Location
    Lisbon
    Posts
    939

    Default

    Hi there,

    Your method of creation for this map is quite impressive, indeed. But there's a few things I should point out, for improvement:

    - at planetary level, any random/blob method for creating landforms will look, well, blobby. Still, your planet isn't the worst case, by far, and the landforms are pretty believable. The placing of mountains, though, isn't.

    - climate wise, if this planet is similar in conditions to Earth, you're getting a lot of things wrong. Simply put, the equator is the greener area of the planet and it's the tropics (23º south and north) that get the deserts - terms like "tropical forests" are confusing, I know.

    - you have a serious misconception about linear scale vs. graticule. One degree east or west close to the poles doesn't span as much land as one degree at the equator, that's because you still need to divide the circle at that latitude by 360 degrees. An equiretangular map like this is very useful, but it seriously distorts dimensions. My suggestion is that you save your map at low resolution and upload as image overlay in Google Earth. There you can measure linear distances.

    Despite all these things, you've got a nice start, I'd be proud to have this as a starting point. Congrats! And, by the way, the site you linked doesn't seem to be functioning right, I gave it a look with both Chrome and Opera.

  5. #5

    Default

    I wasn't especially worried; I just hoped 4 days and 220+ views would have been enough for a response. I thought maybe it lacked information, so I added some.
    No accusation of any sort was being made

    Anyway, getting back on track;
    As for why the desert is bound by the land (and thus not fully spreading across to the west half), my reasoning was as follows:
    Earth has this constant wind high in the atmosphere: the jet-stream, if my memory serves me well. One of the biggest carriers of clouds and therefor most rain. Of course there are still other streams caused by atmospheric pressure differences, but let's keep to this main one.
    Most of the clouds on Aerth drift from east to west and, around the equator, have a big ocean to form clouds, storms, and other wet phenomena. These hit the land on the east coast and start releasing their energy, causing rainfall, which results in lush green regions.
    Am I still making sense?
    This continues until these clouds float over the mountain range some 600km inland. The lower wind flows push the air up, causing these clouds to get pushed into higher atmospheres, cool down, condense and again let their energies out in the form of rain and snow.
    Almost all the clouds are dissipated like this because of the loss of energy and, by the time they get past these mountains (if they even make it), they have very little water left. They either fully disperse or let out a bit of rain, probably enough to sustain the extreme forms of life (cacti, desert creatures, and other low-moisture-requiring lifeforms).
    Again, I've no experience in meteorology, so this is just a train of thought without much scientific base to rely on.

    The reason the desert ends near the sea almost splitting the biggest continent in half, that's taken from the example set by the sea that separates a part of Russia from Kazakhstan; the Caspian Sea.
    The distance between the coasts range from 400km to 300km (rough estimates).
    Now I don't know the science behind it, but as I see it, it is possible. Though I'll agree with you that it wouldn't hurt to show some desert-forming near the east coast of the sea.

    The same logic can also be confusing though, just by looking at satellite images:
    The lands of Saudi Arabia, Yemen, Oman and East Arabia are deserts directly west of a large body of water.
    The lands of Vietnam, Laos, Cambodia and Burma are not deserts, but also directly west of a large body of water.
    One could appoint the difference due to the shape of the land (mountainous) or to something of previous eras, I don't know.

    As for the risky (pushing the 'standard' limit) position of the Equator, I could just place it lower on the map. This would set the desert at 17.5-25 N and the cold regions starting at 50N (roughly the south of Canada and Alaska)

    Assuming things is easy; maybe just as easy as assuming them wrong. The problem lies in knowing the difference.
    I'm willing to research the shebang out of it, but only if I know what to look for. Feel free to lay the science smack down.

    -- During the writing of this post, Pixie placed a reaction, so I'll elaborate to that too.

    Quote Originally Posted by Pixie
    - at planetary level, any random/blob method for creating landforms will look, well, blobby. Still, your planet isn't the worst case, by far, and the landforms are pretty believable. The placing of mountains, though, isn't.
    Ah yes, that's due to the generator not knowing a thing about tectonics. It uses a tetrahedron to form land, so the original mass was mostly one giant mountain. After more lurking around, and seeing others start out with a coastline, followed by plate defining, I had already planned to make a plausible guess at drawing the plates and decide the direction they're going, as well as where the land goes up and down into the earth..
    I hope to have something by the end of tomorrow.
    Quote Originally Posted by Pixie
    - climate wise, if this planet is similar in conditions to Earth, you're getting a lot of things wrong. Simply put, the equator is the greener area of the planet and it's the tropics (23º south and north) that get the deserts - terms like "tropical forests" are confusing, I know.
    Assume the equator the lowest line on the map and this should fix that a bit. Truth be told, I never realized that the equator was not the 'designated desert' zone.
    Quote Originally Posted by Pixie
    - you have a serious misconception about linear scale vs. graticule. One degree east or west close to the poles doesn't span as much land as one degree at the equator, that's because you still need to divide the circle at that latitude by 360 degrees. An equirectangular map like this is very useful, but it seriously distorts dimensions. My suggestion is that you save your map at low resolution and upload as image overlay in Google Earth. There you can measure linear distances.
    As long as the latitudes are close to the equator, there's roughly little difference. Yes I'm aware that the distance between 90N0E and 90N180E is zero, but for the sake of not requiring a calculator or a scale every 10 degrees latitude, assuming equal distances was simplest to do.
    I have the math around here somewhere to accurately calculate the distance between two points on a sphere, using the longitude and latitude of the two points, and the radius of the sphere.. as the bird flies at least.

    Weird that the site does not work... do you have JavaScript active? I just tried using Firefox and it seems fine to me.
    IE did blow up on me. I'll take a look and make it work cross-browsers. Thanks for the heads-up!

  6. #6
    Guild Artisan Pixie's Avatar
    Join Date
    Jan 2014
    Location
    Lisbon
    Posts
    939

    Default

    Quote Originally Posted by Shikotei View Post
    Assume the equator the lowest line on the map and this should fix that a bit. Truth be told, I never realized that the equator was not the 'designated desert' zone.
    As long as the latitudes are close to the equator, there's roughly little difference.
    My bad, I assumed this was a whole-planet map, 90ºN to 90ºS and 180ºW to 180ºE, and I just realised that isn't the case. Ignore at least half of what I wrote.

  7. #7

    Default

    Quote Originally Posted by Pixie View Post
    My bad, I assumed this was a whole-planet map, 90ºN to 90ºS and 180ºW to 180ºE, and I just realised that isn't the case. Ignore at least half of what I wrote.
    A full-planet map would have been (4600px/30º*360º) 55200px by 27600px at the current scale.
    The map I'm making is 14032px by 9922px, which would be 91.51º by 64.71º.

    --edit
    Fixed the external website. It should now work on Firefox, Chrome and IE.
    I haven't tried Opera, but it screws up in Safari (5.1.7) because of the way DOM objects are added (in reverse...).
    Last edited by Shikotei; 10-28-2015 at 05:00 AM.

  8. #8

    Default

    Excuse the double posting, but I felt the additional information was a valid enough reason to add a new post rather than run the risk of it going unnoticed as an edit.

    Two of the added images contain the (initial) tectonic plates and the relevant information. I've pointed the general direction (dark blue) of each plate, the local direction (light blue arrows) and their resulting activity (red).
    Trying to make some sense of the existing mountain formations, I tried to create the plates based on these, rather than the reverse.
    If these mountains make no sense after this, it might be time to remake the map (don't worry, this doesn't involve THAT much), so I could accept a verdict the likes of 'maybe you should build upon the plates and remake a large part of your map'.
    One image includes only the coastline, while the second has the map included.
    Both of them have a little legend to explain the what's what.

    The third image is an expanded version of the earlier posted one. I have calculated the distance between two points on the same latitude with a 5 degree difference in longitude (as is the size of the grid).
    I have also marked the equator as a red line.

    The math used (in Excel):
    ACOS(COS(RADIAN(90-latA))*COS(RADIAN(90-latB)) + SIN(RADIAN(90-latA))*SIN(RADIAN(90-latB))*COS(RADIAN(lonA-lonB)))*radius
    Where:
    - latA is the latitude of point A
    - lonA is the longitude of point A
    - latB is the latitude of point B
    - lonB is the longitude of point B
    - radius is the radius of the planet (mine is 36000/(2*pi)
    All points are in degrees.
    Attached Images Attached Images

  9. #9

    Default

    Quote Originally Posted by Shikotei View Post
    Excuse the double posting, but I felt the additional information was a valid enough reason to add a new post rather than run the risk of it going unnoticed as an edit.

    Two of the added images contain the (initial) tectonic plates and the relevant information. I've pointed the general direction (dark blue) of each plate, the local direction (light blue arrows) and their resulting activity (red).
    Trying to make some sense of the existing mountain formations, I tried to create the plates based on these, rather than the reverse.
    If these mountains make no sense after this, it might be time to remake the map (don't worry, this doesn't involve THAT much), so I could accept a verdict the likes of 'maybe you should build upon the plates and remake a large part of your map'.
    One image includes only the coastline, while the second has the map included.
    Both of them have a little legend to explain the what's what.

    The third image is an expanded version of the earlier posted one. I have calculated the distance between two points on the same latitude with a 5 degree difference in longitude (as is the size of the grid).
    I have also marked the equator as a red line.

    The math used (in Excel):
    ACOS(COS(RADIAN(90-latA))*COS(RADIAN(90-latB)) + SIN(RADIAN(90-latA))*SIN(RADIAN(90-latB))*COS(RADIAN(lonA-lonB)))*radius
    Where:
    - latA is the latitude of point A
    - lonA is the longitude of point A
    - latB is the latitude of point B
    - lonB is the longitude of point B
    - radius is the radius of the planet (mine is 36000/(2*pi)
    All points are in degrees.
    On the tectonics map I think you may have forgot to liable the rift cutting across the northern tip of the continent, it shows spreading but you didn't indicate it was a rift.

    Also your lone rifts with no plate isn't going to happen, they would create plates, even if they are small. Hopefully someone with more knowledge can pipe in.

  10. #10
    Guild Artisan Pixie's Avatar
    Join Date
    Jan 2014
    Location
    Lisbon
    Posts
    939

    Default

    Hey there again.

    Your knowledge of tectonics still needs to grow a bit to get a successful attempt. Here's one thing you should try: reverse time using your map. Have a look at your map and backtrack the continents, reducing the size of the oceans created by ocean rifts and so. You would have a hard time doing this with your current draft - the shapes of the land wouldn't fit with what was going on before.

    If you want to read a little bit on the topic, do a quick google search on the Wilson Cycle, it will help a lot.

    As for your map, there's a very good aspect and a not so good one, when it comes to creating imaginary tectonics: - it isn't a whole globe map, so you have plenty of freedom (that's the positive) - it doesn't have long pensinsulas or convex shoreline facing open ocean (that's the negative). Those convex shorelines, with mountains preferably, are obvious limits for a continental plate and they are the first thing to look for.

    Alas, since you don't have that convex coast but you have a long inner sea, wider on one side, I think the base foundation for your tectonics could be a major oceanic rift forming to split that continent in two. That's what you have, already, but the rest doesn't fit in. Retrospectively, this large continent which is breaking in two halves was moving in which direction? Which is it's passive margin, where was its front boundary?
    I just tried to come up with a better fit for your landmasses and I couldn't get anything worth sharing. So, good luck
    Last edited by Pixie; 10-29-2015 at 07:38 PM.

Page 1 of 2 12 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •