Results 1 to 7 of 7

Thread: 3d perspective maps

  1. #1
    Guild Novice Facebook Connected
    Join Date
    Nov 2017
    Location
    Buffalo, NY
    Posts
    22

    Default 3d perspective maps

    I am more math nerd than artist and I struggle with most of the techniques unless their is some solid math or process behind them. I can do basic perspective drawing but that's because it has a method and rules to follow. For the really ncie 3d overhead perspective pieces is there a formula to determine how much smaller you make the base from the top of the walls to achieve that top down look? I understand the perspective concepts but every attempt I have made at making the walls looks awful. any ideas?

  2. #2
    Guild Master Falconius's Avatar
    Join Date
    Jul 2013
    Location
    Israel
    Posts
    2,727

    Default

    The problem with top down perspective with a 3d map is that the top will obscure the bottom for anything you are directly over and at the edges will give you quite a lot of distortion. Top down (assuming straight walls) is basically single point perspective right in the center of the page. So what most people do (in video games and such) is actually make the top smaller than the base, so you can see the wall. In other words there is nothing particularly mathematic about it, its mostly just what looks good to the eye. Even if they don't do that, with top down people do usually use orthographic perspective, meaning vanishing points simply do not exist, but that won't give you a 3d look without making the tops smaller than the bases.

  3. #3
    Guild Novice Facebook Connected
    Join Date
    Nov 2017
    Location
    Buffalo, NY
    Posts
    22

    Default

    wait a second!! you may have solved it for me. Is it just an optical illusion where the top and the bottom are the same size but the graphic for the top of the wall is smaller than the full width of the object?? I never considered that and it sound brilliant. I'm no being sarcastic either I really hadn't thought of that. So if a wall section is say, 10 squares wide and 1 square deep I don't mess with the base of the wall piece at all I just cut the wall top down to based on the forced perspective? So the bottom wall which I assume the viewer is directly on top of would be 100% wall top while the side walls would be say 2/3 and the top wall which is the most revealed would be about 1/3 of the square. So in pixles if I go with 70 the im looking at 50/20 top wall texture/top texture. I will have to test this out. If i produce anything that isn't laughable I will post it. Thanks

  4. #4
    Guild Novice Facebook Connected
    Join Date
    Nov 2017
    Location
    Buffalo, NY
    Posts
    22

    Default

    Nope that didn't work.. I messed around with trying to use light and shadow it looks terrible. I may be bad at it but I double checked some reference material and the base of the wall is smaller than the top. I guess i just nee to figure out the angle or ration to get a good look.

  5. #5
    Guild Master Falconius's Avatar
    Join Date
    Jul 2013
    Location
    Israel
    Posts
    2,727

    Default

    What's the reference material? Also let's see the results. It's hard to figure out visual problems without seeing them.

    If you check out this discussion over here: https://www.cartographersguild.com/s...ad.php?t=40553 You can see in those maps that to achieve that look he either had to take individual perspective shots of each room and hallway and then stitch them together over an orthographic shot of the floor, or that he had to manipulate the model itself to work with an orthographic shot from the top. Top down in perspective is single point perspective (for the most part, really it's 5 point perspective but it's easier to ignore the other four) which leads to a lot of distortion which doesn't work for a map.

  6. #6
    Guild Novice Facebook Connected
    Join Date
    Nov 2017
    Location
    Buffalo, NY
    Posts
    22

    Default

    Here is what I want, vs what I get. I understand that this is not something that would be perfect the first go and that most of these are based on computer aided drafting models that are turned into skinned images. That being said there must be some math that the computer uses to go from wire frame to image, some formula of distance from center point equating to some angle between top corner and bottom corner. That is what Im trying to figure out so I can replicate the effect with my limited tools.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Pieper-Properties_Mayer-Building_Unit-205_1x1_Studio_Renovation.png 
Views:	16 
Size:	1.00 MB 
ID:	104348   Click image for larger version. 

Name:	3d test room flat.jpg 
Views:	13 
Size:	235.4 KB 
ID:	104349  

  7. #7
    Guild Master Falconius's Avatar
    Join Date
    Jul 2013
    Location
    Israel
    Posts
    2,727

    Default

    The computer is essentially using single point vanishing lines in the first one (there is obviously a second point way off the page going up, but it is not important really).
    Click image for larger version. 

Name:	Pieper-Properties_Mayer-Building_Unit-205_1x1_Studio_Renovation.png 
Views:	33 
Size:	1.29 MB 
ID:	104357

    The camera is pointing almost straight down in the area of the blue square and they've cropped the image to show the import bit where the floor layout is.

    If you want to draw that I'd suggest you lay out your floor plan first perfectly flat as it would appear in a totally 2d plan, choose your camera position (the vanishing point) and then run the lines through the corners of your plan where ever they occur visibly (like the red lines in the example). Determining wall height is kinda of just going by whatever looks good (it's kind of arbitrary since in real life it's determined by the focal length of the camera or eye or whatever). To get them all looking of equal height is a lot more work, but is basically done by intersections all based off of your original height, you plot out horizontal lines (perfectly flat since we are only using one point) and then vertical lines (again totally vertical since we are not using that second point like they do in your example) for perpendicular (both these represented by the green lines) and then draw the corner line where two walls meet from our single point and wherever that intersects our horizontal (yellow line) and vertical lines is the even height of the wall. You should note that even all that won't achieve true perspective, but it'll be close enough to get away with it.

    This method is an extraordinarily painful way to achieve this result however. This is why, among many other reasons, most people don't go for a realistic perspective for maps. The other major reason is usability since it will necessarily obscure useful information, which is generally the opposite of what maps try to achieve. That is why mot people just use isometric representations or why they warp the reality to display a more user friendly experience.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Pieper-Properties_Mayer-Building_Unit-205_1x1_Studio_Renovation.png 
Views:	15 
Size:	1.27 MB 
ID:	104356  

Posting Permissions

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