Angband Forums

Angband Forums (http://angband.oook.cz/forum/index.php)
-   Development (http://angband.oook.cz/forum/forumdisplay.php?f=10)
-   -   Angband 64 x 64 pixel tileset (http://angband.oook.cz/forum/showthread.php?t=4012)

takkaria January 4, 2011 19:59

Quote:

Originally Posted by Magnate (Post 46259)
Is the colouring a property of the tile then? Or could we not use 16 different tiles, and display them in three different shades for the lighting?

At the moment, each lighting level has its own tile. We could use shading, but I'm not writing the code :P Plus, I think it's also a pretty good idea to give tile designers control over lighting levels.

Magnate January 4, 2011 20:04

Quote:

Originally Posted by takkaria (Post 46261)
At the moment, each lighting level has its own tile. We could use shading, but I'm not writing the code :P Plus, I think it's also a pretty good idea to give tile designers control over lighting levels.

Sure, that sounds sensible. It's not as if changing the lighting on the 2nd and 3rd sets of 16 is the same amount of work as creating the tiles in the first place.

Shockbolt January 4, 2011 20:27

Another screenshot to see how the tiles are put together/grid placed over them.

I'll adress the replies in here as soon as I can, and I'm working on a less 3D'ish version now(for the wall tiles that is), to cut down on the number of tiles needed. I must say, I did not have the huge number of tiles in mind when I made this first tileset concept :)

Question to the programmers though: "real building artwork" for the above village level, does that require alot of work to code into the game? I had tiles featuring walls/roof etc in mind for those.

http://www.digitalartwork.no/Angband...e_preview2.jpg

fph January 4, 2011 20:33

Quote:

Originally Posted by d_m (Post 46246)
then instead we need 260 tiles to cover all possible kinds of connectedness (1 + 8 + 28 + 56 + 70 + 56 + 28 + 8 + 1).

Nitpick: that sums to 256, not 260. You can get that figure, as well as the 16 above, in a simpler way: each edge or vertex can be either connected or not connected, so 2 possibility for each of the 8 edges/vertices, 2^8=256. :)

Shockbolt January 4, 2011 20:33

Quote:

Originally Posted by d_m (Post 46246)
Currently we have one wall character. We would need to support (at least) 16 (1 + 4 + 6 + 4 + 1) different wall tiles for all the different connectedness options:

x1 wall touches walls on all sides (embedded in other walls
x4 wall touching walls on 3 sides
x6 wall touching walls on 2 sides
x4 wall touching wall on 1 sides
x1 wall touching wall on 0 sides (pillar)

I don't know if we need to worry about diagonal connectedness. If we do, then instead we need 260 tiles to cover all possible kinds of connectedness (1 + 8 + 28 + 56 + 70 + 56 + 28 + 8 + 1). That is probably not feasible.

The graphics look great!

Do you think having 16 kinds of wall tiles will work (e.g. that we can do the 3d walls you're doing)? keep in mind that we would need to "cut up" your graphic into squares and then randomly piece them back together based on the map. I can't visually tell if this will work or not.

I didn't realize the codes would be so complex when I sat painting these today. I'm working on a second edition now, one less 3D than this one, posting it later tonight.

Derakon January 4, 2011 21:03

For what it's worth, I've done some work on faced wall tile generation. Here's one of the tilesets I made:

http://derakon.dyndns.org/~chriswei/...ls/block11.png

That's 20 different tiles (there's 9 in that large center block) and it covers every combination I wanted for the project I was working on. Angband doesn't need the 45 sloped tiles, which brings you down to 16. Lighting should be handled through some kind of postprocessing filter that tints the tiles IMO.

Probably you could get away with only rendering the faces of the walls, and using a procedural texture fill for the interiors; I've been meaning to give that a shot since it would have the added benefit of making the wall interiors less repetitive. Then you'd only need 12 images: four faces (north/east/west/south), 4 corners joining faces (northeast, northwest, southeast, southwest), and four corners that don't join faces (likewise). Probably you could get rid of the latter set of corners and bring it down to only 8 manually-created images.

Though honestly, at this point I start recommending you just use OpenGL and 3D geometries that actually do what you're trying to simulate with 2D.

Shockbolt January 4, 2011 21:09

Quote:

Originally Posted by Derakon (Post 46270)

Though honestly, at this point I start recommending you just use OpenGL and 3D geometries that actually do what you're trying to simulate with 2D.

Now that comment I liked :p

Shockbolt January 4, 2011 23:21

64 x 64 pixel tileset "V2.0".

This tileset has 2 types of walls and 2 types of doors, I'll explain my thoughts about them:

Walls:
There's now a solid wall for vertical walls and "fillings" between rooms/tunnels,
the "south" end of this wall needs to end in a "3D-type" wall.
"3D-type" walls are shown for horizontal walls and "south" corners.

Doors:
Horizontal doors are just 1 tile as shown in the concept,vertical doors need to be 2 tiles stacked vertically. As seen in the concept, the vertical door to the left is shown as open, the one to the right is closed. The vertical door should also have a version that is flipped 180 degrees.

This should generate less work than my initial approach to the tileset...hopefully.

I'm also thinking of creating different player tiles for when the player is standing in the doorway (some of the player being blocked by the wall/doorway), it shouldn't be too hard to add this feature to the game?

http://www.digitalartwork.no/Angband...e_preview3.jpg

dos350 January 5, 2011 03:09

still looking good , ee!

d_m January 5, 2011 04:00

Quote:

Originally Posted by fph (Post 46265)
Nitpick: that sums to 256, not 260. You can get that figure, as well as the 16 above, in a simpler way: each edge or vertex can be either connected or not connected, so 2 possibility for each of the 8 edges/vertices, 2^8=256. :)

Indeed you're correct.


All times are GMT +1. The time now is 04:06.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.