Hi,
I’m trying to understand what I am doing wrong with some images I want to put on a map. These images are just for testing for now but a good example of what my final ones will look like.
I have buildings that will be spread accross several tiles (let’s say 2x2) so I split the image into images of 128x64 (my tile size). When I select the group of 8 images composing the building and add it on the map, I get the following result:
If I add the images one by one I get the correct result:
When I load the map in my game simulator (Corona SDK), both representations are worng:
Can you see what I am doing wrong here?
Thanks!
PS: My images are imported based on a tileset image and my map is isometric type. I tried to upload the file but I don’t have the rights yet.
Could you share your tileset as well, so we have a good idea of what’s going on?
In general, it’s easiest to work with the tiles if they are logical pieces. So, this building for example could be 2x2 tiles or 3x3 tiles (the latter would allow it to grow to arbitrary sizes). But the tiles would need to be relatively high to cover the entire height of the building.
Right, that’s a good example of what the tiles should look like to make editing more convenient. Though it should be noted that Tiled currently does not support tiles of different sizes within the same image. You’d need to either have multiple tileset images, one for each tile size, or use an image collection tileset where each tile has its own image file.
At some point I hope to get around to the following issue, which would allow tilesets like the above work out of the box:
The reason why I would like to split my buildings/images into several tiles is that I want one row per tile in my database so I am able to check if a tile is occupied. See what was suggested to me here
@Javo I’d like to suggest something but not sure if your current setup will allow it.
I highly recommend using Objects for buildings or any large element, and then using generic (flat) tiles as markers for occupied spaces/tiles, or other special meanings (like identifying landmarks or terrain properties, e.g. water, mud, etc)
However, when placing an Object in an isometric layout you need to transform the ‘orthographic’ coordinates written in the TMX to an ‘isometric-to-orthographic projection’ in the game engine. For example, assuming 32x16 tiles, the X/Y coordinates of tile 1,1 will be 16,16. But when used in the game engine (assuming you are placing them orthographically on-screen) the actual X/Y coordinates should be 0,16 (assuming x=0 is offsetted accordingly to center). This is not too difficult to do (you can generate your vectors from the tile proportions:
E.g.
tile_w = 32
tile_h = 16
uox_x = tile_w/2
uox_y = tile_h/2
uoy_x = -(tile_w/2)
uoy_y = tile_h/2
ux = in_x/tile_h
uy = in_y/tile_h # Yes it's still height!
out_x = (ux * uox_x) + (uy * uoy_x) # final x
out_y = (ux * uox_y) + (uy * uoy_y) # final y
I did it this way because the art workflow is more efficient (e.g. revisions, modularity) especially for big sprites.
This is the prototype that shows the result of the technique (using C2 as the framework).
Just a suggestion, though I realise it might be a totally different way to work.
@faulknermano thanks for the reply but I’m using a plugin called Qiso, with Corona SDK and that plugin already handles the conversion so I probably don’t need to do it myself.
Also, if I add a single image on a single tile, it is added correctly on my map. The problem is only when dividing an image into several ones.
Oh, and I will look into using objects for my buildings, for I guess it won’t change my current problem?
Edit: Actually I just found out that it’s a rendering problem with Qiso. If I add the images (through my app) without the grass background, the building shows up properly. I will look on the Qiso side.
Edit 2: The problem was that the tiles where not added in the right location on the map.
