Super Mario Bros level editing vs Tiled

Questions
1

I’ve been using Tiled as a level editor for a Super Mario Bros clone and it’s been great to work with. I’ve been referring to Reggie (the unofficial New Super Mario Bros Wii level editor) and Mario Maker along the way, and I’ve noticed they both have interesting tile/object editing features, based around a more abstract tileset.

In Reggie, the tileset seems to be more “object” based rather than tile based. Indeed the “tilsets” are not individual 1x1 tiles, but NxM objects that are painted into the scene. When you select an “object” from the tileset and paint with it or resize an existing “instance”, it automatically tiles and repeats in a pre-defined pattern.

So it’s like a Tiled Object, but it always snapped to the course tile grid, and is rendered as a pattern tiles from the tileset image. A bit of a mix between an object layer and a tile layer. However, such “objects” do not appear to have any per-instance properties. Once an NxM rectangle of tiles has been added it can be translated and resized as a whole. I guess the final export format does not store gids, but stores a set of objects with ids that refer to object definitions that refer to actual gids.

In Super Mario Maker, when you select an “object” such as “grass” and start painting with it, the editor automatically tiles the interior with a N*M pattern fixes up the edges and corners.

Is this something the Tiled terrain brushes can be used for?

I haven’t been working on tiled games for long, so any insight into different methods and terminology would be much appreciated. Thanks.

2

Hmm, I’ve tried to launch Reggie since it indeed seems to have some nice features, but it asked me for some data I don’t know where to find and then told me:

This folder doesn’t seem to have the required files. In order to use Reggie, you need the Stage folder from the game, including the Texture folder and the level files contained within it.

To me it always seems a little strange when such tools are developed that will not work without first obtaining some data that is probably not legally distributable. Also it looks like an enormous effort went into it, but unfortunately its development history was cut off in 2010, when they dumped release 1 into Git.

Maybe, but it may be better to set up automapping rules instead. It’s more complicated but it can do more for you as well.

1 Like