Rive Guide
Search
⌃K

Tilemaps

Some techniques to tile your Rive games
In this tutorial, we’ll explore a simple 2D-level concept and demonstrate a few approaches to tile your game. In principle, it works the same as a traditional sprite tilemap system.
💡 With the Rive Renderer you have the flexibility to make every scene component a vector element. Meaning you can make your world fully interactive and animated.
The picture below demonstrates a sample tilemap with a number of different Rive artboards. The small artboards represent a single tile, while the medium artboards are composited from the small ones using Nested Artboards. The same applies to the final large artboard that is made from the four medium-sized ones.
Any change applied to the nested artboards will reflect in the parent artboards. This is extremely beneficial at design time as it’ll allow you to quickly see what the larger game scene looks like.
Let’s pretend the biggest artboard in the picture above is the final desired world size. It has a size of 4096 x 4096 and is comprised of 64 smaller artboards (8 x 8).
The player, size 300 x 300, is allowed to move freely in the world. If you’re intending to create a mobile game with this world size, it’ll be a waste of rendering resources to draw the whole world at once (the largest artboard) - as most of the scene won’t be in the camera view.
Take a look at the following code that randomly selects one of the smaller tiles and lays them out on the scene, in a 8 x 8 grid:
import 'dart:math';
import 'package:flutter/material.dart';
import 'package:rive_gamekit/rive_gamekit.dart' as rive;
​
class SceneObject {
final rive.Artboard artboard;
final rive.AABB bounds;
final rive.Vec2D position;
​
SceneObject(
this.artboard, {
required this.bounds,
required this.position,
});
}
​
class TilemapPainterDemo extends rive.RenderTexturePainter {
final rive.File riveFile;
final List<rive.Artboard> artboards = [];
final List<SceneObject> sceneObjects = [];
late rive.AABB tileSize;
late Size worldSize;
​
final random = Random();
​
TilemapPainterDemo(this.riveFile) {
// You can reuse the artboard reference for each tile. But if your tile
// as a state machine that needs to be operated independently, you'll
// need to reinitialize the artboard to get a unique state machine.
artboards.add(riveFile.artboard('tile1')!);
artboards.add(riveFile.artboard('tile2')!);
artboards.add(riveFile.artboard('tile3')!);
artboards.add(riveFile.artboard('tile4')!);
artboards.add(riveFile.artboard('tile5')!);
​
const rows = 8;
const columns = 8;
​
// 5 rows
for (var i = 0; i < rows; i++) {
// 5 columns
for (var j = 0; j < columns; j++) {
final artboard = artboards[random.nextInt(artboards.length)];
sceneObjects.add(
SceneObject(
artboard,
bounds: artboard.bounds,
position: rive.Vec2D.fromValues(
j * artboard.bounds.width,
i * artboard.bounds.height,
),
),
);
}
}
​
tileSize = artboards.first.bounds; // assuming all tiles are the same size
worldSize = Size(tileSize.width * columns, tileSize.height * rows);
}
​
@override
Color get background => const Color(0xffffffff);
​
@override
bool paint(rive.RenderTexture texture, Size size, double elapsedSeconds) {
final renderer = rive.Renderer.make();
​
_drawTiles(renderer);
​
return true;
}
​
void _drawTiles(rive.Renderer renderer) {
for (var sceneObject in sceneObjects) {
renderer.save();
renderer.translate(sceneObject.position.x, sceneObject.position.y);
sceneObject.artboard.draw(renderer);
renderer.restore();
}
}
​
@override
void dispose() {
for (var artboard in artboards) {
artboard.dispose();
}
riveFile.dispose();
super.dispose();
}
}
The randomly generated tilemap will look something like this:
Tilemap - 9 tiles visible
About nine tiles are visible in the screenshot above, but 64 are drawn.
To illustrate this, let’s simulate “zooming out the camera” by applying a scale transform:
@override
bool paint(rive.RenderTexture texture, Size size, double elapsedSeconds) {
final renderer = rive.Renderer.make();
​
renderer.save();
// Apply a transform to make the whole game scene smaller.
renderer.transform(rive.Mat2D.fromScale(0.2, 0.2));
​
_drawTiles(renderer);
​
renderer.restore();
return true;
}
Now all 64 tiles are visible.
Tilemap - all 64 tiles visible
The size you choose for your artboards and when to render them will be dependent on your particular game and the screen sizes you target.
For example, in the code given above, you could simply not draw the tile if its position is bigger than the provided size in the paint method. However, if you add a moving player that the camera follows, this calculation becomes more complex, even more so if you’re applying other types of transformations to the renderer, such as the zoom above.

AABB Tree

There are a variety of efficient techniques to cull objects that are not visible in the camera; one such technique is an AABB Tree. See our documentation if you’re unfamiliar with the concept and for additional examples.
The following continues with the code above and demonstrates how to use an AABB Tree to ensure only the visible tiles in the camera view are drawn.
Here is the complete code. You can read the comments and see if you can understand what is going on, after we'll walk through it step by step.
import 'dart:math';
​
import 'package:flutter/material.dart';
import 'package:rive_gamekit/rive_gamekit.dart' as rive;
import 'package:zombie_skins_gamekit/aabb_tree.dart';
​
class SceneObject {
final rive.Artboard artboard;
final rive.AABB bounds;
final rive.Vec2D position;
int proxyId = -1;
​
SceneObject(
this.artboard, {
required this.bounds,
required this.position,
});
}
​
class TilemapPainterDemo extends rive.RenderTexturePainter {
final rive.File riveFile;
final List<rive.Artboard> artboards = [];
final List<SceneObject> sceneObjects = [];
late rive.AABB tileSize;
late Size worldSize;
​
final random = Random();
​
// The AABB tree is used to efficiently query for objects that intersect with
// a given AABB (the camera's view in this example).
late AABBTree<SceneObject> _tree;
// The view transform is used to transform the world into the camera's view.
final rive.Mat2D _viewTransform = rive.Mat2D();
// The inverse view transform is used to transform the camera's view into the
// world.
final rive.Mat2D _inverseViewTransform = rive.Mat2D();
​
TilemapPainterDemo(this.riveFile) {
// STEP 1: Create an AABBTree
_tree = AABBTree();
​
artboards.add(riveFile.artboard('tile1')!);
artboards.add(riveFile.artboard('tile2')!);
artboards.add(riveFile.artboard('tile3')!);
artboards.add(riveFile.artboard('tile4')!);
artboards.add(riveFile.artboard('tile5')!);
​
const rows = 8;
const columns = 8;
​
for (var i = 0; i < rows; i++) {
for (var j = 0; j < columns; j++) {
final artboard = artboards[random.nextInt(artboards.length)];
// STEP 2: Create an object to store information about the tile for
// easy reference.
final sceneObject = SceneObject(
artboard,
bounds: artboard.bounds,
position: rive.Vec2D.fromValues(
j * artboard.bounds.width,
i * artboard.bounds.height,
),
);
​
// STEP 3: Add the object to the tree
// The proxy ID is used to reference the object later if it moves or is
// deleted.
final proxyID = _tree.createProxy(
artboard.bounds.translate(sceneObject.position), sceneObject);
sceneObject.proxyId = proxyID;
​
sceneObjects.add(sceneObject);
}
}
​
tileSize = artboards.first.bounds; // assuming all tiles are the same size
worldSize = Size(tileSize.width * columns, tileSize.height * rows);
}
​
@override
Color get background => const Color(0xffffffff);
​
@override
bool paint(rive.RenderTexture texture, Size size, double elapsedSeconds) {
final renderer = rive.Renderer.make();
​
// Modify the camera's zoom and translation to simulate a camera moving
// around the world.
late double cameraZoom = 1;
late rive.Vec2D cameraTranslation = rive.Vec2D.fromValues(0, 0);
​
// STEP 4: Create a view transform (Mat2D) that represents the visible area
// of the world - depending on the camera zoom and translation.
// The camera's view is a rectangle that is centered on the camera's
// translation and is scaled by the camera's zoom.
_viewTransform[0] = cameraZoom;
_viewTransform[1] = 0;
_viewTransform[2] = 0;
_viewTransform[3] = cameraZoom;
_viewTransform[4] = -cameraTranslation.x * cameraZoom;
_viewTransform[5] = -cameraTranslation.y * cameraZoom;
​
// STEP 5: Create an AABB that represents the camera's view according to
// the window size. This is used to determine which tiles are visible on
// screen.
rive.Mat2D.invert(_inverseViewTransform, _viewTransform);
final cameraAABB = rive.AABB.fromPoints(
[
rive.Vec2D.fromValues(0, 0),
rive.Vec2D.fromValues(size.width, 0),
rive.Vec2D.fromValues(size.width, size.height),
rive.Vec2D.fromValues(0, size.height),
],
transform: _inverseViewTransform,
);
​
renderer.save();
renderer.transform(_viewTransform);
​
// STEP 6: Query the AABB Tree to find all the objects who's bounds intersect
// with the cameraAABB. These will be the tiles that are visible on screen.
// Finally, draw the tiles.
final visibleSceneObjects = <SceneObject>[];
_tree.query(cameraAABB, (id, object) {
visibleSceneObjects.add(object);
return true;
});
_drawTiles(renderer, visibleSceneObjects);
​
renderer.restore();
return true;
}
​
var numberOfVisibleTiles = 0;
void _drawTiles(rive.Renderer renderer, List<SceneObject> objects) {
// This is just to show how many tiles are visible for demo purposes
if (numberOfVisibleTiles != objects.length) {
numberOfVisibleTiles = objects.length;
print('number of visible tiles: ${objects.length}');
}
​
for (var sceneObject in objects) {
renderer.save();
renderer.translate(sceneObject.position.x, sceneObject.position.y);
sceneObject.artboard.draw(renderer);
renderer.restore();
}
}
​
@override
void dispose() {
for (var artboard in artboards) {
artboard.dispose();
}
riveFile.dispose();
super.dispose();
}
}
The important bits can be broken down into seven steps:
  1. 1.
    Create an AABBTree, with a generic SceneObject type. This is used to store information about all of the scene objects and to efficiently query for objects that are within a provided AABB space.
  2. 2.
    Create all the tile objects, which store information about the tile, artboard and position for easy reference. This could be any object that stores the needed information to render your scene.
  3. 3.
    Add all the objects to the AABBTree using the createProxy method. This takes in an AABB and a SceneObject.
  4. 4.
    Create a view transform (Mat2D) that represents the visible area of the world (camera) - depending on the camera zoom and translation.
  5. 5.
    Create an AABB that represents the camera's view according to the window size. This is used to determine which tiles are visible on screen.
  6. 6.
    Query the AABB Tree to find all the objects who's bounds intersect with the cameraAABB. These will be the tiles that are visible on screen. Once you have these, you can draw them to the renderer.
Running the above code you'll see a print that indicates how many tiles are currently visible on screen. This will print a different result if you resize the window, zoom in or out, or moving the camera's translation around. Try updating the cameraZoom and cameraTranslation and performing a hot reload. Or add keyboard and mouse interactions to move the camera.