An A-Frame component for displaying OGC 3D Tiles based on nytimes/three-loader-3dtiles.
Demos — API — Installation — Usage — Contributing — Roadmap
The loader uses the loaders.gl library, which is part of the vis.gl platform, openly governed by the Urban Computing Foundation. Development of the loader started at The New York Times R&D as an effort deliver massive 3D and Geographical journalism to readers. This A-Frame components makes 3D Tiles more accessible to creators and brings immersitve journalism into mixed realities on the web.
Demos are available here.
The following example loads the library from CDN, as well as missing required components for handling KTX2/basis_universal glTFs.
<head>
<title>My A-Frame Scene</title>
<script src="https://aframe.io/releases/1.3.0/aframe.min.js"></script>
<script src='https://cdn.jsdelivr.net/npm/three@0.137.0/examples/js/utils/WorkerPool.js'></script>
<script src='https://cdn.jsdelivr.net/npm/three@0.137.0/examples/js/loaders/KTX2Loader.js'></script>
<script src="https://cdn.jsdelivr.net/npm/aframe-loader-3dtiles-component/dist/aframe-loader-3dtiles-component.min.js"></script>
</head>
The following example sets up a basic scene with a RealityCapture photogrammetry model.
<body>
<a-scene environment="preset: default">
<a-camera id="camera"></a-camera>
<a-entity
id="freeman-tiles"
position="0 2.6 0"
rotation="-90 180 0"
scale="2 2 2"
loader-3dtiles="
url: https://int.nyt.com/data/3dscenes/ONA360/TILESET/0731_FREEMAN_ALLEY_10M_A_36x8K__10K-PN_50P_DB/tileset_tileset.json;
maximumSSE: 48;
cameraEl: #camera;
"
>
</a-entity>
</a-scene>
</body>
Install via npm:
npm install aframe-loader-3dtiles-component
Then require and use.
require('aframe');
require('aframe-loader-3dtiles-component');
Property | Type | Description | Default Value |
---|---|---|---|
url | string | The URL of the tileset. For example if using Cesium ION, it would have the form: https://assets.cesium.com/[ASSET_ID]/tileset.json. Can be updated at runtime. | '' |
cameraEl | selector | A selector for the camera whose position is used to determine which tiles to display at which LOD. If a cameraEl value is not provided then component uses document.querySelector('a-scene').camera to find a-camera primitive or entity with camera component. For more info see Camera Requirement note below. The component automatically switches the camera in response to the camera-set-active event. |
'' |
maximumSSE | int | maximumScreenSpaceError (Optional) determines the distance from tiles in which they are refined, depending on their geometrical size. Increase the value to load lower LOD tiles from the same view distance (increases performance). Cannot be updated at runtime. | 16 |
maximumMem | int | maximumMemoryUsage (Optional) determines maximum GPU memory (MB) to use for displaying tiles. May go over the limit if tiles within the camera viewport exceed that amount. Cannot be updated at runtime. | 32 |
distanceScale | number | (Optional) 0-1 scale for the LOD quality. A lower value loads tiles from lower LODs (increases performance). Can be updated at runtime. | 1.0 |
pointcloudColoring | string | (Optional) When viewing point clouds, which datum determines the color of a point. Possible values: 'white', 'intensity', 'classification', 'elevation', 'rgb'. | 'white' |
pointcloudElevationRange | array | (Optional) When viewing point clouds colored by their elevation, what are the minimum and maximum height values. | 0, 400 |
wireframe | boolean | (Optional) When viewing b3dm (mesh) tiles, show meshes as wireframe. Can be updated at runtime. | false |
showStats | boolean | (Optional) Attaches a box with tilestats to the scene, useful for debugging. | false |
cesiumIONToken | string | (Optional) A Cesium ION access token when loading tilesets from Cesium ION. | '' |
googleApiKey | string | (Optional) Google API Key for loading Google Maps 3D Tiles. | '' |
lat | int | (Optional) latitude coordinate for google 3D Tiles API. | '' |
long | int | (Optional) longitude coordinate for google 3D Tiles API. | '' |
height | int | (Optional) camera height for google 3D Tiles API. | '' |
geoTransform | string | (Optional) How to handle geo transformations: Reset any geo location and place the model at (0,0,0), Apply Mercator projection (for use with ccommon 2D mapping applications, or convert WGS84 long/lat to 3D cartesian coordinates). Possible values: 'Reset', 'Mercator', 'WGS84Cartesian' | 'Reset' |
To use with the Google Maps 3D Map Tiles API, you need to specify googleApiKey
, geoTransform
: 'Mercator' or 'WGS84Cartesian' (see geoTransform property). To set coordinates use: lat
, long
, height
.
To get a 'Google API key', go to the Google Cloud Console and follow the instructions to create an Maps JavaScript API key. Then you need to enable the Google Map Tiles API option on the APIs & Services page.
Please note there are costs from Google to provide this service to users of your application. Pricing starts at $6 per 1,000 loads.
To prevent huge surprise bills, Google 3D Tiles has a low default quota cap of 300 root JSON load events per day which effectively means a limit of roughly 300 unique users per day. This cap is ok for development but you may wish to request a quota increase to support more users in a production application. The fastest route to approval is to directly contact the Maps support team, which is separate from GCP billing or other teams. You will get a quick response but it may take up to 2 business days to lift the quota. Expect to provide additional information about your product, use case and target audience. Caching 3d tiles responses is not permitted by the terms of use.
To use the loader-3dtiles
component, your A-Frame scene must have a camera defined using <a-camera>
or the camera
component . The component will attempt to use one of these automatically, or you can specify your own selector with the cameraEl
property.
The camera is used to provide dynamic levels of detail (LOD) for 3D tiles. If no suitable camera is found the tiles will not display and an error is shown on the console 3D Tiles: Please add an active camera or specify the target camera via the cameraEl property
.
Currently, three-loader-3dtiles
does not support the use of THREE.Cache which is enabled by default in A-Frame. In its default setting, the cache does not recognize repeated loads of tiles, which causes the cache to overflow. Therefore, the loader-3dtiles
component disables THREE.Cache
if it is enabled and displays a warning: 3D Tiles loader cannot work with THREE.Cache, disabling
.
Refer to CONTRIBUTING.MD for general contribution instructions.
Suggestions for future work include:
- More elaborate test cases.
- Better integration with the inspector.
- Test AR scenarios.
- Test point-cloud scenarios.
- Could
THREE.Cache
be used manually? - Conversion to Typescript.
This repository is maintained by the Research & Development team at The New York Times and is provided as-is for your own use. For more information about R&D at the Times visit rd.nytimes.com