troika-three-text

What is troika-three-text?

The troika-three-text npm package is a library for rendering high-quality, performant text in Three.js scenes. It provides a simple and flexible way to add 3D text to your WebGL applications, with support for advanced features like custom fonts, text layout, and text effects.

What are troika-three-text's main functionalities?

Basic Text Rendering

This feature allows you to render basic 3D text in a Three.js scene. You can set the text content, font size, and position.

const { Text } = require('troika-three-text');
const text = new Text();
text.text = 'Hello, World!';
text.fontSize = 1;
text.position.set(0, 0, 0);
scene.add(text);

Custom Fonts

This feature allows you to use custom fonts by specifying the URL to a font file. This can be useful for branding or stylistic purposes.

const { Text } = require('troika-three-text');
const text = new Text();
text.text = 'Custom Font';
text.font = 'https://example.com/path/to/font.woff';
text.fontSize = 1;
text.position.set(0, 0, 0);
scene.add(text);

Text Effects

This feature allows you to apply various effects to the text, such as changing the color or adding glow effects. This can enhance the visual appeal of the text in your scene.

const { Text } = require('troika-three-text');
const text = new Text();
text.text = 'Glowing Text';
text.fontSize = 1;
text.position.set(0, 0, 0);
text.material = new THREE.MeshBasicMaterial({ color: 0xff0000 });
scene.add(text);

Other packages similar to troika-three-text

three-bmfont-text

The three-bmfont-text package is another library for rendering text in Three.js. It uses Bitmap fonts for text rendering, which can be more performant for certain use cases. However, it lacks some of the advanced features and flexibility of troika-three-text, such as support for custom fonts and text effects.

three-mesh-ui

The three-mesh-ui package is designed for creating user interfaces in Three.js scenes. It supports text rendering along with other UI components like buttons and panels. While it offers a broader range of UI elements, it may not be as specialized or optimized for text rendering as troika-three-text.

README

troika-three-text

This package provides high quality text rendering in Three.js scenes, using signed distance fields (SDF) and antialiasing using standard derivatives.

Rather than relying on pre-generated SDF textures, this parses font files (.ttf, .otf, .woff) directly using Typr.js, and generates the SDF atlas for glyphs on-the-fly as they are used. It also handles proper kerning and ligature glyph substitution. All font parsing, SDF generation, and glyph layout is performed in a web worker to prevent frame drops.

Once the SDFs are generated, it assembles a geometry that positions all the glyphs, and patches any Three.js Material with the proper shader code for rendering the SDFs. This means you can still benefit from all the features of Three.js's built-in materials like lighting, physically-based rendering, shadows, and fog.

Demos

• With the Troika scene management framework
• With react-three-fiber

With Other Frameworks

• In the drei utilities for react-three-fiber
• As an A-Frame component

Screenshots

Installation

Get it from NPM:

npm install troika-three-text

You will also need to install a compatible version of Three.js; see the notes in the Troika 3D Readme for details.

Usage

import {Text} from 'troika-three-text'

You can then use the Text class like any other Three.js mesh:

// Create:
const myText = new Text()
myScene.add(myText)

// Set properties to configure:
myText.text = 'Hello world!'
myText.fontSize = 0.2
myText.position.z = -2
myText.color = 0x9966FF

// Update the rendering:
myText.sync()

It's a good idea to call the .sync() method after changing any properties that would affect the text's layout. If you don't, it will be called automatically on the next render frame, but calling it yourself can get the result sooner.

When you're done with the Text instance, be sure to call dispose on it to prevent a memory leak:

myScene.remove(myText)
myText.dispose()

Supported properties

Instances of Text support the following configuration properties:

text

The string of text to be rendered. Newlines and repeating whitespace characters are honored.

Default: none

anchor

This property is deprecated as of version 0.24.0; use anchorX and anchorY instead.

anchorX

Defines the horizontal position in the text block that should line up with the local origin. Can be specified as a numeric x position in local units, a string percentage of the total text block width e.g. '25%', or one of the following keyword strings: 'left', 'center', or 'right'.

Default: 0

anchorY

Defines the vertical position in the text block that should line up with the local origin. Can be specified as a numeric y position in local units (note: down is negative y), a string percentage of the total text block height e.g. '25%', or one of the following keyword strings: 'top', 'top-baseline', 'middle', 'bottom-baseline', or 'bottom'.

Default: 0

clipRect

If specified, defines the [minX, minY, maxX, maxY] of a rectangle outside of which all pixels will be discarded. This can be used for example to clip overflowing text when whiteSpace='nowrap'.

Default: none

color

This is a shortcut for setting the color of the text's material. You can use this if you don't want to specify a whole custom material and just want to change its color.

Use the material property if you want to control aspects of the material other than its color.

Default: none - uses the color of the material

depthOffset

This is a shortcut for setting the material's polygonOffset and related properties, which can be useful in preventing z-fighting when this text is laid on top of another plane in the scene. Positive numbers are further from the camera, negatives closer.

Be aware that while this can help with z-fighting, it does not affect the rendering order; if the text renders before the content behind it, you may see antialiasing pixels that appear too dark or light. You may need to also change the text mesh's renderOrder, or set its z position a fraction closer to the camera, to ensure the text renders after background objects.

Default: 0

font

The URL of a custom font file to be used. Supported font formats are:

• .ttf
• .otf
• .woff (.woff2 is not supported)

Default: The Roboto font loaded from Google Fonts CDN

fontSize

The em-height at which to render the font, in local world units.

Default: 0.1

glyphGeometryDetail

The number of vertical/horizontal segments that make up each glyph's rectangular plane. This can be increased to provide more geometrical detail for custom vertex shader effects, for example.

Default: 1

letterSpacing

Sets a uniform adjustment to spacing between letters after kerning is applied, in local world units. Positive numbers increase spacing and negative numbers decrease it.

Default: 0

lineHeight

Sets the height of each line of text. Can either be 'normal' which chooses a reasonable height based on the chosen font's ascender/descender metrics, or a number that is interpreted as a multiple of the fontSize.

Default: 'normal'

material

Defines a Three.js Material instance to be used as a base when rendering the text. This material will be automatically replaced with a new material derived from it, that adds shader code to decrease the alpha for each fragment (pixel) outside the text glyphs, with antialiasing.

By default it will derive from a simple white `MeshBasicMaterial, but you can use any of the other mesh materials to gain other features like lighting, texture maps, etc.

Also see the color shortcut property.

Note that because your material instance is replaced by a derived material instance, any changes you make to your original material will not be reflected in the derived version. If you need to modify properties of the material afterward, be sure you get a new reference to the derived version:

// Bad:
text.material = myOrigMaterial
myOrigMaterial.opacity = 0.5

// Good:
text.material = myOrigMaterial
text.material.opacity = 0.5

Default: a MeshBasicMaterial instance

maxWidth

The maximum width of the text block, above which text may start wrapping according to the whiteSpace and overflowWrap properties.

Default: Infinity, meaning text will never wrap

overflowWrap

Defines how text wraps if the whiteSpace property is 'normal'. Can be either 'normal' to break at whitespace characters, or 'break-word' to allow breaking within words.

Default: 'normal'

sdfGlyphSize

Allows overriding the default size of each glyph's SDF (signed distance field) used when rendering this text instance. This must be a power-of-two number. Larger sizes can improve the quality of glyph rendering by increasing the sharpness of corners and preventing loss of very thin lines, at the expense of increased memory footprint and longer SDF generation time.

Default: 64

textAlign

The horizontal alignment of each line of text within the overall text bounding box. Can be one of 'left', 'right', 'center', or 'justify'.

Default: 'left'

textIndent

An indentation applied to the first character of each hard newline. Behaves like CSS text-indent.

Default: 0

whiteSpace

Defines whether text should wrap when a line reaches the maxWidth. Can be either 'normal', to allow wrapping according to the overflowWrap property, or 'nowrap' to prevent wrapping.

Note that 'normal' in this context does honor newline characters to manually break lines, making it behave more like 'pre-wrap' does in CSS.

Default: 'normal'

Handling Asynchronous Updates

Since the text processing occurs in a web worker, it is by definition asynchronous. This means that you can't rely on the text being visible or having a complete geometry immediately. If you need to do things like access the geometry's boundingSphere or the textRenderInfo, you will have to listen for completion. You can do this two ways:

1. Pass a callback function when you call the sync method:

   myText.sync(() => {
     // code to execute after sync completes...
   })
   

   This is best when you want to only react to that specific sync call. Keep in mind that the callback will not execute if the text is already fully synced.

2. Add a listener for the synccomplete event:

   myText.addEventListener('synccomplete', () => {
     // code to execute after sync completes...
   })
   

   This will fire after every sync, no matter who invoked it. This is best if you need to react to all syncs, for example to trigger a manual canvas render.

   You can also listen for the syncstart event if you need to react to the initiation of a sync call, e.g. to set some sort of "waiting" state while the text is being processed.

Preloading

To avoid long pauses when first displaying a piece of text in your scene, you can preload fonts and optionally pre-generate the SDF textures for particular glyphs up front:

import {preloadFont} from 'troika-three-text'

myApp.showLoadingScreen()

preloadFont(
  {
    font: 'path/to/myfontfile.woff', 
    characters: 'abcdefghijklmnopqrstuvwxyz'
  },
  () => {
    myApp.showScene()
  }
)

The arguments are:

• options

  • options.font - The URL of the font file to preload. If null is passed, this will preload the default font.

  • options.characters - A string or array of string character sequences for which to pre-generate glyph SDF textures. Note that this will honor ligature substitution, so you may need to specify ligature sequences in addition to their individual characters to get all possible glyphs, e.g. ["t", "h", "th"] to get the "t" and "h" glyphs plus the "th" glyph.

  • options.sdfGlyphSize - The size at which to prerender the SDFs for the characters glyphs. See the sdfGlyphSize config property on Text for details about SDF sizes. If not specified, will use the default SDF size.

• callback - A function that will be called when the preloading is complete.

Postprocessing

It is possible to use Text within scenes that utilize the postprocessing library for applying image effects. However, you must enable a special mode in that library that allows Text's custom material to be honored. Just do the following once somewhere in your code:

import { OverrideMaterialManager } from 'postprocessing'

OverrideMaterialManager.workaroundEnabled = true

Carets and Selection Ranges

In addition to rendering text, it is possible to access positioning information for caret placement and selection ranges. To access that info, use the getCaretAtPoint and getSelectionRects utility functions. Both of these functions take a textRenderInfo object as input, which you can get from the Text object's textRenderInfo property after sync has completed. See "Handling Asynchronous Updates" above for how to react to sync completion events.

getCaretAtPoint(textRenderInfo, x, y)

This returns the caret position nearest to a given x/y position in the local text plane. This is useful for placing an editing caret based on a click or ther raycasted event. The return value is an object with the following properties:

• x - x position of the caret
• y - y position of the caret's bottom
• height - height of the caret, based on the current fontSize and lineHeight
• charIndex - the index in the original input string of this caret's target character. The caret will be for the position before that character. For the final caret position, this will be equal to the string length. For ligature glyphs, this will be for the first character in the ligature sequence.

getSelectionRects(textRenderInfo, start, end)

This returns a list of rectangles covering all the characters within a given character range. This is useful for highlighting a selection range. The return value is an array of objects, each with {left, top, right, bottom} properties in the local text plane.