four

four

精简 3D 图形渲染库 支持 WebGL 和 WebGPU

Four 是一个精简的 3D 图形渲染库,提供简洁 API 支持 WebGL 和 WebGPU。包含核心 3D 对象、几何体、材质和渲染功能,注重性能和易用性。适用于快速构建各类 3D 图形应用,从基础场景到复杂可视化项目。

WebGLWebGPU3D渲染图形编程three.jsGithub开源项目

Size Version Downloads

four

Minimal three.js alternative.

Table of Contents

Installation

To install, use your preferred package manager or CDN:

npm install four@npm:fourwastaken yarn add four@npm:fourwastaken pnpm add four@npm:fourwastaken
<script type="module"> import * as FOUR from 'https://unpkg.com/fourwastaken' </script>

Note: Vite may have issues consuming WebGPU code which relies on top-level await via ESM. This is well supported since 2021, but you may need to use vite-plugin-top-level-await to use this library with vite.optimizeDeps.

Getting Started

The following creates a renderer, camera, and renders a red cube:

<details> <summary>Show WebGL example</summary>
import { WebGLRenderer, PerspectiveCamera, Geometry, Material, Mesh } from 'four' const renderer = new WebGLRenderer() renderer.setSize(window.innerWidth, window.innerHeight) document.body.appendChild(renderer.canvas) const camera = new PerspectiveCamera(45, window.innerWidth / window.innerHeight) camera.position.z = 5 const geometry = new Geometry({ position: { size: 3, data: new Float32Array([ 0.5, 0.5, 0.5, 0.5, -0.5, 0.5, 0.5, 0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, -0.5, -0.5, 0.5, -0.5, -0.5, -0.5, -0.5, -0.5, 0.5, 0.5, -0.5, -0.5, -0.5, -0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, 0.5, -0.5, -0.5, 0.5, 0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, -0.5, 0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, -0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, -0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, 0.5, -0.5, -0.5, -0.5, -0.5, -0.5, -0.5, 0.5, -0.5, ]), }, }) const material = new Material({ vertex: /* glsl */ `#version 300 es uniform mat4 projectionMatrix; uniform mat4 modelViewMatrix; in vec3 position; void main() { gl_Position = projectionMatrix * modelViewMatrix * vec4(position, 1); } `, fragment: /* glsl */ `#version 300 es out lowp vec4 color; void main() { color = vec4(1, 0, 0, 1); } `, }) const mesh = new Mesh(geometry, material) renderer.render(mesh, camera)
</details> <details> <summary>Show WebGPU example</summary>
import { WebGPURenderer, PerspectiveCamera, Geometry, Material, Mesh } from 'four' const renderer = new WebGPURenderer() renderer.setSize(window.innerWidth, window.innerHeight) document.body.appendChild(renderer.canvas) const camera = new PerspectiveCamera(45, window.innerWidth / window.innerHeight) camera.position.z = 5 const geometry = new Geometry({ position: { size: 3, data: new Float32Array([ 0.5, 0.5, 0.5, 0.5, -0.5, 0.5, 0.5, 0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, -0.5, -0.5, 0.5, -0.5, -0.5, -0.5, -0.5, -0.5, 0.5, 0.5, -0.5, -0.5, -0.5, -0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, 0.5, -0.5, -0.5, 0.5, 0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, -0.5, 0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, 0.5, 0.5, -0.5, -0.5, 0.5, 0.5, -0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, -0.5, 0.5, -0.5, -0.5, -0.5, 0.5, -0.5, 0.5, -0.5, -0.5, -0.5, -0.5, -0.5, -0.5, 0.5, -0.5, ]), }, }) const material = new Material({ vertex: /* wgsl */ ` struct Uniforms { projectionMatrix: mat4x4<f32>, modelViewMatrix: mat4x4<f32>, }; @group(0) @binding(0) var<uniform> uniforms: Uniforms; @vertex fn main(@location(0) position: vec3<f32>) -> @builtin(position) vec4<f32> { return uniforms.projectionMatrix * uniforms.modelViewMatrix * vec4(position, 1); } `, fragment: /* wgsl */ ` @fragment fn main() -> @location(0) vec4<f32> { return vec4(1, 0, 0, 1); } `, }) const mesh = new Mesh(geometry, material) renderer.render(mesh, camera)
</details>

Object3D

An Object3D represents a basic 3D object and its transforms. Objects are linked via their parent and children properties, constructing a rooted scene-graph.

const object = new Object3D() object.add(new Object3D(), new Object3D()) object.traverse((node) => { if (node !== object) object.remove(node) if (!node.visible) return true })

Vector3

A Vector3 represents a three-dimensional (x, y, z) vector and describes local position in Object3D.position. It is also used to control local scale in Object3D.scale.

object.position.set(1, 2, 3) object.position.x = 4 object.position[0] = 5

Quaternion

A Quaternion represents a four-dimensional vector with a rotation axis (x, y, z) and magnitude (w) and describes local orientation in Object3D.quaternion.

object.quaternion.set(0, 0, 0, 1) object.quaternion.fromEuler(Math.PI / 2, 0, 0) object.quaternion.x *= -1 object.quaternion[0] *= -1

Matrix4

A Matrix4 represents a 4x4 transformation matrix and describes world transforms in Object3D.matrix.

object.matrix.set(1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 2, 3, 1) object.matrix[12] = 4 object.matrix.invert() object.matrix.identity()

Mesh

A Mesh contains a Geometry and Material to describe visual behavior, and can be manipulated in 3D as an Object3D.

const geometry = new Geometry({ ... }) const material = new Material({ ... }) const mesh = new Mesh(geometry, material)

Geometry

A Geometry contains an Attribute list of vertex or storage buffer data, with a GPU buffer allocated for each Attribute.

const geometry = new Geometry({ position: { size: 2, data: new Float32Array([-1, -1, 3, -1, -1, 3]) }, uv: { size: 2, data: new Float32Array([0, 0, 2, 0, 0, 2]) }, index: { size: 1, data: new Uint16Array([0, 1, 2]) }, })

A DrawRange can also be configured to control rendering without submitting vertex data. This is useful for GPU-computed geometry or vertex pulling, as demonstrated in the fullscreen demos.

const geometry = new Geometry() geometry.drawRange = { start: 0, count: 3 } // renders 3 vertices at starting index 0

Attribute

An Attribute defines a data view, its per-vertex size, and an optional per-instance divisor (see instancing).

// Creates a 4x4 instance matrix for 2 instances { data: new Float32Array([ 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, ]), size: 16, divisor: 1, }

Material

A Material describes a program or shader interface for rasterization and compute (see compute), defining a vertex and fragment or compute shader, respectively.

<details> <summary>Show WebGL example</summary>
const material = new Material({ vertex: /* glsl */ `#version 300 es uniform mat4 projectionMatrix; uniform mat4 modelViewMatrix; in vec3 position; void main() { gl_Position = projectionMatrix * modelViewMatrix * vec4(position, 1); } `, fragment: /* glsl */ `#version 300 es out lowp vec4 color; void main() { color = vec4(1, 0, 0, 1); } `, side: 'front', transparent: false, depthTest: true, depthWrite: true, })
</details> <details> <summary>Show WebGPU example</summary>
const material = new Material({ vertex: /* wgsl */ ` struct Uniforms { projectionMatrix: mat4x4<f32>, modelViewMatrix: mat4x4<f32>, }; @group(0) @binding(0) var<uniform> uniforms: Uniforms; @vertex fn main(@location(0) position: vec3<f32>) -> @builtin(position) vec4<f32> { return uniforms.projectionMatrix * uniforms.modelViewMatrix * vec4(position, 1); } `, fragment: /* wgsl */ ` @fragment fn main() -> @location(0) vec4<f32> { return vec4(1, 0, 0, 1); } `, side: 'front', transparent: false, depthTest: true, depthWrite: true, })
</details>

Uniforms

The following uniforms are built-in and will be automatically populated when specified:

TypeNameDescriptionConversion
mat4x4modelMatrixworld-space mesh transformlocal space => world space
mat4x4projectionMatrixclip-space camera projectionview space => clip space
mat4x4viewMatrixinverse camera transformworld space => view space
mat4x4modelViewMatrixpremultiplied model-view transformlocal space => view space
mat4x4normalMatrixisotropic inverse model-view or "normal" transformlocal space => view space

In WebGPU, uniforms are bound to a single uniform buffer, preceded by storage buffers, and followed by sampler-texture for texture uniforms.

// Storage buffers @group(0) @binding(0) var<storage, read_write> data: array<vec2<f32>>; // Uniform buffer struct Uniforms { time: f32, }; @group(0) @binding(1) var<uniform> uniforms: Uniforms; // Texture bindings @group(0) @binding(2) var sample: sampler; @group(0) @binding(3) var color: texture_2d<f32>; @group(0) @binding(4) var sample_2: sampler; @group(0) @binding(5) var color_2: texture_2d<f32>;

Blending

By default, opaque meshes do not blend but replace values, and transparent meshes alpha blend by the following blend equation:

material.blending = { color: { operation: 'add', srcFactor: 'src-alpha', dstFactor: 'one-minus-src-alpha', }, alpha: { operation: 'add', srcFactor: 'one', dstFactor: 'one-minus-src-alpha', }, }

This gets applied to the final fragment color as src * srcFactor + dst * dstFactor, assuming a premultiplied alpha.

Custom blending can be used for postprocessing and various VFX. The following are the most common configurations:

Blend ModeBlendOperationBlendFactor (src)BlendFactor (dst)
Additiveaddsrc-alphaone
Subtractivereverse-subtractsrc-alphaone
Multiplyadd

编辑推荐精选

博思AIPPT

博思AIPPT

AI一键生成PPT,就用博思AIPPT!

博思AIPPT,新一代的AI生成PPT平台,支持智能生成PPT、AI美化PPT、文本&链接生成PPT、导入Word/PDF/Markdown文档生成PPT等,内置海量精美PPT模板,涵盖商务、教育、科技等不同风格,同时针对每个页面提供多种版式,一键自适应切换,完美适配各种办公场景。

AI办公办公工具AI工具博思AIPPTAI生成PPT智能排版海量精品模板AI创作热门
潮际好麦

潮际好麦

AI赋能电商视觉革命,一站式智能商拍平台

潮际好麦深耕服装行业,是国内AI试衣效果最好的软件。使用先进AIGC能力为电商卖家批量提供优质的、低成本的商拍图。合作品牌有Shein、Lazada、安踏、百丽等65个国内外头部品牌,以及国内10万+淘宝、天猫、京东等主流平台的品牌商家,为卖家节省将近85%的出图成本,提升约3倍出图效率,让品牌能够快速上架。

iTerms

iTerms

企业专属的AI法律顾问

iTerms是法大大集团旗下法律子品牌,基于最先进的大语言模型(LLM)、专业的法律知识库和强大的智能体架构,帮助企业扫清合规障碍,筑牢风控防线,成为您企业专属的AI法律顾问。

SimilarWeb流量提升

SimilarWeb流量提升

稳定高效的流量提升解决方案,助力品牌曝光

稳定高效的流量提升解决方案,助力品牌曝光

Sora2视频免费生成

Sora2视频免费生成

最新版Sora2模型免费使用,一键生成无水印视频

最新版Sora2模型免费使用,一键生成无水印视频

Transly

Transly

实时语音翻译/同声传译工具

Transly是一个多场景的AI大语言模型驱动的同声传译、专业翻译助手,它拥有超精准的音频识别翻译能力,几乎零延迟的使用体验和支持多国语言可以让你带它走遍全球,无论你是留学生、商务人士、韩剧美剧爱好者,还是出国游玩、多国会议、跨国追星等等,都可以满足你所有需要同传的场景需求,线上线下通用,扫除语言障碍,让全世界的语言交流不再有国界。

讯飞绘文

讯飞绘文

选题、配图、成文,一站式创作,让内容运营更高效

讯飞绘文,一个AI集成平台,支持写作、选题、配图、排版和发布。高效生成适用于各类媒体的定制内容,加速品牌传播,提升内容营销效果。

热门AI辅助写作AI工具讯飞绘文内容运营AI创作个性化文章多平台分发AI助手
TRAE编程

TRAE编程

AI辅助编程,代码自动修复

Trae是一种自适应的集成开发环境(IDE),通过自动化和多元协作改变开发流程。利用Trae,团队能够更快速、精确地编写和部署代码,从而提高编程效率和项目交付速度。Trae具备上下文感知和代码自动完成功能,是提升开发效率的理想工具。

AI工具TraeAI IDE协作生产力转型热门
商汤小浣熊

商汤小浣熊

最强AI数据分析助手

小浣熊家族Raccoon,您的AI智能助手,致力于通过先进的人工智能技术,为用户提供高效、便捷的智能服务。无论是日常咨询还是专业问题解答,小浣熊都能以快速、准确的响应满足您的需求,让您的生活更加智能便捷。

imini AI

imini AI

像人一样思考的AI智能体

imini 是一款超级AI智能体,能根据人类指令,自主思考、自主完成、并且交付结果的AI智能体。

下拉加载更多