Project Icon

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。

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:

Show WebGL example
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)
Show WebGPU example
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)

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.

Show WebGL example
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,
})
Show WebGPU example
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,
})

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
项目侧边栏1项目侧边栏2
推荐项目
Project Cover

豆包MarsCode

豆包 MarsCode 是一款革命性的编程助手,通过AI技术提供代码补全、单测生成、代码解释和智能问答等功能,支持100+编程语言,与主流编辑器无缝集成,显著提升开发效率和代码质量。

Project Cover

AI写歌

Suno AI是一个革命性的AI音乐创作平台,能在短短30秒内帮助用户创作出一首完整的歌曲。无论是寻找创作灵感还是需要快速制作音乐,Suno AI都是音乐爱好者和专业人士的理想选择。

Project Cover

有言AI

有言平台提供一站式AIGC视频创作解决方案,通过智能技术简化视频制作流程。无论是企业宣传还是个人分享,有言都能帮助用户快速、轻松地制作出专业级别的视频内容。

Project Cover

Kimi

Kimi AI助手提供多语言对话支持,能够阅读和理解用户上传的文件内容,解析网页信息,并结合搜索结果为用户提供详尽的答案。无论是日常咨询还是专业问题,Kimi都能以友好、专业的方式提供帮助。

Project Cover

阿里绘蛙

绘蛙是阿里巴巴集团推出的革命性AI电商营销平台。利用尖端人工智能技术,为商家提供一键生成商品图和营销文案的服务,显著提升内容创作效率和营销效果。适用于淘宝、天猫等电商平台,让商品第一时间被种草。

Project Cover

吐司

探索Tensor.Art平台的独特AI模型,免费访问各种图像生成与AI训练工具,从Stable Diffusion等基础模型开始,轻松实现创新图像生成。体验前沿的AI技术,推动个人和企业的创新发展。

Project Cover

SubCat字幕猫

SubCat字幕猫APP是一款创新的视频播放器,它将改变您观看视频的方式!SubCat结合了先进的人工智能技术,为您提供即时视频字幕翻译,无论是本地视频还是网络流媒体,让您轻松享受各种语言的内容。

Project Cover

美间AI

美间AI创意设计平台,利用前沿AI技术,为设计师和营销人员提供一站式设计解决方案。从智能海报到3D效果图,再到文案生成,美间让创意设计更简单、更高效。

Project Cover

AIWritePaper论文写作

AIWritePaper论文写作是一站式AI论文写作辅助工具,简化了选题、文献检索至论文撰写的整个过程。通过简单设定,平台可快速生成高质量论文大纲和全文,配合图表、参考文献等一应俱全,同时提供开题报告和答辩PPT等增值服务,保障数据安全,有效提升写作效率和论文质量。

投诉举报邮箱: service@vectorlightyear.com
@2024 懂AI·鲁ICP备2024100362号-6·鲁公网安备37021002001498号