Minimal three.js alternative.
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
.
The following creates a renderer, camera, and renders a red cube:
<details> <summary>Show WebGL example</summary></details> <details> <summary>Show WebGPU 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>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)
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 })
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
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
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()
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)
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
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, }
A Material
describes a program or shader interface for rasterization and compute (see compute), defining a vertex
and fragment
or compute
shader, respectively.
</details> <details> <summary>Show WebGPU 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>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, })
The following uniforms are built-in and will be automatically populated when specified:
Type | Name | Description | Conversion |
---|---|---|---|
mat4x4 | modelMatrix | world-space mesh transform | local space => world space |
mat4x4 | projectionMatrix | clip-space camera projection | view space => clip space |
mat4x4 | viewMatrix | inverse camera transform | world space => view space |
mat4x4 | modelViewMatrix | premultiplied model-view transform | local space => view space |
mat4x4 | normalMatrix | isotropic inverse model-view or "normal" transform | local 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>;
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 Mode | BlendOperation | BlendFactor (src) | BlendFactor (dst) |
---|---|---|---|
Additive | add | src-alpha | one |
Subtractive | reverse-subtract | src-alpha | one |
Multiply | add |
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小说写作助手,一站式润色、改写、扩写
蛙蛙写作—国内先进的AI写作平台,涵盖小说、学术、社交媒体等多场景。提供续写、改写、润色等功能,助力创作者高效优化写作流程。界面简洁,功能全面,适合各类写作者提升内容品质和工作效率。
全能AI智能助手,随时解答生活与工作的多样问题
问小白,由元石科技研发的AI智能助手,快速准确地解答各种生活和工作问题,包括但不限于搜索、规划和社交互动,帮助用户在日常生活中提高效率,轻松管理个人事务。
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。
一键生成PPT和Word,让学习生活更轻松
讯飞智文是一个利用 AI 技术的项目,能够帮助用户生成 PPT 以及各类文档。无论是商业领域的市场分析报告、年度目标制定,还是学生群体的职业生涯规划、实习避坑指南,亦或是活动策划、旅游攻略等内容,它都能提供支持,帮助用户精准表达,轻松呈现各种信息。
深度推理能力全新升级,全面对标OpenAI o1
科大讯飞的星火大模型,支持语言理解、知识问答和文本创作等多功能,适用于多种文件和业务场景,提升办公和日常生活的效率。讯飞星火是一个提供丰富智能服务的平台,涵盖科技资讯、图像创作、写作辅助、编程解答、科研文献解读等功能,能为不同需求的用户提供便捷高效的帮助,助力用户轻松获取信息、解决问题,满足多样化使用场景。
一种基于大语言模型的高效单流解耦语音令牌文本到语音合成模型
Spark-TTS 是一个基于 PyTorch 的开源文本到语音合成项目,由多个知名机构联合参与。该项目提供了高效的 LLM(大语言模型)驱动的语音合成方案,支持语音克隆和语音创建功能,可通过命令行界面(CLI)和 Web UI 两种方式使用。用户可以根据需求调整语音的性别、音高、速度等参数,生成高质量的语音。该项目适用于多种场景,如有声读物制作、智能语音助手开发等。
AI助力,做PPT更简单!
咔片是一款轻量化在线演示设计工具,借助 AI 技术,实现从内容生成到智能设计的一站式 PPT 制作服务。支持多种文档格式导入生成 PPT,提供海量模板、智能美化、素材替换等功能,适用于销售、教师、学生等各类人群,能高效制作出高品质 PPT,满足不同场景演示需求。
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。
专业的AI公文写作平台,公文写作神器
AI 材料星,专业的 AI 公文写作辅助平台,为体制内工作人员提供高效的公文写作解决方案。拥有海量公文文库、9 大核心 AI 功能,支持 30 + 文稿类型生成,助力快速完成领导讲话、工作总结、述职报告等材料,提升办公效率,是体制打工人的得力写作神器。
最新AI工具、AI资讯
独家AI资源、AI项目落地
微信扫一扫关注公众号