GLTF数据格式解析，自制模型转换工具
GLTF是新的一种高效传输和加载3D场景和模型的格式。可以支持很多实现三维的引擎。
GLTF是由Khronos Group设计和规定的。
GLTF其实就是一个json文件，描述的就是3D场景数据的组成和构造。
GLTF最外层的属性有：
1.scenes,nodes,cameras,animations(这些属性描述的就是模型的基本结构）
2.mesh,textures,images,samplers(这些属性描述的模型的纹理信息，用于展示场景）
3.buffer,bufferViews,accessors(这些属性描述的是场景、模型的点的信息，点的位置，法线和纹理坐标值）
4.materials，techniques,programs,shaders(这些属性是场景/模型的材质信息，用于渲染3D模型）
5.skins(结合点的表面信息？？没用过，给的都是空值）
说到这里，给大家看一个gltf的文件。
{ "accessors": { "accessor_index0": { "bufferView": "bufferView_index", "byteOffset": 0, "byteStride": 0, "componentType": 5123, "count": 231432, "type": "SCALAR" }, "accessor_index1": { "bufferView": "bufferView_index", "byteOffset": 462864, "byteStride": 0, "componentType": 5123, "count": 1020, "type": "SCALAR" }, "accessor_position": { "bufferView": "bufferView_vertex", "byteOffset": 0, "byteStride": 12, "componentType": 5126, "count": 58900, "max": [ 308.304321, 159.551575, 145.747 ], "min": [ 179.30542, -72.13728, 71.62281 ], "type": "VEC3" }, "accessor_normal": { "bufferView": "bufferView_vertex", "byteOffset": 706800, "byteStride": 12, "componentType": 5126, "count": 58900, "max": [ 1.0, 1.0, 1.0 ], "min": [ 1.0, 1.0, 1.0 ], "type": "VEC3" }, "accessor_uv": { "bufferView": "bufferView_vertex", "byteOffset": 1413600, "byteStride": 8, "componentType": 5126, "count": 58900, "max": [ 0.997774363, 0.8115263 ], "min": [ 0.00197040662, 0.00195908546 ], "type": "VEC2" } }, "animations": {}, "asset": { "generator": "infoearth", "premultipliedAlpha": true, "profile": { "api": "webgl", "version": "1.0.2" }, "version": 1 }, "bufferViews": { "bufferView_index": { "buffer": "model_buffer", "byteLength": 464904, "byteOffset": 0, "target": 34963 }, "bufferView_vertex": { "buffer": "model_buffer", "byteLength": 1884800, "byteOffset": 464904, "target": 34962 } }, "buffers": { "model_buffer": { "byteLength": 2349704, "type": "arraybuffer", "uri": "Tile_+004_+004.bin" } }, "images": { "texture0": { "name": "texture0", "uri": "Tile_+004_+004_0.jpg" }, "texture1": { "name": "texture1", "uri": "notexture.png" } }, "materials": { "material0": { "name": "material0", "technique": "technique0", "values": { "diffuse": "texture_texture0" } }, "material1": { "name": "material1", "technique": "technique0", "values": { "diffuse": "texture_texture1" } } }, "meshes": { "mesh_model": { "name": "mesh_model", "primitives": [ { "attributes": { "NORMAL": "accessor_normal", "POSITION": "accessor_position", "TEXCOORD_0": "accessor_uv" }, "indices": "accessor_index0", "material": "material0", "mode": 4 }, { "attributes": { "NORMAL": "accessor_normal", "POSITION": "accessor_position", "TEXCOORD_0": "accessor_uv" }, "indices": "accessor_index1", "material": "material1", "mode": 4 } ] } }, "nodes": { "node": { "children": [], "matrix": [ 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1 ], "meshes": [ "mesh_model" ], "name": "node" } }, "programs": { "program_0": { "attributes": [ "a_normal", "a_position", "a_texcoord0" ], "fragmentShader": "FS", "vertexShader": "VS" } }, "samplers": { "sampler_0": { "magFilter": 9729, "minFilter": 9729, "wrapS": 10497, "wrapT": 10497 } }, "scene": "defaultScene", "scenes": { "defaultScene": { "nodes": [ "node" ] } }, "shaders": { "FS": { "type": 35632, "uri": "FS.glsl" }, "VS": { "type": 35633, "uri": "VS.glsl" } }, "skins": {}, "techniques": { "technique0": { "attributes": { "a_normal": "normal", "a_position": "position", "a_texcoord0": "texcoord0" }, "parameters": { "diffuse": { "type": 35678 }, "modelViewMatrix": { "semantic": "MODELVIEW", "type": 35676 }, "normal": { "semantic": "NORMAL", "type": 35665 }, "normalMatrix": { "semantic": "MODELVIEWINVERSETRANSPOSE", "type": 35675 }, "position": { "semantic": "POSITION", "type": 35665 }, "projectionMatrix": { "semantic": "PROJECTION", "type": 35676 }, "texcoord0": { "semantic": "TEXCOORD_0", "type": 35664 } }, "program": "program_0", "states": { "enable": [ 2929, 2884 ] }, "uniforms": { "u_diffuse": "diffuse", "u_modelViewMatrix": "modelViewMatrix", "u_normalMatrix": "normalMatrix", "u_projectionMatrix": "projectionMatrix" } } }, "textures": { "texture_texture0": { "format": 6408, "internalFormat": 6408, "sampler": "sampler_0", "source": "texture0", "target": 3553, "type": 5121 }, "texture_texture1": { "format": 6408, "internalFormat": 6408, "sampler": "sampler_0", "source": "texture1", "target": 3553, "type": 5121 } } }
accessors属性里面存放的是点的信息。其中accessor_index0和accessor_index1存储的是点的索引，数据类型是scaler,bufferView属性对应的是存储点的值存储地，可在bufferView里面去找对应的id的buffer信息。byteOffset是当前索引值对应bufferView的偏移量，如果是索引，byteStride是0，如果是点的位置坐标，byteStride是3*4即12。count是数据的总个数，componentType是点存储的字节数，如果顶点的个数小于65535，及componentType用两个字节就可以，用5123，否则就四个字节用5126。对应的其他属性可以以此类推。
关于GLTF的其他属性的值就不一一介绍了，根据上面的例子可以大概看明白，不明白的也可以问我。
由于场景数据没有运动，所以没有给animations的值。