Parameters
- engine: Engine
- isVertexStage: boolean
Returns
| ""
| "\n"
| "\n return g_distance;\n}\nfn calcNormal(p: vec3f) -> vec3f {\n let e = vec2f(1.0,-1.0)*.0005;\n return vec3f(\n normalize(\n e.xyy*map(p+e.xyy)+\n e.yyx*map(p+e.yyx)+\n e.yxy*map(p+e.yxy)+\n e.xxx*map(p+e.xxx)\n )\n );\n}\n\nstruct FragmentOutput {\n @location(0) color: vec4<f32>,\n @builtin(frag_depth) depth: f32,\n}\n\nvar<private> rt0: vec4<f32> = vec4<f32>(0.0, 0.0, 0.0, 1.0);\n@fragment\nfn main(\n input: VertexOutput,\n) -> FragmentOutput {\n var output: FragmentOutput;\n let cameraSID = uniformDrawParameters.cameraSID;\n let viewMatrix = get_viewMatrix(cameraSID);\n let projectionMatrix = get_projectionMatrix(cameraSID);\n\n // Calculate inverse matrices for ray generation\n let invViewMatrix = inverseMat4(viewMatrix);\n let invProjectionMatrix = inverseMat4(projectionMatrix);\n\n // Ray origin is the camera position in world space (extracted from inverse view matrix)\n let ro = invViewMatrix[3].xyz;\n\n // Calculate ray direction from screen coordinates\n var uv = (input.texcoord_0 - 0.5) * 2.0; // NDC coordinates (-1 to 1)\n uv.y = -uv.y; // flip y coordinate in WebGPU because of the coordinate system difference between WebGL and WebGPU\n // Transform from NDC to view space using inverse projection matrix\n let rayClip = vec4f(uv, -1.0, 1.0);\n var rayView = invProjectionMatrix * rayClip;\n rayView = vec4f(rayView.xy, -1.0, 0.0); // Set z to -1 (forward direction in view space)\n // Transform from view space to world space\n let rd = normalize((invViewMatrix * rayView).xyz);\n\n // March the distance field until a surface is hit.\n var h: f32;\n var t: f32 = 1.0;\n for(var i=0;i<256;i++){\n h=map(ro+rd*t);\n t+=h;\n if(h<.01) { break; }\n }\n\n if(h<.01){\n let p = ro+rd*t;\n let normal=calcNormal(p);\n let light=vec3f(0,2,0);\n\n // Calculate diffuse lighting by taking the dot product of\n // the light direction (light-p) and the normal.\n var dif=clamp(dot(normal,normalize(light-p)),0.,1.);\n\n // Multiply by light intensity (5) and divide by the square\n // of the distance to the light.\n dif*=5.0/dot(light-p,light-p);\n\n rt0=vec4f(vec3f(pow(dif,.4545)),1.0);// Gamma correction\n\n // Calculate depth from raymarching hit position\n let clipPos = projectionMatrix * viewMatrix * vec4f(p, 1.0);\n output.depth = clipPos.z / clipPos.w; // WebGPU depth range is [0, 1]\n }else{\n rt0=vec4f(0.0,0.0,0.0,1.0);\n output.depth = 1.0; // Maximum depth for background\n }\n\n output.color = rt0;\n return output;\n}\n"
| "\n return g_distance;\n}\nvec3 calcNormal(vec3 p){\n vec2 e=vec2(1.,-1.)*.0005;\n return normalize(\n e.xyy*map(p+e.xyy)+\n e.yyx*map(p+e.yyx)+\n e.yxy*map(p+e.yxy)+\n e.xxx*map(p+e.xxx)\n );\n}\n\nvoid main() {\n uint cameraSID = uint(u_currentComponentSIDs[/* shaderity: @{WellKnownComponentTIDs.CameraComponentTID} */]);\n #if defined(WEBGL2_MULTI_VIEW) && defined(RN_IS_VERTEX_SHADER)\n cameraSID += uint(gl_ViewID_OVR);\n #endif\n\n mat4 viewMatrix = get_viewMatrix(cameraSID);\n mat4 projectionMatrix = get_projectionMatrix(cameraSID);\n\n // Calculate inverse matrices for ray generation\n mat4 invViewMatrix = inverse(viewMatrix);\n mat4 invProjectionMatrix = inverse(projectionMatrix);\n\n // Ray origin is the camera position in world space (extracted from inverse view matrix)\n vec3 ro = invViewMatrix[3].xyz;\n\n // Calculate ray direction from screen coordinates\n vec2 uv = (v_texcoord_0 - 0.5) * 2.0; // NDC coordinates (-1 to 1)\n // Transform from NDC to view space using inverse projection matrix\n vec4 rayClip = vec4(uv, -1.0, 1.0);\n vec4 rayView = invProjectionMatrix * rayClip;\n rayView = vec4(rayView.xy, -1.0, 0.0); // Set z to -1 (forward direction in view space)\n // Transform from view space to world space\n vec3 rd = normalize((invViewMatrix * rayView).xyz);\n\n // March the distance field until a surface is hit.\n float h,t=1.;\n for(int i=0;i<256;i++){\n h=map(ro+rd*t);\n t+=h;\n if(h<.01)break;\n }\n\n if(h<.01){\n vec3 p=ro+rd*t;\n vec3 normal=calcNormal(p);\n vec3 light=vec3(0,2,0);\n\n // Calculate diffuse lighting by taking the dot product of\n // the light direction (light-p) and the normal.\n float dif=clamp(dot(normal,normalize(light-p)),0.,1.);\n\n // Multiply by light intensity (5) and divide by the square\n // of the distance to the light.\n dif*=5./dot(light-p,light-p);\n\n rt0=vec4(vec3(pow(dif,.4545)),1);// Gamma correction\n\n // Calculate depth from raymarching hit position\n vec4 clipPos = projectionMatrix * viewMatrix * vec4(p, 1.0);\n // Convert from NDC [-1, 1] to depth buffer range [0, 1]\n gl_FragDepth = (clipPos.z / clipPos.w) * 0.5 + 0.5;\n }else{\n rt0=vec4(0,0,0,1);\n gl_FragDepth = 1.0; // Maximum depth for background\n }\n}\n"
The shader code string for the main function ending
RaymarchingShaderPart is a class that provides common shader functionality for both WebGL and WebGPU rendering approaches. This class handles shader code generation, vertex/fragment shader prerequisites, and cross-platform compatibility between WebGL and WebGPU shader languages (GLSL and WGSL).