轻量封装WebGPU渲染系统示例<24>- Rendering Pass Graph基本用法(源码)

当前示例源码github地址:

https://github.com/vilyLei/voxwebgpu/blob/feature/rendering/src/voxgpu/sample/PassNodeGraphTest.ts

此示例基于此渲染系统实现,当前示例TypeScript源码如下:

javascript 复制代码
class PassGraph extends WGRPassNodeGraph {
	constructor() { super(); }
	runBegin(): void {
		super.runBegin();
	}
	run(): void {
		let pass = this.passes[0];
		for (let i = 0; i < 16; ++i) {
			pass.colorAttachments[0].clearEnabled = i < 1;
			pass.render();
		}
	}
}

export class PassNodeGraphTest {
	private mRscene = new RendererScene();

	initialize(): void {

		let multisampleEnabled = true;
		let depthTestEnabled = false;
		let rpassparam = { multisampleEnabled, depthTestEnabled };
		this.mRscene.initialize({ rpassparam });
		this.initEvent();
		this.initScene();
	}

	private applyNewRPass(texUUID: string, entities: FixScreenPlaneEntity[], clearColor: ColorDataType, extent = [0.4, 0.3, 0.5, 0.5]): void {

		let rs = this.mRscene;
		let rttTex = { diffuse: { uuid: texUUID, rttTexture: {} } };
		let colorAttachments = [
			{
				texture: rttTex,
				clearValue: clearColor,
				loadOp: "clear",
				storeOp: "store"
			}
		];
		let rPass = rs.renderer.appendRenderPass({ separate: true, colorAttachments });
		for (let i = 0; i < entities.length; ++i) {
			rPass.addEntity(entities[i]);
		}

		let graph = new PassGraph();
		graph.passes = [rPass];
		rs.setPassNodeGraph(graph);

		let entity = new FixScreenPlaneEntity({ extent, flipY: true, textures: [rttTex] });
		rs.addEntity(entity);
	}
	private initEvent(): void {
		const rs = this.mRscene;
		new MouseInteraction().initialize(rs, 0, false).setAutoRunning(true);
	}
	private initScene(): void {

		const rs = this.mRscene;
		let entity: FixScreenPlaneEntity;

		const diffuseTex = { diffuse: { url: "static/assets/guangyun_40.png", flipY: true } };

		let blendModes = ['add'];
		let entities: FixScreenPlaneEntity[] = [];
		entity = new FixScreenPlaneEntity({ extent: [-0.8, -0.8, 0.8, 0.8], textures: [diffuseTex], blendModes });
		entity.setColor([0.9, 0.3, 0.9]);
		rs.addEntity(entity);
		entities.push(entity);

		this.applyNewRPass('rtt0', entities, [0.2, 0.2, 0.2, 1.0], [-0.2, 0.1, 0.8, 0.8]);
	}

	run(): void {
		this.mRscene.run();
	}
}
相关推荐
Solange__1 小时前
航天科普怎么用AI生成3D场景?从科学约束、尺度分层到网页展示的完整方案
图像处理·人工智能·游戏·3d·ai·自然语言处理·关卡设计
Carl_奕然3 小时前
【机器视觉】最强算法之:3D高斯泼溅(3D Gaussian Splatting)
人工智能·python·算法·目标检测·计算机视觉·3d
河图洛水8 小时前
UniDex-VLA:用人类第一视角视频,训出能跨 8 种灵巧手的 3D VLA
3d·大模型·具身智能·vla·ego
云飞云共享云桌面9 小时前
苏州半导体工厂研发部门多个SolidWorks 如何共享服务器来进行研发设计
运维·服务器·3d·自动化·制造
坚定信念,勇往无前9 小时前
3d建模整个流程-完整工作流总结(Maya + Marmoset Toolbag 八猴 + Substance Painter + Unity URP)
3d
FansAI_art10 小时前
3D潮玩手办的材质还原与多角色同屏防崩坏
3d·材质
Solange__10 小时前
AI模型编辑器、Blender和3D Agent怎么选?按5类修改任务比较局部编辑、拓扑与自动化
人工智能·3d·自然语言处理·编辑器·音视频·blender·游戏策划
杀生丸学AI11 小时前
【世界模型】AEON:统一视频与 3D 世界模型
3d·三维重建·空间智能·动态重建·大场景重建·流式重建·前馈式重建
GlobalInfo1 天前
全球包装检测用3D轮廓传感器市场调研:2026年规模预计达1.87亿美元,食品医药与电商物流需求推动增长
大数据·人工智能·3d
AI浩1 天前
BEVNeXt:复兴用于3D目标检测的稠密BEV框架
人工智能·目标检测·3d