在鼠标移动上创建涂抹/液化效果，使用 webgl 连续动画回原始状态 [英] Creating a smudge/liquify effect on mouse move that continuously animates back to the original state using webgl
问题描述
我正在尝试查找可用于创建持续动画回到原始状态的涂抹/液化效果的信息或示例.
I am trying to find information or examples that I can use to create a smudge/liquify effect that continuously animates back to the original state.
最初我正在考虑使用 three.js 或 pixi.js 来渲染一些文本，然后使用鼠标事件和光线投射将网格拖出位置，我发现最接近的是这个.
Initially I was looking at using three.js or pixi.js to render some text and then use mouse events and ray casting to drag the mesh out of position, the closest thing I have found is this.
https://codepen.io/shshaw/pen/qqVgbg
let renderer = PIXI.autoDetectRenderer(window.innerWidth,
window.innerHeight, { transparent: true });
我认为理想情况下，我会将文本渲染为图像，然后将涂抹效果应用于像素，它们会慢慢动画回其原始状态.与此类似.
I think that ideally I would render the text as an image and then the smudge effect would be applied to the pixels and they would slowly animate back to their original states. Similar to this.
http://www.duhaihang.com/#/work/
我想我可能需要使用自定义 GLSL 着色器和某种缓冲区来保存构成图像的像素的原始状态和当前状态.
I think I may need to use a custom GLSL shader and some kind of buffer to hold the original and the current state of the pixels making up the image.
任何帮助或指导将不胜感激.
Any help or direction would be much appreciated.
推荐答案
两者看起来都比较简单.
Both seem relatively straightforward.
第一个，就像你提到的那样，你制作了一个绘制平面的顶点网格(网格).您将面纹理映射到平面，当您四处拖动鼠标时，为鼠标接触的每个顶点添加位移.随着时间的推移，将位移重置为 0(如 0 位移量)
The first one, like you mentioned, you make a mesh (grid) of vertices that draw a plane. You texture map the face to the plane, as you drag the mouse around add a displacement to the each vertex the mouse touches. Over time reset the displacement back to 0 (as in 0 amount of displacement)
这里有一个例子:它只是随机替换单个顶点而不是更可预测的东西.最后，我只是节省了位移应该淡出的时间，然后在着色器中我做了一个简单的线性 lerp(可以使用更高级的 lerp 来实现反弹或其他东西).几乎所有事情都发生在着色器中.
here's an example: It's only displacing a single vertex a random amount instead of something more predictable. Finally I'm just saving the time at which the displacement should fade out by, then in the shader I do a simple linear lerp (could use a fancier lerp for a bounce or something). This is so pretty much everything happens in the shader.
const m4 = twgl.m4;
const gl = document.querySelector("canvas").getContext("webgl");
const vs = `
attribute vec4 position;
attribute vec3 displacement;
uniform mat4 u_matrix;
uniform float u_time;
uniform float u_timeToGoBack;
varying vec2 v_texcoord;
void main() {
// because position goes -1 <-> 1 we can just use
// it for texture coords
v_texcoord = position.xy * .5 + .5;
// displacement.z is the time at which it should be undisplaced
float displaceTime = displacement.z - u_time;
float lerp = clamp(displaceTime / u_timeToGoBack, 0., 1.);
vec2 displace = displacement.xy * lerp;
gl_Position = u_matrix * (position + vec4(displace, 0, 0));
}
`;
const fs = `
precision mediump float;
uniform sampler2D texture;
varying vec2 v_texcoord;
void main() {
gl_FragColor = texture2D(texture, v_texcoord);
}
`;
const programInfo = twgl.createProgramInfo(gl, [vs, fs]);
// create a grid of points in a -1 to +1 quad
const positions = [];
const displacements = [];
const indices = [];
const res = 100;
for (var y = 0; y < res; ++y) {
var v = (y / (res - 1)) * 2 - 1;
for (var x = 0; x < res; ++x) {
var u = (x / (res - 1)) * 2 - 1;
positions.push(u, v);
displacements.push(0, 0, 0);
}
}
for (var y = 0; y < res - 1; ++y) {
var off0 = (y + 0) * res;
var off1 = (y + 1) * res;
for (var x = 0; x < res - 1; ++x) {
indices.push(
off0 + x + 0, off0 + x + 1, off1 + x + 0,
off1 + x + 0, off0 + x + 1, off1 + x + 1
);
}
}
// create buffers and fills them in.
// (calls gl.createBuffer and gl.bufferData for each array)
const bufferInfo = twgl.createBufferInfoFromArrays(gl, {
position: { numComponents: 2, data: positions, },
displacement: { numComponents: 3, data: displacements, },
indices: indices,
});
// this will be replaced when the image has loaded;
var img = { width: 1, height: 1 };
const tex = twgl.createTexture(gl, {
src: 'https://farm6.staticflickr.com/5078/14032935559_8c13e9b181_z_d.jpg',
crossOrigin: '',
}, function(err, texture, source) {
img = source;
});
var currentTime = 0;
var currentMatrix;
const timeToGoBack = 2; // in seconds;
function render(time) {
time *= 0.001; // convert to seconds
currentTime = time;
twgl.resizeCanvasToDisplaySize(gl.canvas);
gl.viewport(0, 0, gl.canvas.width, gl.canvas.height);
gl.useProgram(programInfo.program);
var aspect = img.width / img.height;
var mat = m4.ortho(0, gl.canvas.clientWidth, gl.canvas.clientHeight, 0, -1, 1);
mat = m4.translate(mat, [gl.canvas.clientWidth / 2, gl.canvas.clientHeight / 2, 0]);
mat = m4.scale(mat, [img.width * .25, img.height * .25, 1]);
currentMatrix = mat;
// calls gl.bindBuffer, gl.vertexAttribPointer to setup
// attributes
twgl.setBuffersAndAttributes(gl, programInfo, bufferInfo);
twgl.setUniforms(programInfo, {
u_matrix: mat,
u_texture: tex,
u_time: currentTime,
u_timeToGoBack: timeToGoBack,
});
gl.drawElements(gl.TRIANGLES, bufferInfo.numElements, gl.UNSIGNED_SHORT, 0);
requestAnimationFrame(render);
}
requestAnimationFrame(render);
const displace = new Float32Array(3);
gl.canvas.addEventListener('mousemove', function(event, target) {
target = target || event.target;
const rect = target.getBoundingClientRect();
const rx = event.clientX - rect.left;
const ry = event.clientY - rect.top;
const x = rx * target.width / target.clientWidth;
const y = ry * target.height / target.clientHeight;
// reverse project the mouse onto the image
var rmat = m4.inverse(currentMatrix);
var s = m4.transformPoint(
rmat, [x / target.width * 2 - 1, y / target.height * 2 - 1, 0]);
// s is now a point in the space of `position`
// lets just move closest point?
var gx = Math.round((s[0] * .5 + .5) * res);
var gy = Math.round((s[1] * .5 + .5) * res);
gx = clamp(gx, 0, res - 1);
gy = clamp(gy, 0, res - 1);
const offset = ((res - gy - 1) * res + gx) * 3 * 4;
displace[0] = rand(-.1, .1);
displace[1] = rand(-.1, .1);
displace[2] = currentTime + timeToGoBack;
gl.bindBuffer(gl.ARRAY_BUFFER, bufferInfo.attribs.displacement.buffer);
gl.bufferSubData(gl.ARRAY_BUFFER, offset, displace);
});
function rand(min, max) {
return Math.random() * (max - min) + min;
}
function clamp(v, min, max) {
return Math.max(min, Math.min(max, v));
}body {
margin: 0;
}
canvas {
width: 100vw;
height: 100vh;
display: block;
}<script src="https://twgljs.org/dist/4.x/twgl-full.min.js"></script>
<canvas></canvas>对于第二个而不是置换顶点，您制作置换纹理，随着时间的推移，您将该置换重置回 0
For the second one instead of displacing vertices you make a displacement texture, over time you reset that displacement back to 0
您可以在此处查看淡出的示例.如果您采用该样本而不是在鼠标下绘制随机正方形，则使用该纹理作为主图像的置换.置换是指通常你在像这样的片段着色器中查找纹理
You can see an example of fading things out here. If you took that sample and instead of drawing random square you draw under the mouse, then use that texture as a displacement to your main image. By displacement I mean normally you look up a texture in a fragment shader like this
vec4 color = texture2D(someTexture, someTextureCoords);
相反，您想用位移来置换顶点坐标，就像这样
Instead you want to displace the vertex coords with a displacement, something like this
// assuming the displacement texture is the same size as
// the main texture you can use the same texture coords
// first look up the displacement and convert to -1 <-> 1 range
// we're only using the R and G channels which will become U and V
// displacements to our texture coordinates
vec2 displacement = texture2D(displacementTexture, someTextureCoords).rg * 2. - 1.;
vec2 uv = someTextureCoords + displacement * displacementRange;
vec4 color = texture2d(someTexture, uv);
这是上面链接的用于位移的示例
Here's the sample linked above being used for displacement
var vs = `
attribute vec4 position;
uniform mat4 u_matrix;
void main() {
gl_Position = u_matrix * position;
}
`;
var fs = `
precision mediump float;
uniform vec4 u_color;
void main() {
gl_FragColor = u_color;
}
`;
var vsQuad = `
attribute vec4 position;
attribute vec2 texcoord;
varying vec2 v_texcoord;
void main() {
gl_Position = position;
v_texcoord = texcoord;
}
`;
var fsFade = `
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D u_texture;
uniform float u_mixAmount;
const float kEpsilon = 2./256.;
void main() {
// convert color from 0.->1. to -1. -> +1. so we can go adjust toward zero
vec4 color = texture2D(u_texture, v_texcoord) * 2. - 1.;
// figure out how much to adjust
vec4 adjust = -color * u_mixAmount;
// If the adjustment is too small (because the texture is only 8bits)
// the adjust the minimum amount.
// Could also solve this by using floating point textures
adjust = mix(adjust, sign(color) * -kEpsilon, step(abs(adjust), vec4(kEpsilon)));
// adjust it
color += adjust;
// write it back converting back to 0 -> 1
gl_FragColor = color * .5 + .5;
}
`;
var fsDisplace = `
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D u_texture;
uniform sampler2D u_displacementTexture;
uniform vec2 u_displacementRange;
void main() {
// assuming the displacement texture is the same size as
// the main texture you can use the same texture coords
// first look up the displacement and convert to -1 <-> 1 range
// we're only using the R and G channels which will become U and V
// displacements to our texture coordinates
vec2 displacement = texture2D(u_displacementTexture, v_texcoord).rg * 2. - 1.;
vec2 uv = v_texcoord + displacement * u_displacementRange;
gl_FragColor = texture2D(u_texture, uv);
}
`;
var $ = document.querySelector.bind(document);
var mixAmount = 0.03;
var gl = $("canvas").getContext("webgl");
var m4 = twgl.m4;
var programInfo = twgl.createProgramInfo(gl, [vs, fs]);
var fadeProgramInfo = twgl.createProgramInfo(gl, [vsQuad, fsFade]);
var displaceProgramInfo = twgl.createProgramInfo(gl, [vsQuad, fsDisplace]);
// this will be replaced when the image has loaded;
var img = { width: 1, height: 1 };
const tex = twgl.createTexture(gl, {
src: 'https://farm6.staticflickr.com/5078/14032935559_8c13e9b181_z_d.jpg',
crossOrigin: '',
flipY: true,
}, function(err, texture, source) {
img = source;
});
// Creates a -1 to +1 quad
var quadBufferInfo = twgl.primitives.createXYQuadBufferInfo(gl);
// Creates 2 RGBA texture + depth framebuffers
var fadeAttachments = [
{ format: gl.RGBA, min: gl.NEAREST, max: gl.NEAREST, wrap: gl.CLAMP_TO_EDGE, },
{ format: gl.DEPTH_STENCIL },
];
var fadeFbi1 = twgl.createFramebufferInfo(gl, fadeAttachments);
var fadeFbi2 = twgl.createFramebufferInfo(gl, fadeAttachments);
function drawThing(gl, x, y, rotation, scale, color) {
var matrix = m4.ortho(0, gl.canvas.width, gl.canvas.height, 0, -1, 1);
matrix = m4.translate(matrix, [x, y, 0]);
matrix = m4.rotateZ(matrix, rotation);
matrix = m4.scale(matrix, [scale, scale, 1]);
gl.useProgram(programInfo.program);
twgl.setBuffersAndAttributes(gl, programInfo, quadBufferInfo);
twgl.setUniforms(programInfo, {
u_matrix: matrix,
u_color: color,
});
twgl.drawBufferInfo(gl, quadBufferInfo);
}
function rand(min, max) {
if (max === undefined) {
max = min;
min = 0;
}
return min + Math.random() * (max - min);
}
function render(time) {
if (twgl.resizeCanvasToDisplaySize(gl.canvas)) {
// set the clear color to 0.5 which is 0 displacement
// for our shader
gl.clearColor(0.5, 0.5, 0.5, 0.5);
// resize the framebuffer's attachments so their the
// same size as the canvas
twgl.resizeFramebufferInfo(gl, fadeFbi1, fadeAttachments);
// clear the color buffer to 0.5
twgl.bindFramebufferInfo(gl, fadeFbi1);
gl.clear(gl.COLOR_BUFFER_BIT);
// resize the 2nd framebuffer's attachments so their the
// same size as the canvas
twgl.resizeFramebufferInfo(gl, fadeFbi2, fadeAttachments);
// clear the color buffer to 0.5
twgl.bindFramebufferInfo(gl, fadeFbi2);
gl.clear(gl.COLOR_BUFFER_BIT);
}
// fade by copying from fadeFbi1 into fabeFbi2 using mixAmount.
// fadeFbi2 will contain mix(fadeFb1, u_fadeColor, u_mixAmount)
twgl.bindFramebufferInfo(gl, fadeFbi2);
gl.useProgram(fadeProgramInfo.program);
twgl.setBuffersAndAttributes(gl, fadeProgramInfo, quadBufferInfo);
twgl.setUniforms(fadeProgramInfo, {
u_texture: fadeFbi1.attachments[0],
u_mixAmount: mixAmount,
});
twgl.drawBufferInfo(gl, quadBufferInfo);
// now draw new stuff to fadeFb2. Notice we don't clear!
twgl.bindFramebufferInfo(gl, fadeFbi2);
var x = rand(gl.canvas.width);
var y = rand(gl.canvas.height);
var rotation = rand(Math.PI);
var scale = rand(10, 20);
var color = [rand(1), rand(1), rand(1), 1];
drawThing(gl, x, y, rotation, scale, color);
// now use fadeFbi2 as a displacement while drawing tex to the canvas
twgl.bindFramebufferInfo(gl, null);
gl.useProgram(displaceProgramInfo.program);
twgl.setBuffersAndAttributes(gl, displaceProgramInfo, quadBufferInfo);
twgl.setUniforms(displaceProgramInfo, {
u_texture: tex,
u_displacementTexture: fadeFbi2.attachments[0],
u_displacementRange: [0.1, 0.1],
});
twgl.drawBufferInfo(gl, quadBufferInfo);
// swap the variables so we render to the opposite textures next time
var temp = fadeFbi1;
fadeFbi1 = fadeFbi2;
fadeFbi2 = temp;
requestAnimationFrame(render);
}
requestAnimationFrame(render);body { margin: 0; }
canvas { display: block; width: 100vw; height: 100vh; }<script src="https://twgljs.org/dist/4.x/twgl-full.min.js"></script>
<canvas></canvas>所以剩下的就是让它在鼠标下绘制而不是随机绘制
So all that's left is to make it draw under the mouse instead of at random
var vs = `
attribute vec4 position;
uniform mat4 u_matrix;
void main() {
gl_Position = u_matrix * position;
}
`;
var fs = `
precision mediump float;
uniform vec4 u_color;
void main() {
gl_FragColor = u_color;
}
`;
var vsQuad = `
attribute vec4 position;
attribute vec2 texcoord;
uniform mat4 u_matrix;
varying vec2 v_texcoord;
void main() {
gl_Position = u_matrix * position;
v_texcoord = texcoord;
}
`;
var fsFade = `
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D u_texture;
uniform float u_mixAmount;
const float kEpsilon = 2./256.;
void main() {
vec4 color = texture2D(u_texture, v_texcoord) * 2. - 1.;
vec4 adjust = -color * u_mixAmount;
adjust = mix(adjust, sign(color) * -kEpsilon, step(abs(adjust), vec4(kEpsilon)));
color += adjust;
gl_FragColor = color * .5 + .5;
}
`;
var fsDisplace = `
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D u_texture;
uniform sampler2D u_displacementTexture;
uniform vec2 u_displacementRange;
void main() {
// assuming the displacement texture is the same size as
// the main texture you can use the same texture coords
// first look up the displacement and convert to -1 <-> 1 range
// we're only using the R and G channels which will become U and V
// displacements to our texture coordinates
vec2 displacement = texture2D(u_displacementTexture, v_texcoord).rg * 2. - 1.;
vec2 uv = v_texcoord + displacement * u_displacementRange;
gl_FragColor = texture2D(u_texture, uv);
}
`;
var $ = document.querySelector.bind(document);
var mixAmount = 0.03;
var gl = $("canvas").getContext("webgl");
var m4 = twgl.m4;
var programInfo = twgl.createProgramInfo(gl, [vs, fs]);
var fadeProgramInfo = twgl.createProgramInfo(gl, [vsQuad, fsFade]);
var displaceProgramInfo = twgl.createProgramInfo(gl, [vsQuad, fsDisplace]);
// this will be replaced when the image has loaded;
var img = { width: 1, height: 1 };
const tex = twgl.createTexture(gl, {
src: 'https://farm6.staticflickr.com/5078/14032935559_8c13e9b181_z_d.jpg',
crossOrigin: '',
}, function(err, texture, source) {
img = source;
});
// Creates a -1 to +1 quad
var quadBufferInfo = twgl.primitives.createXYQuadBufferInfo(gl);
// Creates 2 RGBA texture + depth framebuffers
var fadeAttachments = [
{ format: gl.RGBA,
min: gl.NEAREST,
max: gl.NEAREST,
wrap: gl.CLAMP_TO_EDGE,
},
];
var fadeFbi1 = twgl.createFramebufferInfo(gl, fadeAttachments);
var fadeFbi2 = twgl.createFramebufferInfo(gl, fadeAttachments);
function drawThing(gl, x, y, rotation, scale, color) {
var matrix = m4.ortho(0, gl.canvas.width, gl.canvas.height, 0, -1, 1);
matrix = m4.translate(matrix, [x, y, 0]);
matrix = m4.rotateZ(matrix, rotation);
matrix = m4.scale(matrix, [scale, scale, 1]);
gl.useProgram(programInfo.program);
twgl.setBuffersAndAttributes(gl, programInfo, quadBufferInfo);
twgl.setUniforms(programInfo, {
u_matrix: matrix,
u_color: color,
});
twgl.drawBufferInfo(gl, quadBufferInfo);
}
function rand(min, max) {
if (max === undefined) {
max = min;
min = 0;
}
return min + Math.random() * (max - min);
}
var drawRect = false;
var rectX;
var rectY;
var currentMatrix;
function render(time) {
if (twgl.resizeCanvasToDisplaySize(gl.canvas)) {
// set the clear color to 0.5 which is 0 displacement
// for our shader
gl.clearColor(0.5, 0.5, 0.5, 0.5);
// resize the framebuffer's attachments so their the
// same size as the canvas
twgl.resizeFramebufferInfo(gl, fadeFbi1, fadeAttachments);
// clear the color buffer to 0.5
twgl.bindFramebufferInfo(gl, fadeFbi1);
gl.clear(gl.COLOR_BUFFER_BIT);
// resize the 2nd framebuffer's attachments so their the
// same size as the canvas
twgl.resizeFramebufferInfo(gl, fadeFbi2, fadeAttachments);
// clear the color buffer to 0.5
twgl.bindFramebufferInfo(gl, fadeFbi2);
gl.clear(gl.COLOR_BUFFER_BIT);
}
// fade by copying from fadeFbi1 into fabeFbi2 using mixAmount.
// fadeFbi2 will contain mix(fadeFb1, u_fadeColor, u_mixAmount)
twgl.bindFramebufferInfo(gl, fadeFbi2);
gl.useProgram(fadeProgramInfo.program);
twgl.setBuffersAndAttributes(gl, fadeProgramInfo, quadBufferInfo);
twgl.setUniforms(fadeProgramInfo, {
u_matrix: m4.identity(),
u_texture: fadeFbi1.attachments[0],
u_mixAmount: mixAmount,
});
twgl.drawBufferInfo(gl, quadBufferInfo);
if (drawRect) {
drawRect = false;
// now draw new stuff to fadeFb2. Notice we don't clear!
twgl.bindFramebufferInfo(gl, fadeFbi2);
var rotation = rand(Math.PI);
var scale = rand(10, 20);
var color = [rand(1), rand(1), rand(1), 1];
drawThing(gl, rectX, rectY, rotation, scale, color);
}
// now use fadeFbi2 as a displacement while drawing tex to the canvas
twgl.bindFramebufferInfo(gl, null);
var mat = m4.ortho(0, gl.canvas.clientWidth, gl.canvas.clientHeight, 0, -1, 1);
mat = m4.translate(mat, [gl.canvas.clientWidth / 2, gl.canvas.clientHeight / 2, 0]);
mat = m4.scale(mat, [img.width * 0.5, img.height * 0.5, 1]);
currentMatrix = mat;
gl.useProgram(displaceProgramInfo.program);
twgl.setBuffersAndAttributes(gl, displaceProgramInfo, quadBufferInfo);
twgl.setUniforms(displaceProgramInfo, {
u_matrix: mat,
u_texture: tex,
u_displacementTexture: fadeFbi2.attachments[0],
u_displacementRange: [0.05, 0.05],
});
twgl.drawBufferInfo(gl, quadBufferInfo);
// swap the variables so we render to the opposite textures next time
var temp = fadeFbi1;
fadeFbi1 = fadeFbi2;
fadeFbi2 = temp;
requestAnimationFrame(render);
}
requestAnimationFrame(render);
gl.canvas.addEventListener('mousemove', function(event, target) {
target = target || event.target;
const rect = target.getBoundingClientRect();
const rx = event.clientX - rect.left;
const ry = event.clientY - rect.top;
const x = rx * target.width / target.clientWidth;
const y = ry * target.height / target.clientHeight;
// reverse project the mouse onto the image
var rmat = m4.inverse(currentMatrix);
var clipspacePoint = [x / target.width * 2 - 1, -(y / target.height * 2 - 1), 0];
var s = m4.transformPoint(rmat, clipspacePoint);
// s is now a point in the space of the image's quad. The quad goes -1 to 1
// and we're going to draw into it using pixels because drawThing takes
// a pixel value and our displacement map is the same size as the canvas
drawRect = true;
rectX = ( s[0] * .5 + .5) * gl.canvas.width;
rectY = (-s[1] * .5 + .5) * gl.canvas.height;
});body { margin: 0; }
canvas { display: block; width: 100vw; height: 100vh; }<script src="https://twgljs.org/dist/4.x/twgl-full.min.js"></script>
<canvas></canvas>获得第二个示例的确切效果看起来像是通过某种噪声函数运行置换.您可以使用 WebGL Inspector 或 着色器编辑器 查看着色器内部，看看它们在做什么.
Getting the exact effect of your second example looks like it's running the displacement through some kind of noise function. You could use something like the WebGL Inspector or the Shader Editor to look inside the shaders and see what they're doing.
这是另一个示例，它创建了一个向中心位移而不是向边缘位移的位移纹理.
Here's another example that creates a displacement texture that displaces more toward the center than the edge.
注意:我应该明确表示，我没有查看您链接到的示例如何工作的详细信息，我只是建议他们正在做与此类似的事情.找出他们真正在做什么的最好方法是查看他们的代码并运行前面段落中提到的工具来查看内部并查看发生了什么.也许他们没有使用直接置换，而是使用法线之类的东西作为置换.也许不是绘制纯色(第二个和第三个示例)或纹理(第四个示例)，而是使用程序生成的图案或使用基于屏幕的纹理坐标来绘制重复的纹理图案.也许置换纹理是一个单独的纹理，并且它们有一个混合蒙版"，它们以白色绘制并渐变为黑色，以决定要应用多少置换纹理.在 WebGL 中有无数种方法可以做事.
NOTE: I should make it clear I didn't look at the details of how the examples you linked to worked, I'm only suggesting they are doing something similar to this. The best way to find out what they're really doing is to look at their code and run the tools mentioned in the previous paragraphs to look inside and see what's going on. Maybe they aren't using direct displacement but instead using something like normals as displacements. Maybe instead of drawing a solid color (the 2nd and 3rd examples) or a texture (the 4th example), they're drawing with a procedurally generated pattern or using screen based texture coordinates for a repeating texture pattern. Maybe the displacement texture is a separate texture and they have a "mix mask" that they draw in white and fade to black to decide how much of the displacement texture to apply. There is an infinite number of ways to do things in WebGL.
这篇关于在鼠标移动上创建涂抹/液化效果，使用 webgl 连续动画回原始状态的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
