我正在尝试使用Sketch.js来学习使用粒子,并且遇到了这个我想要重做的例子:
http://soulwire.github.io/Plasmatic-Isosurface/
有问题的代码片段就是这个热点:
GLSL = {
vert: "\n#ifdef GL_ES\nprecision mediump float;\n#endif\n\n// Uniforms\nuniform vec2 u_resolution;\n\n// Attributes\nattribute vec2 a_position;\n\nvoid main() {\n gl_Position = vec4 (a_position, 0, 1);\n}\n",
frag: "\n#ifdef GL_ES\nprecision mediump float;\n#endif\n\nuniform bool u_scanlines;\nuniform vec2 u_resolution;\n\nuniform float u_brightness;\nuniform float u_blobiness;\nuniform float u_particles;\nuniform float u_millis;\nuniform float u_energy;\n\n\nfloat noise( vec2 co ){\n return fract( sin( dot( co.xy, vec2( 12.9898, 78.233 ) ) ) * 43758.5453 );\n}\n\nvoid main( void ) {\n\n vec2 position = ( gl_FragCoord.xy / u_resolution.x );\n float t = u_millis * 0.001 * u_energy;\n \n float a = 0.0;\n float b = 0.0;\n float c = 0.0;\n\n vec2 pos, center = vec2( 0.5, 0.5 * (u_resolution.y / u_resolution.x) );\n \n float na, nb, nc, nd, d;\n float limit = u_particles / 40.0;\n float step = 1.0 / u_particles;\n float n = 0.0;\n \n for ( float i = 0.0; i <= 1.0; i += 0.025 ) {\n\n if ( i <= limit ) {\n\n vec2 np = vec2(n, 1-1);\n \n na = noise( np * 1.1 );\n nb = noise( np * 2.8 );\n nc = noise( np * 0.7 );\n nd = noise( np * 3.2 );\n\n pos = center;\n pos.x += sin(t*na) * cos(t*nb) * tan(t*na*0.15) * 0.3;\n pos.y += tan(t*nc) * sin(t*nd) * 0.1;\n \n d = pow( 1.6*na / length( pos - position ), u_blobiness );\n \n if ( i < limit * 0.3333 ) a += d;\n else if ( i < limit * 0.6666 ) b += d;\n else c += d;\n\n n += step;\n }\n }\n \n vec3 col = vec3(a*c,b*c,a*b) * 0.0001 * u_brightness;\n \n if ( u_scanlines ) {\n col -= mod( gl_FragCoord.y, 2.0 ) < 1.0 ? 0.5 : 0.0;\n }\n \n gl_FragColor = vec4( col, 1.0 );\n\n}\n"};
着色器是GLSL的一个奇怪的嵌入位,我无法弄清楚如何调试。我无法理解的这个具体问题是如何设置颜色。 col
变量似乎是一个四维向量,但看起来它可以用二维向量构建。我真的不明白发生了什么,这导致了我在这里问的问题,也就是说,有没有办法调试这样的嵌入式代码?
我尝试过一个名为WebGL Insight的Chrome扩展程序,但它不能识别GLSL代码,我不确定它是我正在寻找的内容。对于类似问题(Debug GLSL code in webgl)有这个答案,但我觉得它没有帮助。我是否需要打破着色器并使用外部编译器? (在这种情况下,我需要什么软件和库?)
答案 0 :(得分:2)
我不知道这是否有帮助,但在Firefox中,您可以看到网页中使用的所有着色器的代码:https://developer.mozilla.org/en-US/docs/Tools/Shader_Editor。
答案 1 :(得分:2)
嗯,简单地重新引入换行符会产生非常易读的代码:
#ifdef GL_ES
precision mediump float;
#endif
uniform bool u_scanlines;
uniform vec2 u_resolution;
uniform float u_brightness;
uniform float u_blobiness;
uniform float u_particles;
uniform float u_millis;
uniform float u_energy;
float noise( vec2 co ){
return fract( sin( dot( co.xy, vec2( 12.9898, 78.233 ) ) ) * 43758.5453 );
}
void main( void ) {
vec2 position = ( gl_FragCoord.xy / u_resolution.x );
float t = u_millis * 0.001 * u_energy;
float a = 0.0;
float b = 0.0;
float c = 0.0;
vec2 pos, center = vec2( 0.5, 0.5 * (u_resolution.y / u_resolution.x) );
float na, nb, nc, nd, d;
float limit = u_particles / 40.0;
float step = 1.0 / u_particles;
float n = 0.0;
for ( float i = 0.0; i <= 1.0; i += 0.025 ) {
if ( i <= limit ) {
vec2 np = vec2(n, 1-1);
na = noise( np * 1.1 );
nb = noise( np * 2.8 );
nc = noise( np * 0.7 );
nd = noise( np * 3.2 );
pos = center;
pos.x += sin(t*na) * cos(t*nb) * tan(t*na*0.15) * 0.3;
pos.y += tan(t*nc) * sin(t*nd) * 0.1;
d = pow( 1.6*na / length( pos - position ), u_blobiness );
if ( i < limit * 0.3333 ) a += d;
else if ( i < limit * 0.6666 ) b += d;
else c += d;
n += step;
}
}
vec3 col = vec3(a*c,b*c,a*b) * 0.0001 * u_brightness;
if ( u_scanlines ) {
col -= mod( gl_FragCoord.y, 2.0 ) < 1.0 ? 0.5 : 0.0;
}
gl_FragColor = vec4( col, 1.0 );
}
由于这似乎是一个屏幕空间着色器,因此可以使用glsl-sandbox或shadertoy之类的东西来玩它,只需重新映射和/或用未定义的制服替换定义:
//uniform bool u_scanlines;
#define u_scanlines false
//uniform float u_brightness;
#define u_brightness 1.
//uniform float u_blobiness;
#define u_blobiness 1.
//uniform float u_particles;
#define u_particles 20.
//uniform float u_millis;
#define u_millis 1.
//uniform float u_energy;
#define u_energy 1.
//uniform vec2 u_resolution;
uniform vec2 resolution;
#define u_resolution resolution
要调试“在app中”你唯一现成的解决方案是Firefoxs Shader-Editor由Tolokoban推荐。对于其他更复杂的工具,还有桌面着色器调试器,如NVIDIAs FX Composer,AMDs RenderMonkey和GLSL-Debugger(formerly GLSLDevil)。