Menu
I am writing a GPU-based real-time raytracing renderer using a GLSL compute shader. So far, it works really well, but I have stumbled into a seemingly unsolvable problem when it comes to having both reflections and refractions simultaneously.
My logic tells me that in order to have reflections and refractions on an object, such as glass, the ray would have to split into two, one ray reflects off the surface, and the other refracts through the surface. The ultimate colours of these rays would then be combined based on some function and ultimately used as the colour of the pixel the ray originated from. The problem I have is that I can't split the rays in shader code, as I would have to use recursion to do so. From my understanding, functions in a shader cannot be recursive because all GLSL functions are like inline functions in C++ due to compatibility issues with older GPU hardware.
Is it possible to simulate or fake recursion in shader code, or can I even achieve reflection and refraction simultaneously without using recursion at all? I can't see how it can happen without recursion, but I might be wrong.
207
I manage to convert back-raytracing to iterative process suitable for GLSL with the method suggested in my comment. It is far from optimized and I do not have all the physical stuff implemented (no Snell's law etc ...) yet but as a proof of concept it works already. I do all the stuff in fragment shader and CPU side code just send the uniforms constants and scene in form of 32 bit non-clamped float texture GL_LUMINANCE32F_ARB The rendering is just single QUAD covering whole screen.
I decided to store the scene in texture so each ray/fragment has direct access to whole scene. The texture is 2D but it is used as linear list of 32 bit floats. I decided this format:
enum _fac_type_enum
{
_fac_triangles=0, // r,g,b,a, n, triangle count, { x0,y0,z0,x1,y1,z1,x2,y2,z2 }
_fac_spheres, // r,g,b,a, n, sphere count, { x,y,z,r }
};
const GLfloat _n_glass=1.561;
const GLfloat _n_vacuum=1.0;
GLfloat data[]=
{
// r, g, b, a, n, type,count
0.2,0.3,0.5,0.5,_n_glass,_fac_triangles, 4, // tetrahedron
// px, py, pz, r, g, b
-0.5,-0.5,+1.0,
0.0,+0.5,+1.0,
+0.5,-0.5,+1.0,
0.0, 0.0,+0.5,
-0.5,-0.5,+1.0,
0.0,+0.5,+1.0,
0.0, 0.0,+0.5,
0.0,+0.5,+1.0,
+0.5,-0.5,+1.0,
0.0, 0.0,+0.5,
+0.5,-0.5,+1.0,
-0.5,-0.5,+1.0,
};
You can add/change any type of object. This example holds just single semi transparent bluish tetrahedron. You could also add transform matrices more coefficients for material properties etc ...
the Vertex shader just initialize corner Rays of the view (start position and direction) which is interpolated so each fragment represents start ray of back ray tracing process.
Iterative back ray tracing
So I created a "static" list of rays and init it with the start ray. The Iteration is done in two steps first the back ray tracing:
store the position, surface normal and material properties into ray struct
also store their indexes to the processed ray struct
Now your rays should hold all the intersection info you need to reconstruct the color. To do that:
so use lighting equations you want. If the ray contains children add their color to the result based on material properties (reflective and refractive coefficients ...)
Now the first ray should contain the color you want to output.
Uniforms used:
tm_eyeview camera matrix
aspectview ys/xs aspect ratio
n0 empty space refraction index (unused yet)
focal_length camera focal length
fac_siz resolution of the scene square texture
fac_num number of floats actually used in the scene texture
fac_txr texture unit for the scene texture
Preview:
The fragment shader contains my debug prints so you will need also the texture if used see the QA:
ToDo:
add matrices for objects, camera etc.
add material properties (shininess, reflection/refraction coefficient)
Snell's law right now the direction of new rays are wrong ...
may be separate R,G,B to 3 start rays and combine at the end
fake SSS Subsurface scattering based on ray lengths
better implement lights (right now they are constants in a code)
implement more primitives (right now only triangles are supported)
[Edit1] code debug and upgrade
I removed old source code to fit inside 30KB limit. If you need it then dig it from edit history. Had some time for more advanced debugging for this and here the result:
this version got resolved some geometrical,accuracy,domain problems and bugs. I got implemented both reflections and refractions as is shown on this debug draw for test ray:
In the debug view only the cube is transparent and last ray that does not hit anything is ignored. So as you can see the ray split ... The ray ended inside cube due to total reflection angle And I disable all reflections inside objects for speed reasons.
The 32bit floats for intersection detection are a bit noisy with distances so you can use 64bit doubles instead but the speed drops considerably in such case. Another option is to rewrite the equation to use relative coordinates which are more precise in this case of use.
Here the float shaders source:
Vertex:
//------------------------------------------------------------------
#version 420 core
//------------------------------------------------------------------
uniform float aspect;
uniform float focal_length;
uniform mat4x4 tm_eye;
layout(location=0) in vec2 pos;
out smooth vec2 txt_pos; // frag position on screen <-1,+1> for debug prints
out smooth vec3 ray_pos; // ray start position
out smooth vec3 ray_dir; // ray start direction
//------------------------------------------------------------------
void main(void)
{
vec4 p;
txt_pos=pos;
// perspective projection
p=tm_eye*vec4(pos.x/aspect,pos.y,0.0,1.0);
ray_pos=p.xyz;
p-=tm_eye*vec4(0.0,0.0,-focal_length,1.0);
ray_dir=normalize(p.xyz);
gl_Position=vec4(pos,0.0,1.0);
}
//------------------------------------------------------------------
Fragment:
//------------------------------------------------------------------
#version 420 core
//------------------------------------------------------------------
// Ray tracer ver: 1.000
//------------------------------------------------------------------
in smooth vec3 ray_pos; // ray start position
in smooth vec3 ray_dir; // ray start direction
uniform float n0; // refractive index of camera origin
uniform int fac_siz; // square texture x,y resolution size
uniform int fac_num; // number of valid floats in texture
uniform sampler2D fac_txr; // scene mesh data texture
out layout(location=0) vec4 frag_col;
//---------------------------------------------------------------------------
//#define _debug_print
#define _reflect
#define _refract
//---------------------------------------------------------------------------
#ifdef _debug_print
in vec2 txt_pos; // frag screen position <-1,+1>
uniform sampler2D txr_font; // ASCII 32x8 characters font texture unit
uniform float txt_fxs,txt_fys; // font/screen resolution ratio
const int _txtsiz=64; // text buffer size
int txt[_txtsiz],txtsiz; // text buffer and its actual size
vec4 txt_col=vec4(0.0,0.0,0.0,1.0); // color interface for txt_print()
bool _txt_col=false; // is txt_col active?
void txt_decimal(vec2 v); // print vec3 into txt
void txt_decimal(vec3 v); // print vec3 into txt
void txt_decimal(vec4 v); // print vec3 into txt
void txt_decimal(float x); // print float x into txt
void txt_decimal(int x); // print int x into txt
void txt_print(float x0,float y0); // print txt at x0,y0 [chars]
#endif
//---------------------------------------------------------------------------
void main(void)
{
const vec3 light_dir=normalize(vec3(0.1,0.1,1.0));
const float light_iamb=0.1; // dot offset
const float light_idir=0.5; // directional light amplitude
const vec3 back_col=vec3(0.2,0.2,0.2); // background color
const float _zero=1e-6; // to avoid intrsection with start point of ray
const int _fac_triangles=0; // r,g,b, refl,refr,n, type, triangle count, { x0,y0,z0,x1,y1,z1,x2,y2,z2 }
const int _fac_spheres =1; // r,g,b, refl,refr,n, type, sphere count, { x,y,z,r }
// ray scene intersection
struct _ray
{
vec3 pos,dir,nor;
vec3 col;
float refl,refr;// reflection,refraction intensity coeficients
float n0,n1,l; // refaction index (start,end) , ray length
int lvl,i0,i1; // recursion level, reflect, refract
};
const int _lvls=5;
const int _rays=(1<<_lvls)-1;
_ray ray[_rays]; int rays;
vec3 v0,v1,v2,pos;
vec3 c,col;
float refr,refl;
float tt,t,n1,a;
int i0,ii,num,id;
// fac texture access
vec2 st; int i,j; float ds=1.0/float(fac_siz-1);
#define fac_get texture(fac_txr,st).r; st.s+=ds; i++; j++; if (j==fac_siz) { j=0; st.s=0.0; st.t+=ds; }
// enque start ray
ray[0].pos=ray_pos;
ray[0].dir=normalize(ray_dir);
ray[0].nor=vec3(0.0,0.0,0.0);
ray[0].refl=0.0;
ray[0].refr=0.0;
ray[0].n0=n0;
ray[0].n1=1.0;
ray[0].l =0.0;
ray[0].lvl=0;
ray[0].i0=-1;
ray[0].i1=-1;
rays=1;
// debug print area
#ifdef _debug_print
bool _dbg=false;
float dbg_x0=45.0;
float dbg_y0= 1.0;
float dbg_xs=12.0;
float dbg_ys=_rays+1.0;
dbg_xs=40.0;
dbg_ys=10;
float x=0.5*(1.0+txt_pos.x)/txt_fxs; x-=dbg_x0;
float y=0.5*(1.0-txt_pos.y)/txt_fys; y-=dbg_y0;
// inside bbox?
if ((x>=0.0)&&(x<=dbg_xs)
&&(y>=0.0)&&(y<=dbg_ys))
{
// prints on
_dbg=true;
// preset debug ray
ray[0].pos=vec3(0.0,0.0,0.0)*2.5;
ray[0].dir=vec3(0.0,0.0,1.0);
}
#endif
// loop all enqued rays
for (i0=0;i0<rays;i0++)
{
// loop through all objects
// find closest forward intersection between them and ray[i0]
// strore it to ray[i0].(nor,col)
// strore it to pos,n1
t=tt=-1.0; ii=1; ray[i0].l=0.0;
ray[i0].col=back_col;
pos=ray[i0].pos; n1=n0;
for (st=vec2(0.0,0.0),i=j=0;i<fac_num;)
{
c.r=fac_get; // RGBA
c.g=fac_get;
c.b=fac_get;
refl=fac_get;
refr=fac_get;
n1=fac_get; // refraction index
a=fac_get; id=int(a); // object type
a=fac_get; num=int(a); // face count
if (id==_fac_triangles)
for (;num>0;num--)
{
v0.x=fac_get; v0.y=fac_get; v0.z=fac_get;
v1.x=fac_get; v1.y=fac_get; v1.z=fac_get;
v2.x=fac_get; v2.y=fac_get; v2.z=fac_get;
vec3 e1,e2,n,p,q,r;
float t,u,v,det,idet;
//compute ray triangle intersection
e1=v1-v0;
e2=v2-v0;
// Calculate planes normal vector
p=cross(ray[i0].dir,e2);
det=dot(e1,p);
// Ray is parallel to plane
if (abs(det)<1e-8) continue;
idet=1.0/det;
r=ray[i0].pos-v0;
u=dot(r,p)*idet;
if ((u<0.0)||(u>1.0)) continue;
q=cross(r,e1);
v=dot(ray[i0].dir,q)*idet;
if ((v<0.0)||(u+v>1.0)) continue;
t=dot(e2,q)*idet;
if ((t>_zero)&&((t<=tt)||(ii!=0)))
{
ii=0; tt=t;
// store color,n ...
ray[i0].col=c;
ray[i0].refl=refl;
ray[i0].refr=refr;
// barycentric interpolate position
t=1.0-u-v;
pos=(v0*t)+(v1*u)+(v2*v);
// compute normal (store as dir for now)
e1=v1-v0;
e2=v2-v1;
ray[i0].nor=cross(e1,e2);
}
}
if (id==_fac_spheres)
for (;num>0;num--)
{
float r;
v0.x=fac_get; v0.y=fac_get; v0.z=fac_get; r=fac_get;
// compute l0 length of ray(p0,dp) to intersection with sphere(v0,r)
// where rr= r^-2
float aa,bb,cc,dd,l0,l1,rr;
vec3 p0,dp;
p0=ray[i0].pos-v0; // set sphere center to (0,0,0)
dp=ray[i0].dir;
rr = 1.0/(r*r);
aa=2.0*rr*dot(dp,dp);
bb=2.0*rr*dot(p0,dp);
cc= rr*dot(p0,p0)-1.0;
dd=((bb*bb)-(2.0*aa*cc));
if (dd<0.0) continue;
dd=sqrt(dd);
l0=(-bb+dd)/aa;
l1=(-bb-dd)/aa;
if (l0<0.0) l0=l1;
if (l1<0.0) l1=l0;
t=min(l0,l1); if (t<=_zero) t=max(l0,l1);
if ((t>_zero)&&((t<=tt)||(ii!=0)))
{
ii=0; tt=t;
// store color,n ...
ray[i0].col=c;
ray[i0].refl=refl;
ray[i0].refr=refr;
// position,normal
pos=ray[i0].pos+(ray[i0].dir*t);
ray[i0].nor=pos-v0;
}
}
}
ray[i0].l=tt;
ray[i0].nor=normalize(ray[i0].nor);
// split ray from pos and ray[i0].nor
if ((ii==0)&&(ray[i0].lvl<_lvls-1))
{
t=dot(ray[i0].dir,ray[i0].nor);
// reflect
#ifdef _reflect
if ((ray[i0].refl>_zero)&&(t<_zero)) // do not reflect inside objects
{
ray[i0].i0=rays;
ray[rays]=ray[i0];
ray[rays].lvl++;
ray[rays].i0=-1;
ray[rays].i1=-1;
ray[rays].pos=pos;
ray[rays].dir=ray[rays].dir-(2.0*t*ray[rays].nor);
ray[rays].n0=ray[i0].n0;
ray[rays].n1=ray[i0].n0;
rays++;
}
#endif
// refract
#ifdef _refract
if (ray[i0].refr>_zero)
{
ray[i0].i1=rays;
ray[rays]=ray[i0];
ray[rays].lvl++;
ray[rays].i0=-1;
ray[rays].i1=-1;
ray[rays].pos=pos;
t=dot(ray[i0].dir,ray[i0].nor);
if (t>0.0) // exit object
{
ray[rays].n0=ray[i0].n0;
ray[rays].n1=n0;
v0=-ray[i0].nor; t=-t;
}
else{ // enter object
ray[rays].n0=n1;
ray[rays].n1=ray[i0].n0;
ray[i0 ].n1=n1;
v0=ray[i0].nor;
}
n1=ray[i0].n0/ray[i0].n1;
tt=1.0-(n1*n1*(1.0-t*t));
if (tt>=0.0)
{
ray[rays].dir=(ray[i0].dir*n1)-(v0*((n1*t)+sqrt(tt)));
rays++;
}
}
#endif
}
else if (i0>0) // ignore last ray if nothing hit
{
ray[i0]=ray[rays-1];
rays--; i0--;
}
}
// back track ray intersections and compute output color col
// lvl is sorted ascending so backtrack from end
for (i0=rays-1;i0>=0;i0--)
{
// directional + ambient light
t=abs(dot(ray[i0].nor,light_dir)*light_idir)+light_iamb;
t*=1.0-ray[i0].refl-ray[i0].refr;
ray[i0].col.rgb*=t;
// reflect
ii=ray[i0].i0;
if (ii>=0) ray[i0].col.rgb+=ray[ii].col.rgb*ray[i0].refl;
// refract
ii=ray[i0].i1;
if (ii>=0) ray[i0].col.rgb+=ray[ii].col.rgb*ray[i0].refr;
}
col=ray[0].col;
// debug prints
#ifdef _debug_print
/*
if (_dbg)
{
txtsiz=0;
txt_decimal(_lvls);
txt[txtsiz]=' '; txtsiz++;
txt_decimal(rays);
txt[txtsiz]=' '; txtsiz++;
txt_decimal(_rays);
txt_print(dbg_x0,dbg_y0);
for (ii=0;ii<rays;ii++)
{
txtsiz=0;
txt_decimal(ray[ii].lvl);
txt_print(dbg_x0,dbg_y0+ii+1);
}
for (ii=0,st=vec2(0.0,0.0),i=j=0;i<fac_num;ii++)
{
c.r=fac_get; // RGBA
txtsiz=0;
txt_decimal(c.r);
txt_print(dbg_x0,dbg_y0+ii+1);
}
if (_txt_col) col=txt_col.rgb;
}
*/
if (_dbg)
{
float x=dbg_x0,y=dbg_y0;
vec3 a=vec3(1.0,2.0,3.0);
vec3 b=vec3(5.0,6.0,7.0);
txtsiz=0; txt_decimal(dot(a,b)); txt_print(x,y); y++;
txtsiz=0; txt_decimal(cross(a,b)); txt_print(x,y); y++;
if (_txt_col) col=txt_col.rgb;
}
#endif
frag_col=vec4(col,1.0);
}
//---------------------------------------------------------------------------
#ifdef _debug_print
//---------------------------------------------------------------------------
void txt_decimal(vec2 v) // print vec2 into txt
{
txt[txtsiz]='('; txtsiz++;
txt_decimal(v.x); txt[txtsiz]=','; txtsiz++;
txt_decimal(v.y); txt[txtsiz]=')'; txtsiz++;
txt[txtsiz]=0; // string terminator
}
//---------------------------------------------------------------------------
void txt_decimal(vec3 v) // print vec3 into txt
{
txt[txtsiz]='('; txtsiz++;
txt_decimal(v.x); txt[txtsiz]=','; txtsiz++;
txt_decimal(v.y); txt[txtsiz]=','; txtsiz++;
txt_decimal(v.z); txt[txtsiz]=')'; txtsiz++;
txt[txtsiz]=0; // string terminator
}
//---------------------------------------------------------------------------
void txt_decimal(vec4 v) // print vec4 into txt
{
txt[txtsiz]='('; txtsiz++;
txt_decimal(v.x); txt[txtsiz]=','; txtsiz++;
txt_decimal(v.y); txt[txtsiz]=','; txtsiz++;
txt_decimal(v.z); txt[txtsiz]=','; txtsiz++;
txt_decimal(v.w); txt[txtsiz]=')'; txtsiz++;
txt[txtsiz]=0; // string terminator
}
//---------------------------------------------------------------------------
void txt_decimal(float x) // print float x into txt
{
int i,j,c; // l is size of string
float y,a;
const float base=10;
// handle sign
if (x<0.0) { txt[txtsiz]='-'; txtsiz++; x=-x; }
else { txt[txtsiz]='+'; txtsiz++; }
// divide to int(x).fract(y) parts of number
y=x; x=floor(x); y-=x;
// handle integer part
i=txtsiz; // start of integer part
for (;txtsiz<_txtsiz;)
{
a=x;
x=floor(x/base);
a-=base*x;
txt[txtsiz]=int(a)+'0'; txtsiz++;
if (x<=0.0) break;
}
j=txtsiz-1; // end of integer part
for (;i<j;i++,j--) // reverse integer digits
{
c=txt[i]; txt[i]=txt[j]; txt[j]=c;
}
// handle fractional part
for (txt[txtsiz]='.',txtsiz++;txtsiz<_txtsiz;)
{
y*=base;
a=floor(y);
y-=a;
txt[txtsiz]=int(a)+'0'; txtsiz++;
if (y<=0.0) break;
}
txt[txtsiz]=0; // string terminator
}
//---------------------------------------------------------------------------
void txt_decimal(int x) // print int x into txt
{
int a,i,j,c; // l is size of string
const int base=10;
// handle sign
if (x<0.0) { txt[txtsiz]='-'; txtsiz++; x=-x; }
else { txt[txtsiz]='+'; txtsiz++; }
// handle integer part
i=txtsiz; // start of integer part
for (;txtsiz<_txtsiz;)
{
a=x;
x/=base;
a-=base*x;
txt[txtsiz]=int(a)+'0'; txtsiz++;
if (x<=0) break;
}
j=txtsiz-1; // end of integer part
for (;i<j;i++,j--) // reverse integer digits
{
c=txt[i]; txt[i]=txt[j]; txt[j]=c;
}
txt[txtsiz]=0; // string terminator
}
//---------------------------------------------------------------------------
void txt_print(float x0,float y0) // print txt at x0,y0 [chars]
{
int i;
float x,y;
// fragment position [chars] relative to x0,y0
x=0.5*(1.0+txt_pos.x)/txt_fxs; x-=x0;
y=0.5*(1.0-txt_pos.y)/txt_fys; y-=y0;
// inside bbox?
if ((x<0.0)||(x>float(txtsiz))||(y<0.0)||(y>1.0)) return;
// get font texture position for target ASCII
i=int(x); // char index in txt
x-=float(i);
i=txt[i];
x+=float(int(i&31));
y+=float(int(i>>5));
x/=32.0; y/=8.0; // offset in char texture
txt_col=texture(txr_font,vec2(x,y));
_txt_col=true;
}
//---------------------------------------------------------------------------
#endif
//---------------------------------------------------------------------------
The code is not optimized yet I wanted to have the physics working correctly first. There are still not Fresnells implemented but refl,refr coefficients of material are used instead.
Also you can ignore the debug prints stuff (they are encapsulated by #define).
I build a small class for the geometry texture so I can easily set up scene objects. This is how the scene was initiated for the preview:
ray.beg();
// r g b rfl rfr n
ray.add_material(1.0,1.0,1.0,0.3,0.0,_n_glass); ray.add_box ( 0.0, 0.0, 6.0,9.0,9.0,0.1);
ray.add_material(1.0,1.0,1.0,0.1,0.8,_n_glass); ray.add_sphere( 0.0, 0.0, 0.5,0.5);
ray.add_material(1.0,0.1,0.1,0.3,0.0,_n_glass); ray.add_sphere( +2.0, 0.0, 2.0,0.5);
ray.add_material(0.1,1.0,0.1,0.3,0.0,_n_glass); ray.add_box ( -2.0, 0.0, 2.0,0.5,0.5,0.5);
ray.add_material(0.1,0.1,1.0,0.3,0.0,_n_glass);
ray.add_tetrahedron
(
0.0, 0.0, 3.0,
-1.0,-1.0, 4.0,
+1.0,-1.0, 4.0,
0.0,+1.0, 4.0
);
ray.end();
It is important so computed normals are facing out of objects because that is used for detecting inside/outside object crossings.
P.S.
If you're interested here is my volumetric 3D back ray tracer:
Here newer version of this "Mesh" Raytracer supporting hemisphere objects:
156
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
