mirror of
https://github.com/OrcaSlicer/OrcaSlicer.git
synced 2026-10-14 11:21:18 +00:00
SoftFever
0473da4ef8
The new Mixed colors setting caps how many mixed filaments a bake adds. They are picked from the texture's colours, and only the ones the bake paints with are created. Previewing no longer creates filament slots, both previews show a mix in its slot's colour, and the bake paints each mix with the slot it actually got.
436 lines
24 KiB
GLSL
436 lines
24 KiB
GLSL
#version 140
|
|
|
|
// Fast, geometry-free preview of texture displacement: perturbs the *shading* normal from the
|
|
// height texture's local gradient, faded out by the per-vertex paint weight. The
|
|
// true, exact result is what "Bake" produces via libslic3r/TextureDisplacement.cpp on the CPU.
|
|
//
|
|
// The bake displaces each surface point along its normal by H = +/- depth_mm * (h(uv) - midlevel),
|
|
// with uv from the layer's projection. The perturbed normal is the analytic
|
|
//
|
|
// N' = normalize(N - (dH/da) * T - (dH/db) * B)
|
|
//
|
|
// over any orthonormal surface tangent pair (T, B), where the two slopes are real mm-per-mm
|
|
// derivatives. Two things have to be right for the preview's apparent depth to match the bake's:
|
|
// the tangent frame the gradient is expressed in, and the uv->mm scale that turns a texel
|
|
// difference into a slope. Getting the scale wrong is a uniform flattening (a raw texel difference
|
|
// is dh over one texel step, not over one mm); getting the frame wrong tilts the relief along the
|
|
// wrong axes.
|
|
//
|
|
// Two projection paths:
|
|
// * Triplanar (use_vertex_uv = 0): uv and the tangent axes are derived in-shader from the dominant
|
|
// normal axis, mirroring libslic3r's project_planar()/apply_uv_transform(), and the slope is
|
|
// formed analytically (there is a closed-form uv, so 1 uv unit is exactly tiling_scale mm). This
|
|
// path also runs a parallax step before shading, see below.
|
|
//
|
|
// Parallax. Perturbing the normal alone shades only, so the pattern is welded to the base surface: it
|
|
// does not shift as the camera orbits and it does not get any deeper as depth_mm grows, which is
|
|
// exactly when the preview stops reading as real geometry. The triplanar path therefore shades at the
|
|
// point the *displaced* surface would show at this pixel rather than at the pixel's own base position.
|
|
//
|
|
// Two cheaper formulations were tried first and both are wrong here, which is worth recording:
|
|
// * Solving Q = P + V * (H(Q) / dot(V, n)) by fixed-point iteration. Geometrically exact, but the
|
|
// divisor goes to zero edge-on, and an unbounded step is not a small error - the sample lands a
|
|
// large fraction of a tile away and the iteration oscillates instead of converging. It reads as a
|
|
// *second, flat copy* of the pattern ghosted over the real one. Clamping the step to one tile does
|
|
// not help either: a tile-sized shift lands on the neighbouring tile, which is the same pattern.
|
|
// * Offset limiting (Welsh): step along the tangential part of V, whose length caps the shift at one
|
|
// depth. Stable and cheap, but it understates parallax by exactly the factor that matters - the
|
|
// relief still flattens as soon as the camera tilts, which is the complaint it was meant to fix.
|
|
//
|
|
// So this ray-marches instead (parallax occlusion mapping). A point at ray parameter s, i.e. P + V * s,
|
|
// sits at height s * dot(V, n) above the undisplaced surface. The displaced surface lives in a shell
|
|
// between the extreme values of amp * (h - midlevel); the march starts at the top of that shell, where
|
|
// the ray is outside the surface by construction, and steps inward until the ray height falls below the
|
|
// sampled height. That crossing *is* the visible point - no divergence, no ghosting, and parallax stays
|
|
// correct at any angle. The hit is interpolated between the last two samples, which is what keeps
|
|
// PARALLAX_STEPS low enough to afford. The march is skipped when sweeping the shell would move the
|
|
// sample point less than half a texel (the head-on case), so the common view pays almost nothing.
|
|
//
|
|
// The gradient/shading below is evaluated at the resulting uv, so the relief both slides correctly
|
|
// under camera motion and visibly deepens with depth_mm. What it still cannot do is change the
|
|
// model's silhouette or cast shadows; for that, switch the View row to Normal.
|
|
// * Precomputed uv (use_vertex_uv = 1, used for LSCM): uv comes per-vertex from the CPU (the LSCM
|
|
// unwrap with island placement + tiling/rotation/offset already folded in), and the perturbed
|
|
// normal is built with Mikkelsen's method -- the surface gradient taken straight from the
|
|
// screen-space derivatives of the sampled height and position. This makes no uv->mm scale
|
|
// assumption, which matters because an LSCM map is conformal, not isometric: the local mm-per-uv
|
|
// varies across the chart, so a single global 1/tiling factor (what an earlier version used) got
|
|
// the apparent depth wrong. This path is also what makes the fast preview follow the UV editor:
|
|
// move an island and its uv -- hence its relief -- moves with it.
|
|
|
|
#define INTENSITY_CORRECTION 0.6
|
|
|
|
#define PARALLAX_STEPS 24
|
|
// Explicit LOD: the march samples inside non-uniform control flow, where implicit
|
|
// derivatives are undefined.
|
|
#define H_AT(uv) textureLod(height_tex, uv, 0.0).r
|
|
|
|
// normalized values for (-0.6/1.31, 0.6/1.31, 1./1.31)
|
|
const vec3 LIGHT_TOP_DIR = vec3(-0.4574957, 0.4574957, 0.7624929);
|
|
#define LIGHT_TOP_DIFFUSE (0.8 * INTENSITY_CORRECTION)
|
|
#define LIGHT_TOP_SPECULAR (0.125 * INTENSITY_CORRECTION)
|
|
#define LIGHT_TOP_SHININESS 20.0
|
|
|
|
// normalized values for (1./1.43, 0.2/1.43, 1./1.43)
|
|
const vec3 LIGHT_FRONT_DIR = vec3(0.6985074, 0.1397015, 0.6985074);
|
|
#define LIGHT_FRONT_DIFFUSE (0.3 * INTENSITY_CORRECTION)
|
|
|
|
#define INTENSITY_AMBIENT 0.3
|
|
|
|
const vec3 ZERO = vec3(0.0, 0.0, 0.0);
|
|
|
|
uniform vec4 uniform_color;
|
|
// The printable palette, in **CIELAB** as well as RGB, and how many entries are real. Lab because the
|
|
// match has to be perceptual - the same reason the CPU side uses CIEDE2000 - and converting the
|
|
// palette once on the CPU is what lets the fragment shader match with a plain squared distance.
|
|
// Count 0 means nothing is colouring, and every fragment falls back to uniform_color as before.
|
|
uniform vec3 palette_lab[64];
|
|
uniform vec3 palette_rgb[64];
|
|
uniform int palette_count;
|
|
uniform bool pure_only; // match against single filaments only (flat-colour image)
|
|
// The entry's two filaments, equal for a single filament - only so a mix can be told apart. An entry's
|
|
// palette_rgb is already the colour it prints in (for a mix, its mixed filament slot's).
|
|
uniform int palette_a[64];
|
|
uniform int palette_b[64];
|
|
uniform float prefer_pure_de; // PREFER_PURE_DE: how much better than a single filament a mix must be
|
|
uniform sampler2D color_tex; // the layer's colour image, sampled at the same uv as the height
|
|
uniform bool has_color_tex;
|
|
uniform bool volume_mirrored;
|
|
|
|
uniform mat4 view_model_matrix;
|
|
uniform mat3 view_normal_matrix;
|
|
|
|
uniform sampler2D height_tex;
|
|
uniform vec2 height_tex_texel; // (1/width, 1/height) of height_tex
|
|
uniform float depth_mm;
|
|
uniform float tiling_scale;
|
|
// Height map width / height. Scales the v axis so a non-square image keeps its proportions
|
|
// instead of being squeezed into a square tile - mirrors libslic3r's apply_uv_transform().
|
|
uniform float tex_aspect;
|
|
uniform float rotation_rad;
|
|
uniform vec2 uv_offset;
|
|
uniform bool invert;
|
|
uniform float midlevel; // the height that means "don't move"; needed by the parallax step
|
|
uniform vec3 eye_model_pos; // camera position in the texture frame (world minus tex_anchor)
|
|
uniform vec3 tex_anchor; // the volume's origin in world space: the texture frame's origin
|
|
uniform bool use_vertex_uv; // true: sample at vertex_uv with a derived tangent frame (LSCM)
|
|
// A 2x3 affine (columns packed as lin = (m00, m01, m10, m11), tr = (m02, m12)) applied to the uv of
|
|
// the island currently being dragged in the UV editor (island_active > 0.5). Identity when nothing is
|
|
// dragged, so this whole path is a no-op then. Lets a UV island drag move the relief on the model with
|
|
// only a uniform update
|
|
uniform vec4 island_delta_lin;
|
|
uniform vec2 island_delta_tr;
|
|
// Which projection to reconstruct in-shader: 0 Triplanar, 1 Cylindrical, 2 Spherical (the low three
|
|
// TextureProjectionMethod values; LSCM and ViewProjected arrive through use_vertex_uv and leave this
|
|
// at 0). The two wrapping projections wrap around the *whole painted patch*, so its centroid - and,
|
|
// for Cylindrical, its axis - are properties no single fragment can derive. They come from the CPU,
|
|
// in this same texture frame, computed with the bake's own texture_displacement_patch_frame().
|
|
uniform int projection_mode;
|
|
uniform vec3 patch_center;
|
|
uniform vec3 patch_axis;
|
|
|
|
in vec3 clipping_planes_dots;
|
|
in vec4 model_pos;
|
|
in vec4 world_pos;
|
|
in float weight;
|
|
in float island_active;
|
|
in vec2 vertex_uv;
|
|
|
|
out vec4 out_color;
|
|
|
|
// The cylinder's own frame, built exactly as libslic3r's project_cylindrical() builds it - including
|
|
// the handedness, which comes out left-handed for an axis of +Z. Copied rather than "corrected", so
|
|
// the preview wraps the texture the same way round as the bake.
|
|
void cylinder_frame(out vec3 up, out vec3 right, out vec3 fwd)
|
|
{
|
|
up = (length(patch_axis) > 1e-8) ? normalize(patch_axis) : vec3(0.0, 0.0, 1.0);
|
|
vec3 arbitrary = (abs(up.z) < 0.9) ? vec3(0.0, 0.0, 1.0) : vec3(1.0, 0.0, 0.0);
|
|
right = normalize(cross(up, arbitrary));
|
|
fwd = normalize(cross(right, up));
|
|
}
|
|
|
|
// The raw, millimetre-valued projection of `p` (a position in the texture frame), before the layer's
|
|
// tiling/rotation/aspect/offset - a term-for-term transcription of libslic3r's project_planar(),
|
|
// project_cylindrical() and project_spherical(). `n` is read by the planar mode only.
|
|
vec2 projection_raw(vec3 p, vec3 n)
|
|
{
|
|
if (projection_mode == 1) { // Cylindrical: (arc length around, distance along)
|
|
vec3 up, right, fwd;
|
|
cylinder_frame(up, right, fwd);
|
|
vec3 rel = p - patch_center;
|
|
float x = dot(rel, right);
|
|
float y = dot(rel, fwd);
|
|
return vec2(atan(y, x) * sqrt(x * x + y * y), dot(rel, up));
|
|
}
|
|
if (projection_mode == 2) { // Spherical: (longitude, latitude) * radius
|
|
vec3 rel = p - patch_center;
|
|
float radius = length(rel);
|
|
if (radius < 1e-8)
|
|
return vec2(0.0);
|
|
vec3 dir = rel / radius;
|
|
return vec2(atan(dir.y, dir.x), asin(clamp(dir.z, -1.0, 1.0))) * radius;
|
|
}
|
|
vec3 an = abs(n); // Triplanar: drop the dominant normal axis
|
|
return (an.x >= an.y && an.x >= an.z) ? p.yz : ((an.y >= an.x && an.y >= an.z) ? p.xz : p.xy);
|
|
}
|
|
|
|
// The two surface directions projection_raw()'s u and v run along at `p`, plus how many raw units one
|
|
// millimetre of travel along each of them covers - the factor that turns the uv-space height gradient
|
|
// into a real mm-per-mm slope. For the planar projection both axes are world axes and the factor is 1.
|
|
// Cylindrical and Spherical are arc-length parametrized, so it is 1 there too, except for the
|
|
// spherical longitude, whose circle shrinks by cos(latitude) toward the poles. Exact where the surface
|
|
// really is the cylinder/sphere the projection assumes - the same assumption libslic3r makes.
|
|
void projection_axes(vec3 p, vec3 n, out vec3 t, out vec3 b, out vec2 units_per_mm)
|
|
{
|
|
units_per_mm = vec2(1.0, 1.0);
|
|
if (projection_mode == 1) {
|
|
vec3 up, right, fwd;
|
|
cylinder_frame(up, right, fwd);
|
|
vec3 rel = p - patch_center;
|
|
vec2 xy = vec2(dot(rel, right), dot(rel, fwd));
|
|
float r = length(xy);
|
|
t = (r > 1e-6) ? (fwd * xy.x - right * xy.y) / r : right; // circumferential: u runs along it
|
|
b = up; // v is the distance along the axis
|
|
return;
|
|
}
|
|
if (projection_mode == 2) {
|
|
vec3 rel = p - patch_center;
|
|
float r = length(rel);
|
|
vec3 dir = (r > 1e-8) ? rel / r : vec3(0.0, 0.0, 1.0);
|
|
float c = length(dir.xy); // cos(latitude)
|
|
t = (c > 1e-6) ? vec3(-dir.y, dir.x, 0.0) / c : vec3(1.0, 0.0, 0.0);
|
|
b = cross(dir, t); // increasing latitude, unit length
|
|
units_per_mm = vec2(1.0 / max(c, 1e-3), 1.0);
|
|
return;
|
|
}
|
|
vec3 an = abs(n);
|
|
if (an.x >= an.y && an.x >= an.z) { // planar = p.yz
|
|
t = vec3(0.0, 1.0, 0.0);
|
|
b = vec3(0.0, 0.0, 1.0);
|
|
} else if (an.y >= an.x && an.y >= an.z) { // planar = p.xz
|
|
t = vec3(1.0, 0.0, 0.0);
|
|
b = vec3(0.0, 0.0, 1.0);
|
|
} else { // planar = p.xy
|
|
t = vec3(1.0, 0.0, 0.0);
|
|
b = vec3(0.0, 1.0, 0.0);
|
|
}
|
|
}
|
|
|
|
vec2 project_uv(vec3 p, vec3 n)
|
|
{
|
|
vec2 planar = projection_raw(p, n);
|
|
planar *= (tiling_scale > 1e-6) ? (1.0 / tiling_scale) : 1.0;
|
|
float cs = cos(rotation_rad);
|
|
float sn = sin(rotation_rad);
|
|
vec2 r = vec2(planar.x * cs - planar.y * sn, planar.x * sn + planar.y * cs);
|
|
// After the rotation, so the rotation stays a rotation rather than becoming a shear.
|
|
r.y *= tex_aspect;
|
|
return r + uv_offset;
|
|
}
|
|
|
|
// sRGB -> CIELAB, matching slic3r/Utils/ColorSpaceConvert's RGB2Lab so this picks the same entry the
|
|
// bake does.
|
|
vec3 srgb_to_lab(vec3 c)
|
|
{
|
|
vec3 v = vec3(c.r > 0.04045 ? pow((c.r + 0.055) / 1.055, 2.4) : c.r / 12.92,
|
|
c.g > 0.04045 ? pow((c.g + 0.055) / 1.055, 2.4) : c.g / 12.92,
|
|
c.b > 0.04045 ? pow((c.b + 0.055) / 1.055, 2.4) : c.b / 12.92);
|
|
vec3 xyz = vec3(dot(v, vec3(0.4124, 0.3576, 0.1805)) / 0.95047,
|
|
dot(v, vec3(0.2126, 0.7152, 0.0722)),
|
|
dot(v, vec3(0.0193, 0.1192, 0.9505)) / 1.08883);
|
|
vec3 f = vec3(xyz.x > 0.008856 ? pow(xyz.x, 1.0 / 3.0) : (7.787 * xyz.x) + 16.0 / 116.0,
|
|
xyz.y > 0.008856 ? pow(xyz.y, 1.0 / 3.0) : (7.787 * xyz.y) + 16.0 / 116.0,
|
|
xyz.z > 0.008856 ? pow(xyz.z, 1.0 / 3.0) : (7.787 * xyz.z) + 16.0 / 116.0);
|
|
return vec3(116.0 * f.y - 16.0, 500.0 * (f.x - f.y), 200.0 * (f.y - f.z));
|
|
}
|
|
|
|
// Nearest printable colour to a sampled one. Quantizing per *fragment* rather than per facet is the
|
|
// whole point of this path: it shows the image at the texture's resolution instead of the mesh's,
|
|
// which is what you need while choosing a texture and placing it. The Normal view is where the
|
|
// facet-resolution truth - what actually bakes - is shown.
|
|
//
|
|
// Squared distance in Lab (CIE76) rather than the CPU's CIEDE2000: the two agree except on near-ties,
|
|
// and CIEDE2000 per fragment across 64 entries is not worth its cost in a preview.
|
|
int nearest_palette_entry(vec3 rgb)
|
|
{
|
|
vec3 lab = srgb_to_lab(rgb);
|
|
int best = 0;
|
|
int best_pure = -1;
|
|
float bd = 1.0e20;
|
|
float bd_pure = 1.0e20;
|
|
for (int i = 0; i < 64; ++i) {
|
|
if (i >= palette_count)
|
|
break;
|
|
if (pure_only && palette_a[i] != palette_b[i])
|
|
continue; // a flat-colour image never takes a mix (see the bake)
|
|
vec3 d = lab - palette_lab[i];
|
|
float d2 = dot(d, d);
|
|
if (palette_a[i] == palette_b[i] && d2 < bd_pure) {
|
|
bd_pure = d2;
|
|
best_pure = i;
|
|
}
|
|
if (d2 < bd) {
|
|
bd = d2;
|
|
best = i;
|
|
}
|
|
}
|
|
// The same bias make_palette_quantizer() applies: a mix is an interleave, so it is only worth taking
|
|
// when it beats the nearest single filament by a visible step - otherwise the preview shows mixes
|
|
// where the bake prints a single filament. Compared on the distances rather than their squares, so
|
|
// the margin means the same thing as it does on the CPU (up to CIE76 against CIEDE2000, the
|
|
// approximation already noted above).
|
|
if (best_pure >= 0 && palette_a[best] != palette_b[best] && sqrt(bd_pure) - sqrt(bd) < prefer_pure_de)
|
|
best = best_pure;
|
|
return best;
|
|
}
|
|
|
|
void main()
|
|
{
|
|
if (any(lessThan(clipping_planes_dots, ZERO)))
|
|
discard;
|
|
|
|
// Everything below runs in world millimetres, like the bake: a tile is tiling_scale mm on the
|
|
// printed part whatever the instance's scale or rotation, so the preview has to project from the
|
|
// world position and perturb the world normal.
|
|
vec3 triangle_normal = normalize(cross(dFdx(world_pos.xyz), dFdy(world_pos.xyz)));
|
|
vec3 tex_pos = world_pos.xyz - tex_anchor; // the frame the texture is projected in, as the bake does
|
|
if (volume_mirrored)
|
|
triangle_normal = -triangle_normal;
|
|
|
|
// Where the colour is read from. Both branches below already compute the uv this fragment's
|
|
// *height* came from - including the parallax-marched one on the triplanar path - and the colour
|
|
// has to follow it exactly, or the colour would slide off the relief as the camera orbits.
|
|
vec2 color_uv = vec2(0.0);
|
|
bool have_uv = false;
|
|
|
|
if (use_vertex_uv) {
|
|
// Precomputed-uv (LSCM) path - Mikkelsen's surface-gradient normal mapping ("Bump Mapping
|
|
// Unparametrized Surfaces on the GPU"). The perturbed normal is derived straight from the
|
|
// screen-space derivatives of the *sampled height* and the position, so it is scale-exact
|
|
// with no uv->mm assumption at all - which is the whole point here: an LSCM map is conformal,
|
|
// not isometric, so the local mm-per-uv varies across the chart and the earlier "one global
|
|
// 1/tiling factor" got the depth visibly wrong. dFdx(h) captures the true on-screen rate of
|
|
// change however the chart is stretched or however fine the tiling is.
|
|
//
|
|
// use_vertex_uv is a uniform, so this whole branch is uniform control flow and the texture
|
|
// derivatives are well defined; the paint weight gates the result by a plain multiply (k)
|
|
// rather than a per-fragment branch, keeping it that way.
|
|
// The dragged island's uv rides a uniform affine so its relief moves without a rebuild; every
|
|
// other vertex (island_active == 0) samples its baked uv unchanged.
|
|
vec2 uv = (island_active > 0.5)
|
|
? vec2(dot(island_delta_lin.xy, vertex_uv), dot(island_delta_lin.zw, vertex_uv)) + island_delta_tr
|
|
: vertex_uv;
|
|
color_uv = uv;
|
|
have_uv = true;
|
|
float h = texture(height_tex, uv).r;
|
|
float k = (invert ? -1.0 : 1.0) * depth_mm * clamp(weight, 0.0, 1.0);
|
|
vec3 sigmaS = dFdx(world_pos.xyz);
|
|
vec3 sigmaT = dFdy(world_pos.xyz);
|
|
vec3 R1 = cross(sigmaT, triangle_normal);
|
|
vec3 R2 = cross(triangle_normal, sigmaS);
|
|
float det = dot(sigmaS, R1);
|
|
float dHdx = k * dFdx(h);
|
|
float dHdy = k * dFdy(h);
|
|
if (abs(det) > 1e-12)
|
|
triangle_normal = normalize(triangle_normal - (dHdx * R1 + dHdy * R2) / det);
|
|
} else if (weight > 0.0) {
|
|
// Triplanar path: uv and the tangent axes are reconstructed in-shader from the dominant
|
|
// normal component (see header). The gradient is expressed analytically because there is a
|
|
// closed-form uv here, unlike the LSCM case.
|
|
vec3 t, b;
|
|
vec2 units_per_mm;
|
|
projection_axes(tex_pos, triangle_normal, t, b, units_per_mm);
|
|
|
|
// Parallax occlusion mapping: march the view ray through the height shell and shade at the
|
|
// first point where it drops below the displaced surface (see header).
|
|
float amp = (invert ? -1.0 : 1.0) * depth_mm * clamp(weight, 0.0, 1.0);
|
|
vec3 view_dir = normalize(eye_model_pos - tex_pos);
|
|
float v_dot_n = dot(view_dir, triangle_normal);
|
|
vec2 uv = project_uv(tex_pos, triangle_normal);
|
|
|
|
// The shell the displaced surface lives inside, as signed heights along the normal. Taken from
|
|
// both ends of h in [0, 1] so it stays correct for an inverted layer or a raised midlevel,
|
|
// where the surface sits *below* the undisplaced one.
|
|
float h_end_a = amp * (0.0 - midlevel);
|
|
float h_end_b = amp * (1.0 - midlevel);
|
|
float h_hi = max(h_end_a, h_end_b);
|
|
float h_lo = min(h_end_a, h_end_b);
|
|
// How far, in mm, sweeping the ray across the shell slides the sample point sideways. Below half
|
|
// a texel there is no parallax to find and the march would be pure cost - which is the common
|
|
// case of looking straight down at a surface.
|
|
float sweep = length(view_dir - triangle_normal * v_dot_n) * (h_hi - h_lo) / max(v_dot_n, 1e-4);
|
|
if (v_dot_n > 0.05 && sweep > 0.5 * tiling_scale * height_tex_texel.x) {
|
|
// A point at ray parameter s (model_pos + view_dir * s) sits at height s * v_dot_n above the
|
|
// undisplaced surface. Start at the top of the shell, where the ray is outside the surface
|
|
// by construction, and step inward; the crossing is what this pixel actually sees.
|
|
float s = h_hi / v_dot_n;
|
|
float ds = (h_hi - h_lo) / (v_dot_n * float(PARALLAX_STEPS));
|
|
vec2 prev_uv = project_uv(tex_pos + view_dir * s, triangle_normal);
|
|
float prev_gap = h_hi - amp * (H_AT(prev_uv) - midlevel); // >= 0 by construction
|
|
for (int i = 0; i < PARALLAX_STEPS; ++i) {
|
|
s -= ds;
|
|
vec2 cur_uv = project_uv(tex_pos + view_dir * s, triangle_normal);
|
|
float gap = s * v_dot_n - amp * (H_AT(cur_uv) - midlevel);
|
|
if (gap <= 0.0) {
|
|
// Crossed between the last two samples - interpolating the hit is what stops it
|
|
// quantising to the step size, and so what keeps the step count affordable.
|
|
uv = mix(prev_uv, cur_uv, clamp(prev_gap / max(prev_gap - gap, 1e-6), 0.0, 1.0));
|
|
break;
|
|
}
|
|
prev_uv = cur_uv;
|
|
prev_gap = gap;
|
|
}
|
|
}
|
|
|
|
color_uv = uv; // after the parallax march, so colour and relief stay registered
|
|
have_uv = true;
|
|
|
|
float hL = texture(height_tex, uv - vec2(height_tex_texel.x, 0.0)).r;
|
|
float hR = texture(height_tex, uv + vec2(height_tex_texel.x, 0.0)).r;
|
|
float hD = texture(height_tex, uv - vec2(0.0, height_tex_texel.y)).r;
|
|
float hU = texture(height_tex, uv + vec2(0.0, height_tex_texel.y)).r;
|
|
|
|
// Central difference, per uv unit (not per texel).
|
|
vec2 dh_duv = vec2((hR - hL) / (2.0 * height_tex_texel.x), (hU - hD) / (2.0 * height_tex_texel.y));
|
|
|
|
// uv -> mm is 1/tiling_scale for the triplanar projection, so this turns the uv-space
|
|
// gradient into a real surface slope.
|
|
float inv_tiling = (tiling_scale > 1e-6) ? (1.0 / tiling_scale) : 1.0;
|
|
float amplitude = (invert ? -1.0 : 1.0) * depth_mm * inv_tiling * clamp(weight, 0.0, 1.0);
|
|
// uv was rotated by project_uv() while t/b are the unrotated model axes, so rotate the
|
|
// gradient back into the axes' frame.
|
|
float cs = cos(rotation_rad);
|
|
float sn = sin(rotation_rad);
|
|
// One uv unit is tiling_scale mm along u but tiling_scale / tex_aspect mm along v, so the v
|
|
// component of the gradient carries the extra factor before being rotated back into t/b.
|
|
vec2 g = vec2(dh_duv.x, dh_duv.y * tex_aspect);
|
|
// ...and back out of raw-projection units into millimetres of travel along t / b, which is a
|
|
// no-op except for the spherical longitude (see projection_axes()). After the inverse rotation,
|
|
// because units_per_mm is expressed in the t/b frame rather than in uv.
|
|
vec2 slope = amplitude * vec2(g.x * cs + g.y * sn, -g.x * sn + g.y * cs) * units_per_mm;
|
|
|
|
vec3 gradient = slope.x * t + slope.y * b;
|
|
gradient -= triangle_normal * dot(triangle_normal, gradient);
|
|
triangle_normal = normalize(triangle_normal - gradient);
|
|
}
|
|
|
|
vec3 eye_normal = normalize(view_normal_matrix * triangle_normal);
|
|
float NdotL = max(dot(eye_normal, LIGHT_TOP_DIR), 0.0);
|
|
|
|
vec2 intensity = vec2(0.0);
|
|
intensity.x = INTENSITY_AMBIENT + NdotL * LIGHT_TOP_DIFFUSE;
|
|
vec3 position = (view_model_matrix * model_pos).xyz;
|
|
intensity.y = LIGHT_TOP_SPECULAR * pow(max(dot(-normalize(position), reflect(-LIGHT_TOP_DIR, eye_normal)), 0.0), LIGHT_TOP_SHININESS);
|
|
|
|
NdotL = max(dot(eye_normal, LIGHT_FRONT_DIR), 0.0);
|
|
intensity.x += NdotL * LIGHT_FRONT_DIFFUSE;
|
|
|
|
// Diffuse albedo: the image's colour at this fragment, snapped to the nearest printable colour.
|
|
// Only the albedo - the specular term (intensity.y) stays white - so a coloured fragment reads as
|
|
// the same material under the same light, and the relief this preview exists to show is unaffected.
|
|
vec3 albedo = uniform_color.rgb;
|
|
if (palette_count > 0 && has_color_tex && have_uv && weight > 0.0)
|
|
albedo = palette_rgb[nearest_palette_entry(texture(color_tex, color_uv).rgb)];
|
|
out_color = vec4(vec3(intensity.y) + albedo * intensity.x, uniform_color.a);
|
|
}
|