daedalus/Source/SysGLES/HLEGraphics/n64.psh

#version 310 es

precision highp float;
precision highp int;

// --------------------------------------------------------
// 1. Macro trick to avoid redefinition of built-in in ES 3.0
// --------------------------------------------------------
#ifdef GL_ES
#if __VERSION__ < 310
    // Only apply if we really end up at ES 3.0 environment
    #define texelFetch myTexelFetch
#endif
#endif

// --------------------------------------------------------
// 2. Provide the actual function body for our "texelFetch" calls
// --------------------------------------------------------
vec4 myTexelFetch(sampler2D tex, ivec2 coords, int lod)
{
    // Emulate texelFetch(...) on a float sampler2D
    ivec2 size = textureSize(tex, 0);
    vec2 uv    = (vec2(coords) + 0.5) / vec2(size);
    return textureLod(tex, uv, 0.0);
}

// --------------------------------------------------------
// Your original uniform declarations
// --------------------------------------------------------
uniform sampler2D uTexture0;
uniform sampler2D uTexture1;

uniform vec2 uTexScale0;		
uniform vec2 uTexScale1;
uniform vec4 uPrimColour;
uniform vec4 uEnvColour;
uniform float uPrimLODFrac;

// Inputs from the vertex shader
in vec2 v_st;
in vec4 v_col;

// Final fragment color
out vec4 fragcol;

// Tiling / Clamping
uniform bvec2 uTileClampEnable0;
uniform ivec2 uTileTL0;    // 10.2 fixed point
uniform ivec2 uTileBR0;    // 10.2 fixed point
uniform vec2  uTileShift0; // floating point
uniform ivec2 uTileMask0;  // 10.5 fixed point
uniform ivec2 uTileMirror0;// 10.5 fixed point

uniform bvec2 uTileClampEnable1;
uniform ivec2 uTileTL1;
uniform ivec2 uTileBR1;
uniform vec2  uTileShift1;
uniform ivec2 uTileMask1;
uniform ivec2 uTileMirror1;

// Example uniform
uniform int uFoo;

// --------------------------------------------------------
// Helper Functions (optimized)
// --------------------------------------------------------

// Replace mix on ivec2 by using the ternary operator for each component.
ivec2 imix(ivec2 a, ivec2 b, bvec2 c)
{
    return ivec2(c.x ? b.x : a.x, c.y ? b.y : a.y);
}

// Multiply an integer vector by a float vector.
ivec2 shift(ivec2 coord, vec2 shift_scale)
{
    return ivec2(float(coord.x) * shift_scale.x, float(coord.y) * shift_scale.y);
}

ivec2 clampPoint(ivec2 coord,
                 ivec2 tile_tl,
                 ivec2 tile_br,
                 bvec2 clamp_enable)
{
    ivec2 coord_clamped = clamp(coord, tile_tl << 3, tile_br << 3);
    ivec2 coord_out     = imix(coord, coord_clamped, clamp_enable);
    ivec2 coord_relative = ((coord_out >> 3) - tile_tl) >> 2;
    return coord_relative;
}

ivec2 clampBilinear(ivec2 coord,
                    ivec2 tile_tl,
                    ivec2 tile_br,
                    bvec2 clamp_enable,
                    out ivec2 frac)
{
    ivec2 tl = tile_tl << 3;  // Precompute shifted tile top‐left.
    ivec2 br = tile_br << 3;  
    ivec2 coord_clamped = clamp(coord, tl, br);
    ivec2 coord_out     = imix(coord, coord_clamped, clamp_enable);
    ivec2 coord_relative = coord_out - tl; // Use tl instead of recomputing (tile_tl << 3)
    frac = coord_relative & 0x1f;      
    return coord_relative >> 5;        
}

ivec2 mask(ivec2 coord, ivec2 mirror_bits, ivec2 mask_bits)
{
    bvec2 mirror = notEqual(coord & mirror_bits, ivec2(0));
    coord = imix(coord, ~coord, mirror);
    coord &= mask_bits;
    return coord;
}

vec4 bilinear(vec4 col_00, vec4 col_01,
              vec4 col_10, vec4 col_11,
              ivec2 frac)
{
    vec2 fracf = vec2(frac) / 32.0;
    vec4 a = mix(col_00, col_10, fracf.x);
    vec4 b = mix(col_01, col_11, fracf.x);
    return mix(a, b, fracf.y);
}

vec4 bilinear_n64(vec4 col_00, vec4 col_01,
                  vec4 col_10, vec4 col_11,
                  ivec2 frac)
{
    bool upper     = (frac.x + frac.y) >= 0x20;
    bvec4 uppersel = bvec4(upper);
    vec4 col0  = mix(col_00, col_11, uppersel);
    vec4 col1  = mix(col_10, col_01, uppersel);
    vec4 col2  = mix(col_01, col_10, uppersel);
    vec2 fracf = vec2(imix(frac, 0x20 - frac, bvec2(upper, upper))) / 32.0;
    vec4 colA = col0 + (fracf.x * (col1 - col0));
    vec4 colB = col0 + (fracf.y * (col2 - col0));
    return clamp(colA + colB - col0, 0.0, 1.0);
}

// --------------------------------------------------------
// Texture Fetch Functions
// --------------------------------------------------------

vec4 fetchBilinear(ivec2 st_in,
                   vec2 shift_scale,
                   ivec2 mirror_bits,
                   ivec2 mask_bits,
                   ivec2 tile_tl,
                   ivec2 tile_br,
                   bvec2 clamp_enable,
                   sampler2D tex,
                   vec2 tex_scale)
{
    ivec2 frac;
    ivec2 uv0 = st_in;
    uv0 = shift(uv0, shift_scale);
    uv0 = clampBilinear(uv0, tile_tl, tile_br, clamp_enable, frac);
    ivec2 uv1 = uv0 + ivec2(1);
    uv0 = mask(uv0, mirror_bits, mask_bits);
    uv1 = mask(uv1, mirror_bits, mask_bits);
    vec4 col_00 = texelFetch(tex, uv0, 0);
    vec4 col_01 = texelFetch(tex, ivec2(uv0.x, uv1.y), 0);
    vec4 col_10 = texelFetch(tex, ivec2(uv1.x, uv0.y), 0);
    vec4 col_11 = texelFetch(tex, uv1, 0);
    return bilinear(col_00, col_01, col_10, col_11, frac);
}

vec4 fetchBilinearClampedCommon(
    vec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale,
    ivec2 bilerp_wrap_enable)
{
    ivec2 frac;
    ivec2 uv0 = ivec2(st_in);
    uv0 = shift(uv0, shift_scale);
    uv0 = clampBilinear(uv0, tile_tl, tile_br, clamp_enable, frac);
    ivec2 uv1 = uv0 + ivec2(1);
    uv0 = mask(uv0, mirror_bits, mask_bits);
    uv1 = mask(uv1, mirror_bits, mask_bits);
    // If uv1 < uv0, remove some fractional bits.
    frac = imix(frac, frac & bilerp_wrap_enable, lessThan(uv1, uv0));
    vec4 col_00 = texelFetch(tex, uv0, 0);
    vec4 col_01 = texelFetch(tex, ivec2(uv0.x, uv1.y), 0);
    vec4 col_10 = texelFetch(tex, ivec2(uv1.x, uv0.y), 0);
    vec4 col_11 = texelFetch(tex, uv1, 0);
    return bilinear(col_00, col_01, col_10, col_11, frac);
}

vec4 fetchBilinearClampedS(
    vec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale)
{
    return fetchBilinearClampedCommon(
        st_in, shift_scale, mirror_bits, mask_bits,
        tile_tl, tile_br, clamp_enable, tex, tex_scale,
        ivec2(0, -1));
}

vec4 fetchBilinearClampedT(
    vec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale)
{
    return fetchBilinearClampedCommon(
        st_in, shift_scale, mirror_bits, mask_bits,
        tile_tl, tile_br, clamp_enable, tex, tex_scale,
        ivec2(-1, 0));
}

vec4 fetchBilinearClampedST(
    vec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale)
{
    return fetchBilinearClampedCommon(
        st_in, shift_scale, mirror_bits, mask_bits,
        tile_tl, tile_br, clamp_enable, tex, tex_scale,
        ivec2(0, 0));
}

vec4 fetchPoint(
    ivec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale)
{
    ivec2 uv = st_in;
    uv = shift(uv, shift_scale);
    uv = clampPoint(uv, tile_tl, tile_br, clamp_enable);
    uv = mask(uv, mirror_bits, mask_bits);
    return texelFetch(tex, uv, 0);
}

vec4 fetchCopy(
    ivec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale)
{
    ivec2 uv = st_in;
    uv = shift(uv, shift_scale);
    uv = (((uv >> 3) - tile_tl) >> 2) & ivec2(0x1fff);
    uv = mask(uv, mirror_bits, mask_bits);
    return texelFetch(tex, uv, 0);
}

vec4 fetchSimple(
    vec2 st_in,
    vec2 shift_scale,
    ivec2 mirror_bits,
    ivec2 mask_bits,
    ivec2 tile_tl,
    ivec2 tile_br,
    bvec2 clamp_enable,
    sampler2D tex,
    vec2 tex_scale)
{
    ivec2 uv = ivec2(st_in);
    uv = shift(uv, shift_scale);
    vec2 uvf = (vec2(uv) - vec2(tile_tl << 3)) * (tex_scale / 32.0);
    return texture(tex, uvf);
}