diff options
Diffstat (limited to 'servers/rendering/renderer_rd/shaders')
10 files changed, 108 insertions, 68 deletions
diff --git a/servers/rendering/renderer_rd/shaders/cluster_render.glsl b/servers/rendering/renderer_rd/shaders/cluster_render.glsl index 2fe230f0bf..8c26a67926 100644 --- a/servers/rendering/renderer_rd/shaders/cluster_render.glsl +++ b/servers/rendering/renderer_rd/shaders/cluster_render.glsl @@ -64,7 +64,7 @@ void main() { #version 450 #VERSION_DEFINES - +#ifndef MOLTENVK_USED // Metal will corrupt GPU state otherwise #if defined(has_GL_KHR_shader_subgroup_ballot) && defined(has_GL_KHR_shader_subgroup_arithmetic) && defined(has_GL_KHR_shader_subgroup_vote) #extension GL_KHR_shader_subgroup_ballot : enable @@ -73,6 +73,7 @@ void main() { #define USE_SUBGROUPS #endif +#endif layout(location = 0) in float depth_interp; layout(location = 1) in flat uint element_index; @@ -141,7 +142,11 @@ void main() { } } #else - if (!gl_HelperInvocation) { +// MoltenVK/Metal fails to compile shaders using gl_HelperInvocation for some GPUs +#ifndef MOLTENVK_USED + if (!gl_HelperInvocation) +#endif + { atomicOr(cluster_render.data[usage_write_offset], usage_write_bit); } #endif @@ -161,7 +166,11 @@ void main() { } } #else - if (!gl_HelperInvocation) { +// MoltenVK/Metal fails to compile shaders using gl_HelperInvocation for some GPUs +#ifndef MOLTENVK_USED + if (!gl_HelperInvocation) +#endif + { atomicOr(cluster_render.data[z_write_offset], z_write_bit); } #endif diff --git a/servers/rendering/renderer_rd/shaders/effects/blur_raster.glsl b/servers/rendering/renderer_rd/shaders/effects/blur_raster.glsl index cb06250cf2..31aabbe9d2 100644 --- a/servers/rendering/renderer_rd/shaders/effects/blur_raster.glsl +++ b/servers/rendering/renderer_rd/shaders/effects/blur_raster.glsl @@ -53,30 +53,31 @@ void main() { #ifdef MODE_GAUSSIAN_BLUR - // Simpler blur uses SIGMA2 for the gaussian kernel for a stronger effect - - // note, for blur blur.luminance_multiplier is irrelavant, we would be multiplying and then dividing by this amount. - - if (bool(blur.flags & FLAG_HORIZONTAL)) { - vec2 pix_size = blur.pixel_size; - pix_size *= 0.5; //reading from larger buffer, so use more samples - vec4 color = texture(source_color, uv_interp + vec2(0.0, 0.0) * pix_size) * 0.214607; - color += texture(source_color, uv_interp + vec2(1.0, 0.0) * pix_size) * 0.189879; - color += texture(source_color, uv_interp + vec2(2.0, 0.0) * pix_size) * 0.131514; - color += texture(source_color, uv_interp + vec2(3.0, 0.0) * pix_size) * 0.071303; - color += texture(source_color, uv_interp + vec2(-1.0, 0.0) * pix_size) * 0.189879; - color += texture(source_color, uv_interp + vec2(-2.0, 0.0) * pix_size) * 0.131514; - color += texture(source_color, uv_interp + vec2(-3.0, 0.0) * pix_size) * 0.071303; - frag_color = color; - } else { - vec2 pix_size = blur.pixel_size; - vec4 color = texture(source_color, uv_interp + vec2(0.0, 0.0) * pix_size) * 0.38774; - color += texture(source_color, uv_interp + vec2(0.0, 1.0) * pix_size) * 0.24477; - color += texture(source_color, uv_interp + vec2(0.0, 2.0) * pix_size) * 0.06136; - color += texture(source_color, uv_interp + vec2(0.0, -1.0) * pix_size) * 0.24477; - color += texture(source_color, uv_interp + vec2(0.0, -2.0) * pix_size) * 0.06136; - frag_color = color; - } + // For Gaussian Blur we use 13 taps in a single pass instead of 12 taps over 2 passes. + // This minimizes the number of times we change framebuffers which is very important for mobile. + // Source: http://www.iryoku.com/next-generation-post-processing-in-call-of-duty-advanced-warfare + vec4 A = texture(source_color, uv_interp + blur.pixel_size * vec2(-1.0, -1.0)); + vec4 B = texture(source_color, uv_interp + blur.pixel_size * vec2(0.0, -1.0)); + vec4 C = texture(source_color, uv_interp + blur.pixel_size * vec2(1.0, -1.0)); + vec4 D = texture(source_color, uv_interp + blur.pixel_size * vec2(-0.5, -0.5)); + vec4 E = texture(source_color, uv_interp + blur.pixel_size * vec2(0.5, -0.5)); + vec4 F = texture(source_color, uv_interp + blur.pixel_size * vec2(-1.0, 0.0)); + vec4 G = texture(source_color, uv_interp); + vec4 H = texture(source_color, uv_interp + blur.pixel_size * vec2(1.0, 0.0)); + vec4 I = texture(source_color, uv_interp + blur.pixel_size * vec2(-0.5, 0.5)); + vec4 J = texture(source_color, uv_interp + blur.pixel_size * vec2(0.5, 0.5)); + vec4 K = texture(source_color, uv_interp + blur.pixel_size * vec2(-1.0, 1.0)); + vec4 L = texture(source_color, uv_interp + blur.pixel_size * vec2(0.0, 1.0)); + vec4 M = texture(source_color, uv_interp + blur.pixel_size * vec2(1.0, 1.0)); + + float base_weight = 0.5 / 4.0; + float lesser_weight = 0.125 / 4.0; + + frag_color = (D + E + I + J) * base_weight; + frag_color += (A + B + G + F) * lesser_weight; + frag_color += (B + C + H + G) * lesser_weight; + frag_color += (F + G + L + K) * lesser_weight; + frag_color += (G + H + M + L) * lesser_weight; #endif #ifdef MODE_GAUSSIAN_GLOW diff --git a/servers/rendering/renderer_rd/shaders/effects/copy.glsl b/servers/rendering/renderer_rd/shaders/effects/copy.glsl index bfe329b8ec..5cc2ed7622 100644 --- a/servers/rendering/renderer_rd/shaders/effects/copy.glsl +++ b/servers/rendering/renderer_rd/shaders/effects/copy.glsl @@ -194,10 +194,10 @@ void main() { color = min(color * feedback, vec4(params.glow_luminance_cap)); } -#endif +#endif // MODE_GLOW imageStore(dest_buffer, pos + params.target, color); -#endif +#endif // MODE_GAUSSIAN_BLUR #ifdef MODE_SIMPLE_COPY @@ -227,7 +227,7 @@ void main() { imageStore(dest_buffer, pos + params.target, color); -#endif +#endif // MODE_SIMPLE_COPY #ifdef MODE_SIMPLE_COPY_DEPTH @@ -239,7 +239,7 @@ void main() { imageStore(dest_buffer, pos + params.target, vec4(color.r)); -#endif +#endif // MODE_SIMPLE_COPY_DEPTH #ifdef MODE_LINEARIZE_DEPTH_COPY @@ -253,7 +253,7 @@ void main() { } imageStore(dest_buffer, pos + params.target, color); -#endif +#endif // MODE_LINEARIZE_DEPTH_COPY #if defined(MODE_CUBEMAP_TO_PANORAMA) || defined(MODE_CUBEMAP_ARRAY_TO_PANORAMA) @@ -276,7 +276,7 @@ void main() { vec4 color = textureLod(source_color, vec4(normal, params.camera_z_far), 0.0); //the biggest the lod the least the acne #endif imageStore(dest_buffer, pos + params.target, color); -#endif +#endif // defined(MODE_CUBEMAP_TO_PANORAMA) || defined(MODE_CUBEMAP_ARRAY_TO_PANORAMA) #ifdef MODE_SET_COLOR imageStore(dest_buffer, pos + params.target, params.set_color); diff --git a/servers/rendering/renderer_rd/shaders/effects/copy_to_fb.glsl b/servers/rendering/renderer_rd/shaders/effects/copy_to_fb.glsl index 1c17eabb56..4d4e983b7f 100644 --- a/servers/rendering/renderer_rd/shaders/effects/copy_to_fb.glsl +++ b/servers/rendering/renderer_rd/shaders/effects/copy_to_fb.glsl @@ -26,7 +26,11 @@ layout(push_constant, std430) uniform Params { bool use_section; bool force_luminance; - uint pad[3]; + bool alpha_to_zero; + bool srgb; + bool alpha_to_one; + + vec4 color; } params; @@ -72,7 +76,9 @@ layout(push_constant, std430) uniform Params { bool force_luminance; bool alpha_to_zero; bool srgb; - uint pad; + bool alpha_to_one; + + vec4 color; } params; @@ -105,6 +111,10 @@ vec3 linear_to_srgb(vec3 color) { } void main() { +#ifdef MODE_SET_COLOR + frag_color = params.color; +#else + #ifdef MULTIVIEW vec3 uv = uv_interp; #else @@ -164,6 +174,10 @@ void main() { if (params.srgb) { color.rgb = linear_to_srgb(color.rgb); } + if (params.alpha_to_one) { + color.a = 1.0; + } frag_color = color; +#endif // MODE_SET_COLOR } diff --git a/servers/rendering/renderer_rd/shaders/effects/taa_resolve.glsl b/servers/rendering/renderer_rd/shaders/effects/taa_resolve.glsl index b0a0839836..02566d8e35 100644 --- a/servers/rendering/renderer_rd/shaders/effects/taa_resolve.glsl +++ b/servers/rendering/renderer_rd/shaders/effects/taa_resolve.glsl @@ -32,7 +32,9 @@ // Based on Spartan Engine's TAA implementation (without TAA upscale). // <https://github.com/PanosK92/SpartanEngine/blob/a8338d0609b85dc32f3732a5c27fb4463816a3b9/Data/shaders/temporal_antialiasing.hlsl> +#ifndef MOLTENVK_USED #define USE_SUBGROUPS +#endif // MOLTENVK_USED #define GROUP_SIZE 8 #define FLT_MIN 0.00000001 diff --git a/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered.glsl b/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered.glsl index 1a8a1f3aa3..7440c5748b 100644 --- a/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered.glsl +++ b/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered.glsl @@ -97,9 +97,7 @@ layout(location = 8) out vec4 prev_screen_position; #ifdef MATERIAL_UNIFORMS_USED layout(set = MATERIAL_UNIFORM_SET, binding = 0, std140) uniform MaterialUniforms{ - #MATERIAL_UNIFORMS - } material; #endif @@ -691,7 +689,7 @@ vec4 fog_process(vec3 vertex) { void cluster_get_item_range(uint p_offset, out uint item_min, out uint item_max, out uint item_from, out uint item_to) { uint item_min_max = cluster_buffer.data[p_offset]; - item_min = item_min_max & 0xFFFF; + item_min = item_min_max & 0xFFFFu; item_max = item_min_max >> 16; item_from = item_min >> 5; @@ -828,7 +826,8 @@ void fragment_shader(in SceneData scene_data) { // alpha hash can be used in unison with alpha antialiasing #ifdef ALPHA_HASH_USED - if (alpha < compute_alpha_hash_threshold(vertex, alpha_hash_scale)) { + vec3 object_pos = (inverse(read_model_matrix) * inv_view_matrix * vec4(vertex, 1.0)).xyz; + if (alpha < compute_alpha_hash_threshold(object_pos, alpha_hash_scale)) { discard; } #endif // ALPHA_HASH_USED @@ -958,9 +957,9 @@ void fragment_shader(in SceneData scene_data) { while (merged_mask != 0) { uint bit = findMSB(merged_mask); - merged_mask &= ~(1 << bit); + merged_mask &= ~(1u << bit); #ifdef USE_SUBGROUPS - if (((1 << bit) & mask) == 0) { //do not process if not originally here + if (((1u << bit) & mask) == 0) { //do not process if not originally here continue; } #endif @@ -1419,9 +1418,9 @@ void fragment_shader(in SceneData scene_data) { while (merged_mask != 0) { uint bit = findMSB(merged_mask); - merged_mask &= ~(1 << bit); + merged_mask &= ~(1u << bit); #ifdef USE_SUBGROUPS - if (((1 << bit) & mask) == 0) { //do not process if not originally here + if (((1u << bit) & mask) == 0) { //do not process if not originally here continue; } #endif @@ -1447,18 +1446,24 @@ void fragment_shader(in SceneData scene_data) { } //finalize ambient light here - ambient_light *= albedo.rgb; - ambient_light *= ao; + { +#if defined(AMBIENT_LIGHT_DISABLED) + ambient_light = vec3(0.0, 0.0, 0.0); +#else + ambient_light *= albedo.rgb; + ambient_light *= ao; + + if (bool(implementation_data.ss_effects_flags & SCREEN_SPACE_EFFECTS_FLAGS_USE_SSIL)) { + vec4 ssil = textureLod(sampler2D(ssil_buffer, material_samplers[SAMPLER_LINEAR_CLAMP]), screen_uv, 0.0); + ambient_light *= 1.0 - ssil.a; + ambient_light += ssil.rgb * albedo.rgb; + } +#endif // AMBIENT_LIGHT_DISABLED + } // convert ao to direct light ao ao = mix(1.0, ao, ao_light_affect); - if (bool(implementation_data.ss_effects_flags & SCREEN_SPACE_EFFECTS_FLAGS_USE_SSIL)) { - vec4 ssil = textureLod(sampler2D(ssil_buffer, material_samplers[SAMPLER_LINEAR_CLAMP]), screen_uv, 0.0); - ambient_light *= 1.0 - ssil.a; - ambient_light += ssil.rgb * albedo.rgb; - } - //this saves some VGPRs vec3 f0 = F0(metallic, specular, albedo); @@ -1775,9 +1780,9 @@ void fragment_shader(in SceneData scene_data) { float shadow = 1.0; #ifndef SHADOWS_DISABLED if (i < 4) { - shadow = float(shadow0 >> (i * 8) & 0xFF) / 255.0; + shadow = float(shadow0 >> (i * 8u) & 0xFFu) / 255.0; } else { - shadow = float(shadow1 >> ((i - 4) * 8) & 0xFF) / 255.0; + shadow = float(shadow1 >> ((i - 4u) * 8u) & 0xFFu) / 255.0; } shadow = shadow * directional_lights.data[i].shadow_opacity + 1.0 - directional_lights.data[i].shadow_opacity; @@ -1839,9 +1844,9 @@ void fragment_shader(in SceneData scene_data) { while (merged_mask != 0) { uint bit = findMSB(merged_mask); - merged_mask &= ~(1 << bit); + merged_mask &= ~(1u << bit); #ifdef USE_SUBGROUPS - if (((1 << bit) & mask) == 0) { //do not process if not originally here + if (((1u << bit) & mask) == 0) { //do not process if not originally here continue; } #endif @@ -1910,9 +1915,9 @@ void fragment_shader(in SceneData scene_data) { while (merged_mask != 0) { uint bit = findMSB(merged_mask); - merged_mask &= ~(1 << bit); + merged_mask &= ~(1u << bit); #ifdef USE_SUBGROUPS - if (((1 << bit) & mask) == 0) { //do not process if not originally here + if (((1u << bit) & mask) == 0) { //do not process if not originally here continue; } #endif @@ -2065,7 +2070,7 @@ void fragment_shader(in SceneData scene_data) { float sGreen = floor((cGreen / pow(2.0f, exps - B - N)) + 0.5f); float sBlue = floor((cBlue / pow(2.0f, exps - B - N)) + 0.5f); //store as 8985 to have 2 extra neighbour bits - uint light_rgbe = ((uint(sRed) & 0x1FF) >> 1) | ((uint(sGreen) & 0x1FF) << 8) | (((uint(sBlue) & 0x1FF) >> 1) << 17) | ((uint(exps) & 0x1F) << 25); + uint light_rgbe = ((uint(sRed) & 0x1FFu) >> 1) | ((uint(sGreen) & 0x1FFu) << 8) | (((uint(sBlue) & 0x1FFu) >> 1) << 17) | ((uint(exps) & 0x1Fu) << 25); imageStore(emission_grid, grid_pos, uvec4(light_rgbe)); imageStore(emission_aniso_grid, grid_pos, uvec4(light_aniso)); @@ -2099,8 +2104,8 @@ void fragment_shader(in SceneData scene_data) { if (bool(instances.data[instance_index].flags & INSTANCE_FLAGS_USE_VOXEL_GI)) { // process voxel_gi_instances uint index1 = instances.data[instance_index].gi_offset & 0xFFFF; uint index2 = instances.data[instance_index].gi_offset >> 16; - voxel_gi_buffer.x = index1 & 0xFF; - voxel_gi_buffer.y = index2 & 0xFF; + voxel_gi_buffer.x = index1 & 0xFFu; + voxel_gi_buffer.y = index2 & 0xFFu; } else { voxel_gi_buffer.x = 0xFF; voxel_gi_buffer.y = 0xFF; diff --git a/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered_inc.glsl b/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered_inc.glsl index 0d36b98645..3a45ab0059 100644 --- a/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered_inc.glsl +++ b/servers/rendering/renderer_rd/shaders/forward_clustered/scene_forward_clustered_inc.glsl @@ -4,14 +4,15 @@ #define MAX_VOXEL_GI_INSTANCES 8 #define MAX_VIEWS 2 +#ifndef MOLTENVK_USED #if defined(has_GL_KHR_shader_subgroup_ballot) && defined(has_GL_KHR_shader_subgroup_arithmetic) #extension GL_KHR_shader_subgroup_ballot : enable #extension GL_KHR_shader_subgroup_arithmetic : enable #define USE_SUBGROUPS - #endif +#endif // MOLTENVK_USED #if defined(USE_MULTIVIEW) && defined(has_VK_KHR_multiview) #extension GL_EXT_multiview : enable diff --git a/servers/rendering/renderer_rd/shaders/forward_mobile/scene_forward_mobile.glsl b/servers/rendering/renderer_rd/shaders/forward_mobile/scene_forward_mobile.glsl index 33fd4c35b1..cc44cff799 100644 --- a/servers/rendering/renderer_rd/shaders/forward_mobile/scene_forward_mobile.glsl +++ b/servers/rendering/renderer_rd/shaders/forward_mobile/scene_forward_mobile.glsl @@ -779,7 +779,8 @@ void main() { // alpha hash can be used in unison with alpha antialiasing #ifdef ALPHA_HASH_USED - if (alpha < compute_alpha_hash_threshold(vertex, alpha_hash_scale)) { + vec3 object_pos = (inverse(read_model_matrix) * inv_view_matrix * vec4(vertex, 1.0)).xyz; + if (alpha < compute_alpha_hash_threshold(object_pos, alpha_hash_scale)) { discard; } #endif // ALPHA_HASH_USED @@ -1172,8 +1173,14 @@ void main() { } //Reflection probes // finalize ambient light here - ambient_light *= albedo.rgb; - ambient_light *= ao; + { +#if defined(AMBIENT_LIGHT_DISABLED) + ambient_light = vec3(0.0, 0.0, 0.0); +#else + ambient_light *= albedo.rgb; + ambient_light *= ao; +#endif // AMBIENT_LIGHT_DISABLED + } // convert ao to direct light ao ao = mix(1.0, ao, ao_light_affect); diff --git a/servers/rendering/renderer_rd/shaders/particles.glsl b/servers/rendering/renderer_rd/shaders/particles.glsl index 3a6dd579b9..9f6aa7adc0 100644 --- a/servers/rendering/renderer_rd/shaders/particles.glsl +++ b/servers/rendering/renderer_rd/shaders/particles.glsl @@ -462,7 +462,7 @@ void main() { if (any(lessThan(uvw_pos, vec3(0.0))) || any(greaterThan(uvw_pos, vec3(1.0)))) { continue; } - vec3 s = texture(sampler3D(sdf_vec_textures[FRAME.attractors[i].texture_index], material_samplers[SAMPLER_LINEAR_CLAMP]), uvw_pos).xyz * 2.0 - 1.0; + vec3 s = texture(sampler3D(sdf_vec_textures[FRAME.attractors[i].texture_index], material_samplers[SAMPLER_LINEAR_CLAMP]), uvw_pos).xyz * -2.0 + 1.0; dir = mat3(FRAME.attractors[i].transform) * safe_normalize(s); //revert direction amount = length(s); diff --git a/servers/rendering/renderer_rd/shaders/scene_forward_aa_inc.glsl b/servers/rendering/renderer_rd/shaders/scene_forward_aa_inc.glsl index 97c913d489..71510ee06a 100644 --- a/servers/rendering/renderer_rd/shaders/scene_forward_aa_inc.glsl +++ b/servers/rendering/renderer_rd/shaders/scene_forward_aa_inc.glsl @@ -11,7 +11,8 @@ float hash_3d(vec3 p) { float compute_alpha_hash_threshold(vec3 pos, float hash_scale) { vec3 dx = dFdx(pos); - vec3 dy = dFdx(pos); + vec3 dy = dFdy(pos); + float delta_max_sqr = max(length(dx), length(dy)); float pix_scale = 1.0 / (hash_scale * delta_max_sqr); @@ -32,9 +33,9 @@ float compute_alpha_hash_threshold(vec3 pos, float hash_scale) { 1.0 - ((1.0 - a_interp) * (1.0 - a_interp) / (2.0 * min_lerp * (1.0 - min_lerp)))); float alpha_hash_threshold = - (lerp_factor < (1.0 - min_lerp)) ? ((lerp_factor < min_lerp) ? cases.x : cases.y) : cases.z; + (a_interp < (1.0 - min_lerp)) ? ((a_interp < min_lerp) ? cases.x : cases.y) : cases.z; - return clamp(alpha_hash_threshold, 0.0, 1.0); + return clamp(alpha_hash_threshold, 0.00001, 1.0); } #endif // ALPHA_HASH_USED |