log.txt for X-Plane 11.30r3 (build 113014 64-bit) compiled on Jan 9 2019 09:57:43 X-Plane started on Sun Jan 13 11:10:38 2019 This log file is generated automatically by Laminar Research applications and contains diagnostics about your graphics hardware, installation, and any error conditions. If you need to contact tech support or file a bug, please send us this file. NOTE: this file is rewritten every time you start ANY of your X-System applications. Windows 10.0 (build 17134/2) This is a 64-bit version of Windows. CPU type: 8664 Physical Memory (total for computer): 34271838208 Maximum Virtual Memory (for X-Plane only): 140737488224256 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 CPU 0: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Speed (mhz): 3492 X-System folder:'C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/', case sensitive=0 WGL_ARB_buffer_region WGL_ARB_create_context WGL_ARB_create_context_no_error WGL_ARB_create_context_profile WGL_ARB_create_context_robustness WGL_ARB_context_flush_control WGL_ARB_extensions_string WGL_ARB_make_current_read WGL_ARB_multisample WGL_ARB_pbuffer WGL_ARB_pixel_format WGL_ARB_pixel_format_float WGL_ARB_render_texture WGL_ATI_pixel_format_float WGL_EXT_colorspace WGL_EXT_create_context_es_profile WGL_EXT_create_context_es2_profile WGL_EXT_extensions_string WGL_EXT_framebuffer_sRGB WGL_EXT_pixel_format_packed_float WGL_EXT_swap_control WGL_EXT_swap_control_tear WGL_NVX_DX_interop WGL_NV_DX_interop WGL_NV_DX_interop2 WGL_NV_copy_image WGL_NV_delay_before_swap WGL_NV_float_buffer WGL_NV_multisample_coverage WGL_NV_render_depth_texture WGL_NV_render_texture_rectangle OpenGL Vendor : NVIDIA Corporation OpenGL Render : GeForce GTX 980 Ti/PCIe/SSE2 OpenGL Version : 4.6.0 NVIDIA 417.35 (460/0) OpenGL Extensions: GL_AMD_multi_draw_indirect GL_AMD_seamless_cubemap_per_texture GL_AMD_vertex_shader_viewport_index GL_AMD_vertex_shader_layer GL_ARB_arrays_of_arrays GL_ARB_base_instance GL_ARB_bindless_texture GL_ARB_blend_func_extended GL_ARB_buffer_storage GL_ARB_clear_buffer_object GL_ARB_clear_texture GL_ARB_clip_control GL_ARB_color_buffer_float GL_ARB_compatibility GL_ARB_compressed_texture_pixel_storage GL_ARB_conservative_depth GL_ARB_compute_shader GL_ARB_compute_variable_group_size GL_ARB_conditional_render_inverted GL_ARB_copy_buffer GL_ARB_copy_image GL_ARB_cull_distance GL_ARB_debug_output GL_ARB_depth_buffer_float GL_ARB_depth_clamp GL_ARB_depth_texture GL_ARB_derivative_control GL_ARB_direct_state_access GL_ARB_draw_buffers GL_ARB_draw_buffers_blend GL_ARB_draw_indirect GL_ARB_draw_elements_base_vertex GL_ARB_draw_instanced GL_ARB_enhanced_layouts GL_ARB_ES2_compatibility GL_ARB_ES3_compatibility GL_ARB_ES3_1_compatibility GL_ARB_ES3_2_compatibility GL_ARB_explicit_attrib_location GL_ARB_explicit_uniform_location GL_ARB_fragment_coord_conventions GL_ARB_fragment_layer_viewport GL_ARB_fragment_program GL_ARB_fragment_program_shadow GL_ARB_fragment_shader GL_ARB_fragment_shader_interlock GL_ARB_framebuffer_no_attachments GL_ARB_framebuffer_object GL_ARB_framebuffer_sRGB GL_ARB_geometry_shader4 GL_ARB_get_program_binary GL_ARB_get_texture_sub_image GL_ARB_gl_spirv GL_ARB_gpu_shader5 GL_ARB_gpu_shader_fp64 GL_ARB_gpu_shader_int64 GL_ARB_half_float_pixel GL_ARB_half_float_vertex GL_ARB_imaging GL_ARB_indirect_parameters GL_ARB_instanced_arrays GL_ARB_internalformat_query GL_ARB_internalformat_query2 GL_ARB_invalidate_subdata GL_ARB_map_buffer_alignment GL_ARB_map_buffer_range GL_ARB_multi_bind GL_ARB_multi_draw_indirect GL_ARB_multisample GL_ARB_multitexture GL_ARB_occlusion_query GL_ARB_occlusion_query2 GL_ARB_parallel_shader_compile GL_ARB_pipeline_statistics_query GL_ARB_pixel_buffer_object GL_ARB_point_parameters GL_ARB_point_sprite GL_ARB_polygon_offset_clamp GL_ARB_post_depth_coverage GL_ARB_program_interface_query GL_ARB_provoking_vertex GL_ARB_query_buffer_object GL_ARB_robust_buffer_access_behavior GL_ARB_robustness GL_ARB_sample_locations GL_ARB_sample_shading GL_ARB_sampler_objects GL_ARB_seamless_cube_map GL_ARB_seamless_cubemap_per_texture GL_ARB_separate_shader_objects GL_ARB_shader_atomic_counter_ops GL_ARB_shader_atomic_counters GL_ARB_shader_ballot GL_ARB_shader_bit_encoding GL_ARB_shader_clock GL_ARB_shader_draw_parameters GL_ARB_shader_group_vote GL_ARB_shader_image_load_store GL_ARB_shader_image_size GL_ARB_shader_objects GL_ARB_shader_precision GL_ARB_shader_storage_buffer_object GL_ARB_shader_subroutine GL_ARB_shader_texture_image_samples GL_ARB_shader_texture_lod GL_ARB_shading_language_100 GL_ARB_shader_viewport_layer_array GL_ARB_shading_language_420pack GL_ARB_shading_language_include GL_ARB_shading_language_packing GL_ARB_shadow GL_ARB_sparse_buffer GL_ARB_sparse_texture GL_ARB_sparse_texture2 GL_ARB_sparse_texture_clamp GL_ARB_spirv_extensions GL_ARB_stencil_texturing GL_ARB_sync GL_ARB_tessellation_shader GL_ARB_texture_barrier GL_ARB_texture_border_clamp GL_ARB_texture_buffer_object GL_ARB_texture_buffer_object_rgb32 GL_ARB_texture_buffer_range GL_ARB_texture_compression GL_ARB_texture_compression_bptc GL_ARB_texture_compression_rgtc GL_ARB_texture_cube_map GL_ARB_texture_cube_map_array GL_ARB_texture_env_add GL_ARB_texture_env_combine GL_ARB_texture_env_crossbar GL_ARB_texture_env_dot3 GL_ARB_texture_filter_anisotropic GL_ARB_texture_filter_minmax GL_ARB_texture_float GL_ARB_texture_gather GL_ARB_texture_mirror_clamp_to_edge GL_ARB_texture_mirrored_repeat GL_ARB_texture_multisample GL_ARB_texture_non_power_of_two GL_ARB_texture_query_levels GL_ARB_texture_query_lod GL_ARB_texture_rectangle GL_ARB_texture_rg GL_ARB_texture_rgb10_a2ui GL_ARB_texture_stencil8 GL_ARB_texture_storage GL_ARB_texture_storage_multisample GL_ARB_texture_swizzle GL_ARB_texture_view GL_ARB_timer_query GL_ARB_transform_feedback2 GL_ARB_transform_feedback3 GL_ARB_transform_feedback_instanced GL_ARB_transform_feedback_overflow_query GL_ARB_transpose_matrix GL_ARB_uniform_buffer_object GL_ARB_vertex_array_bgra GL_ARB_vertex_array_object GL_ARB_vertex_attrib_64bit GL_ARB_vertex_attrib_binding GL_ARB_vertex_buffer_object GL_ARB_vertex_program GL_ARB_vertex_shader GL_ARB_vertex_type_10f_11f_11f_rev GL_ARB_vertex_type_2_10_10_10_rev GL_ARB_viewport_array GL_ARB_window_pos GL_ATI_draw_buffers GL_ATI_texture_float GL_ATI_texture_mirror_once GL_S3_s3tc GL_EXT_texture_env_add GL_EXT_abgr GL_EXT_bgra GL_EXT_bindable_uniform GL_EXT_blend_color GL_EXT_blend_equation_separate GL_EXT_blend_func_separate GL_EXT_blend_minmax GL_EXT_blend_subtract GL_EXT_compiled_vertex_array GL_EXT_Cg_shader GL_EXT_depth_bounds_test GL_EXT_direct_state_access GL_EXT_draw_buffers2 GL_EXT_draw_instanced GL_EXT_draw_range_elements GL_EXT_fog_coord GL_EXT_framebuffer_blit GL_EXT_framebuffer_multisample GL_EXTX_framebuffer_mixed_formats GL_EXT_framebuffer_multisample_blit_scaled GL_EXT_framebuffer_object GL_EXT_framebuffer_sRGB GL_EXT_geometry_shader4 GL_EXT_gpu_program_parameters GL_EXT_gpu_shader4 GL_EXT_multi_draw_arrays GL_EXT_packed_depth_stencil GL_EXT_packed_float GL_EXT_packed_pixels GL_EXT_pixel_buffer_object GL_EXT_point_parameters GL_EXT_polygon_offset_clamp GL_EXT_post_depth_coverage GL_EXT_provoking_vertex GL_EXT_raster_multisample GL_EXT_rescale_normal GL_EXT_secondary_color GL_EXT_separate_shader_objects GL_EXT_separate_specular_color GL_EXT_shader_image_load_formatted GL_EXT_shader_image_load_store GL_EXT_shader_integer_mix GL_EXT_shadow_funcs GL_EXT_sparse_texture2 GL_EXT_stencil_two_side GL_EXT_stencil_wrap GL_EXT_texture3D GL_EXT_texture_array GL_EXT_texture_buffer_object GL_EXT_texture_compression_dxt1 GL_EXT_texture_compression_latc GL_EXT_texture_compression_rgtc GL_EXT_texture_compression_s3tc GL_EXT_texture_cube_map GL_EXT_texture_edge_clamp GL_EXT_texture_env_combine GL_EXT_texture_env_dot3 GL_EXT_texture_filter_anisotropic GL_EXT_texture_filter_minmax GL_EXT_texture_integer GL_EXT_texture_lod GL_EXT_texture_lod_bias GL_EXT_texture_mirror_clamp GL_EXT_texture_object GL_EXT_texture_shared_exponent GL_EXT_texture_sRGB GL_EXT_texture_sRGB_R8 GL_EXT_texture_sRGB_decode GL_EXT_texture_storage GL_EXT_texture_swizzle GL_EXT_timer_query GL_EXT_transform_feedback2 GL_EXT_vertex_array GL_EXT_vertex_array_bgra GL_EXT_vertex_attrib_64bit GL_EXT_window_rectangles GL_EXT_import_sync_object GL_IBM_rasterpos_clip GL_IBM_texture_mirrored_repeat GL_KHR_context_flush_control GL_KHR_debug GL_EXT_memory_object GL_EXT_memory_object_win32 GL_EXT_win32_keyed_mutex GL_KHR_parallel_shader_compile GL_KHR_no_error GL_KHR_robust_buffer_access_behavior GL_KHR_robustness GL_EXT_semaphore GL_EXT_semaphore_win32 GL_KTX_buffer_region GL_NV_alpha_to_coverage_dither_control GL_NV_bindless_multi_draw_indirect GL_NV_bindless_multi_draw_indirect_count GL_NV_bindless_texture GL_NV_blend_equation_advanced GL_NV_blend_equation_advanced_coherent GL_NV_blend_minmax_factor GL_NV_blend_square GL_NV_command_list GL_NV_compute_program5 GL_NV_conditional_render GL_NV_conservative_raster GL_NV_conservative_raster_dilate GL_NV_copy_depth_to_color GL_NV_copy_image GL_NV_depth_buffer_float GL_NV_depth_clamp GL_NV_draw_texture GL_NV_draw_vulkan_image GL_NV_ES1_1_compatibility GL_NV_ES3_1_compatibility GL_NV_explicit_multisample GL_NV_feature_query GL_NV_fence GL_NV_fill_rectangle GL_NV_float_buffer GL_NV_fog_distance GL_NV_fragment_coverage_to_color GL_NV_fragment_program GL_NV_fragment_program_option GL_NV_fragment_program2 GL_NV_fragment_shader_interlock GL_NV_framebuffer_mixed_samples GL_NV_framebuffer_multisample_coverage GL_NV_geometry_shader4 GL_NV_geometry_shader_passthrough GL_NV_gpu_program4 GL_NV_internalformat_sample_query GL_NV_gpu_program4_1 GL_NV_gpu_program5 GL_NV_gpu_program5_mem_extended GL_NV_gpu_program_fp64 GL_NV_gpu_shader5 GL_NV_half_float GL_NV_light_max_exponent GL_NV_memory_attachment GL_NV_multisample_coverage GL_NV_multisample_filter_hint GL_NV_occlusion_query GL_NV_packed_depth_stencil GL_NV_parameter_buffer_object GL_NV_parameter_buffer_object2 GL_NV_path_rendering GL_NV_path_rendering_shared_edge GL_NV_pixel_data_range GL_NV_point_sprite GL_NV_primitive_restart GL_NV_query_resource GL_NV_query_resource_tag GL_NV_register_combiners GL_NV_register_combiners2 GL_NV_sample_locations GL_NV_sample_mask_override_coverage GL_NV_shader_atomic_counters GL_NV_shader_atomic_float GL_NV_shader_atomic_fp16_vector GL_NV_shader_atomic_int64 GL_NV_shader_buffer_load GL_NV_shader_storage_buffer_object GL_NV_texgen_reflection GL_NV_texture_barrier GL_NV_texture_compression_vtc GL_NV_texture_env_combine4 GL_NV_texture_multisample GL_NV_texture_rectangle GL_NV_texture_rectangle_compressed GL_NV_texture_shader GL_NV_texture_shader2 GL_NV_texture_shader3 GL_NV_transform_feedback GL_NV_transform_feedback2 GL_NV_uniform_buffer_unified_memory GL_NV_vertex_array_range GL_NV_vertex_array_range2 GL_NV_vertex_attrib_integer_64bit GL_NV_vertex_buffer_unified_memory GL_NV_vertex_program GL_NV_vertex_program1_1 GL_NV_vertex_program2 GL_NV_vertex_program2_option GL_NV_vertex_program3 GL_NV_viewport_array2 GL_NV_viewport_swizzle GL_NVX_conditional_render GL_NVX_gpu_memory_info GL_NVX_multigpu_info GL_NVX_nvenc_interop GL_NV_shader_thread_group GL_NV_shader_thread_shuffle GL_KHR_blend_equation_advanced GL_KHR_blend_equation_advanced_coherent GL_OVR_multiview GL_OVR_multiview2 GL_SGIS_generate_mipmap GL_SGIS_texture_lod GL_SGIX_depth_texture GL_SGIX_shadow GL_SUN_slice_accum GL_WIN_swap_hint WGL_EXT_swap_control threaded_avail : 1 aniso_avail : 1 timer_avail : 1 sync/fence : 1 pbo_avail : 1 frameblit_avail : 1 framebuffer_msaa : 1 gpu_shad4_avail : 1 shad_lod_avail : 1 array_tex_avail : 1 texture_rg : 1 tex_float_avail : 1 seamless_avail : 1 drawbuf2_avail : 1 packed_stencil : 1 framebuffer_srgb : 1 copy_buf_avail : 1 ubo_avail : 1 gshader_avail : 1 base_vert_avail : 1 tex_multi_avail : 1 depth_float_avail: 1 instance_avail : 1 tess_avail : 1 viewport_arr : 1 tex_storage : 1 buf_storage : 1 multi_bind : 1 clip_control : 1 program binary : 1 debug_avail : 1 max tex units : 32 (32/8) max iso filtering: 16.000000 max samples : 32 max texture size : 16384 (hardware limit) max point size : 189.875000 idx in vram : 1 GLSL Version :4.60 NVIDIA/460 (16/4096/124/192/32/4096/4096/65536) This video card is: DX11+ VR Ready (found glMapBufferRange) CPU count : 12 OpenAL loaded: Resources/dlls/64/openal32.dll OpenAL devices: OpenAL Soft Open AL default device:OpenAL Soft OpenAL vendor : OpenAL Community OpenAL renderer : OpenAL Soft OpenAL version : 1.1 OpenAL hardware : OpenAL Soft OpenAL extensions: ALC_ENUMERATE_ALL_EXT ALC_ENUMERATION_EXT ALC_EXT_CAPTURE ALC_EXT_DEDICATED ALC_EXT_disconnect ALC_EXT_EFX ALC_EXT_thread_local_context ALC_SOFTX_device_clock ALC_SOFTX_HRTF ALC_SOFT_loopback ALC_SOFTX_midi_interface ALC_SOFT_pause_device AL_EXT_ALAW AL_EXT_DOUBLE AL_EXT_EXPONENT_DISTANCE AL_EXT_FLOAT32 AL_EXT_IMA4 AL_EXT_LINEAR_DISTANCE AL_EXT_MCFORMATS AL_EXT_MULAW AL_EXT_MULAW_MCFORMATS AL_EXT_OFFSET AL_EXT_source_distance_model AL_LOKI_quadriphonic AL_SOFT_block_alignment AL_SOFT_buffer_samples AL_SOFT_buffer_sub_data AL_SOFT_deferred_updates AL_SOFT_direct_channels AL_SOFT_loop_points AL_SOFT_MSADPCM AL_SOFT_source_latency AL_SOFT_source_length Fetching plugins for C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/PluginAdmin/64/win.xpl (xpsdk.examples.pluginadmin). Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/AutoGate/64/win.xpl (Marginal.AutoGate). CustomSBDatarefs004 - parsing setup file CustomSBDatarefs004 - parsing successful Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/CustomSBDatarefs004/64/win.xpl (sandybarbour.projects.customsbdatarefs004). Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/FlyInsideXP/64/win.xpl (flyinside.xpintegrator). G64: 0.000: Gizmo64 Script Engine v18.02.27.0200 Windows G64: 0.000: built: Feb 27 2018, 02:00:08 Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/Gizmo64.plugin/64/win.xpl (gizmo.x-plugins.com). Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/HangarOps/64/win.xpl (BlueSideUpBob.STMA_HangarOps). Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/PilotEdge/64/win.xpl (com.pilotedge.plugin.xplane). PythonInterfaceVersionNumber = 2.73.06 Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/PythonInterface/64/win.xpl (sandybarbour.projects.pythoninterface). [SKUNKCRAFTS UPDATER INFO]: Starting X-Plane SASL plugin v3.4.2+3b8dca9 [Commercial Edition] [SKUNKCRAFTS UPDATER INFO]: SkunkCrafts Updater | Global | Start Full Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/SkunkCraftsUpdater/64/win.xpl (1-sim SkunkCrafts Updater). G64: 0.390: Boot Count: 1 G64: 0.733: Plugin Enabled. [SKUNKCRAFTS UPDATER INFO]: Enabling... [SKUNKCRAFTS UPDATER INFO]: Sound output device: OpenAL Soft [SKUNKCRAFTS UPDATER INFO]: Sound input device: OpenAL Soft on Headset Microphone (Razer Kraken 7.1 Chroma) [SKUNKCRAFTS UPDATER INFO]: Initializing... [SKUNKCRAFTS UPDATER INFO]: Path to module: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/SkunkCraftsUpdater/data/modules/main.secured [SKUNKCRAFTS UPDATER INFO]: Loading module... [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading module [SKUNKCRAFTS UPDATER INFO]: [LUA]: including helpers.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Resources/plugins/SkunkCraftsUpdater/data/modules/Custom Module/helpers [SKUNKCRAFTS UPDATER INFO]: [LUA]: including math.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including colors.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including string.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including preferences.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including properties.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including datarefs.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_queue.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_listpart.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_image.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_command.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_font.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_spy.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including utils.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: including class_addon.lua [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" Version 2.2 [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" set opt:audio_feedback to 1 [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" set opt:audio_startup to 1 [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" set opt:run_at_startup to 1 [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" set opt:scan_sceneries to 0 [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" set opt:sound_gain to 150 [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" set opt:verbose_logging to 0 [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading tabs [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading report [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading list [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading button [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading dummy [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading detail [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading checkbox [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading options [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading label [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading radioarray [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading radiobox [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading slider [SKUNKCRAFTS UPDATER INFO]: [LUA]: loading about [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" Remote version of updater: 2.2 [SKUNKCRAFTS UPDATER INFO]: Loading cursors... [SKUNKCRAFTS UPDATER INFO]: Module loaded I found the following scenery packages (prioritized in this order): 0 Custom Scenery/X-Plane Landmarks - Dubai/ 1 Custom Scenery/0L7 Jean/ 2 Custom Scenery/0O2 Baker/ 3 Custom Scenery/3L2 Sky Ranch/ 4 Custom Scenery/Aerosoft - EGBB Birmingham/ 5 Custom Scenery/Aerosoft - EGKK London-Gatwick/ 6 Custom Scenery/Aerosoft - EGPF Glasgow/ 7 Custom Scenery/Aerosoft - EGSS London-Stansted/ 8 Custom Scenery/Aerosoft - LEMD Madrid/ 9 Custom Scenery/Aerosoft - LFPG Paris CDG/ 10 Custom Scenery/flags_of_USA_states/ 11 Custom Scenery/KAVX - Catalina Airport/ 12 Custom Scenery/KGCN Grand Canyon NP/ 13 Custom Scenery/KIFP Bullhead City/ 14 Custom Scenery/KLAS-GLITTER_GULCH 1.5/ 15 Custom Scenery/KTRM - Jacqueline Cochran Regional Airport/ 16 Custom Scenery/LFPG Paris - Charles de Gaulle/ 17 Custom Scenery/MisterX_Library/ 18 Custom Scenery/OpenSceneryX/ 19 Custom Scenery/Aerosoft - EBBR Brussels/ 20 Custom Scenery/Aerosoft - EDDF Frankfurt/ 21 Custom Scenery/Aerosoft - EDLP Paderborn-Lippstadt/ 22 Custom Scenery/Aerosoft - EGLL Heathrow/ 23 Custom Scenery/Aerosoft - LFMN Nice Cote d Azur X/ 24 Custom Scenery/Aerosoft - LFPO Paris Orly/ 25 Custom Scenery/Aerosoft - LPFR Faro/ 26 Custom Scenery/Aerosoft - LSGG Genf/ 27 Custom Scenery/Global Airports/ 28 Custom Scenery/KLAS Las Vegas Mc Carran/ 29 Custom Scenery/KSEA Demo Area/ 30 Custom Scenery/LOWI Demo Area/ 31 Custom Scenery/X-Plane Landmarks - Chicago/ 32 Custom Scenery/X-Plane Landmarks - Las Vegas/ 33 Custom Scenery/X-Plane Landmarks - London/ 34 Custom Scenery/X-Plane Landmarks - Sydney/ 35 Global Scenery/X-Plane 11 Demo Areas/ 36 Global Scenery/X-Plane 11 Global Scenery/ 37 Resources/default scenery/1000 autogen/ 38 Resources/default scenery/1000 decals/ 39 Resources/default scenery/1000 forests/ 40 Resources/default scenery/1000 roads/ 41 Resources/default scenery/1000 urban terrain/ 42 Resources/default scenery/1000 world terrain/ 43 Resources/default scenery/700 roads/ 44 Resources/default scenery/900 beaches/ 45 Resources/default scenery/900 europe objects/ 46 Resources/default scenery/900 forests/ 47 Resources/default scenery/900 roads/ 48 Resources/default scenery/900 us objects/ 49 Resources/default scenery/900 world object placeholders/ 50 Resources/default scenery/airport decals/ 51 Resources/default scenery/airport scenery/ 52 Resources/default scenery/default apt dat/ 53 Resources/default scenery/default atc dat/ 54 Resources/default scenery/sim objects/ 55 Resources/tutorials/tutorials_Alaska/ 56 Resources/tutorials/tutorials_Hawaii/ 57 Resources/tutorials/tutorials_Seattle/ 0:00:00.000 D/HID: HID Bridge Running Fetching plugins for Custom Scenery/KAVX - Catalina Airport/plugins Loaded: Custom Scenery/KAVX - Catalina Airport/plugins/GroundTraffic/64/win.xpl (Marginal.GroundTraffic.KAVX - Catalina Airport). Fetching plugins for Custom Scenery/KLAS-GLITTER_GULCH 1.5/plugins CustomSBDatarefs004 - parsing setup file CustomSBDatarefs004 - parsing successful Warning: Custom Scenery/KLAS-GLITTER_GULCH 1.5/plugins/CustomSBDatarefs004/64/win.xpl Signature: sandybarbour.projects.customsbdatarefs004. (A version of this plugin is already loaded.) Custom Scenery/KLAS-GLITTER_GULCH 1.5/plugins/CustomSBDatarefs004/win.xpl : Error Code = 193 : %1 is not a valid Win32 application. Loaded: Custom Scenery/KLAS-GLITTER_GULCH 1.5/plugins/GroundTraffic/64/win.xpl (Marginal.GroundTraffic.KLAS-GLITTER_GULCH 1.5). Initializing off screen memory. Initializing off screen memory complete. 0:00:18.606 I/REN: Setting texture resolution read from prefs as 4 0:00:18.606 I/REN: Setting |draw view indicator read from prefs as 1. 0:00:18.606 I/REN: Setting |dim under high G-load or hypoxia read from prefs as 1. 0:00:18.606 I/REN: Setting |draw hi-res planet textures from orbit read from prefs as 1. 0:00:18.606 I/REN: Setting |draw forest fires and balloons read from prefs as 0. 0:00:18.606 I/REN: Setting |draw birds and deer in nice weather read from prefs as 1. 0:00:18.606 I/REN: Setting |draw aircraft carriers and frigates read from prefs as 1. 0:00:18.606 I/REN: Setting |draw Aurora Borealis read from prefs as 0. 0:00:18.606 I/REN: Setting |runways follow terrain contours read from prefs as 0. 0:00:18.606 I/REN: Setting |3-d drawing detail read from prefs as 3. 0:00:18.606 I/REN: Setting |water reflection detail read from prefs as 0. 0:00:18.606 I/REN: Setting |draw parked aircraft at airports read from prefs as 1. 0:00:18.606 I/REN: Setting compress textures to save VRAM read from prefs as 1. 0:00:18.606 I/REN: Setting |Effects read from prefs as 4. 0:00:18.606 I/REN: Setting |3-d shadows read from prefs as 0. 0:00:18.606 I/REN: Setting |HDR anti-aliasing read from prefs as 0. 0:00:18.606 I/REN: Setting |draw volumetric fog read from prefs as 1. 0:00:18.606 I/REN: Setting |draw per pixel lighting read from prefs as 1. 0:00:18.606 I/REN: Setting |atmospheric scattering read from prefs as 1. 0:00:19.033 I/NET: OS has RFC 4941 Privacy Extensions for IPv6 *enabled*, using TEMPORARY IPv6 addresses 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 192.168.1.167 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:15d:b4df:97ed:d551 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:715c:25d:40bc:f7bd 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:8987:11b1:c463:1c0a 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:9589:718:8c9c:28f0 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:a535:57b:6ff4:7d52 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:bd56:4459:d246:306f 0:00:19.033 I/NET: Ethernet (Intel(R) Ethernet Connection (2) I218-LM) 2601:647:4000:76:e870:2efe:395e:28cf 0:00:19.461 D/HID: Hardware ADDED(068e, c011) - Control Manager ID #00 at path \\?\hid#vid_068e&pid_c011#3&2f3f2f1c&1&0000#{4d1e55b2-f16f-11cf-88cb-001111000030} 0:00:19.461 I/JOY: REGISTER Joystick device: Control Manager ID #00 - VID:1678PID:49169, this is a familiar device. 0:00:19.461 D/HID: AXIS Added Index: 0 (HID_Page_GenericDesktop/HID_Usage_GD_X) for handle: 000001EE4F956690. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 1 (HID_Page_GenericDesktop/HID_Usage_GD_Y) for handle: 000001EE4F956690. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 2 (HID_Page_GenericDesktop/HID_Usage_GD_Z) for handle: 000001EE4F956690. Min/Max 0 - 255 0:00:19.461 D/HID: END HARDWARE 0:00:19.461 D/HID: Hardware ADDED(068e, c012) - Control Manager ID #00 at path \\?\hid#vid_068e&pid_c012#3&6d48153&0&0000#{4d1e55b2-f16f-11cf-88cb-001111000030} 0:00:19.461 I/JOY: REGISTER Joystick device: Control Manager ID #00 - VID:1678PID:49170, this is a familiar device. 0:00:19.461 D/HID: BUTTON Added Index: 0 (HID_Page_Button/01) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 1 (HID_Page_Button/02) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 2 (HID_Page_Button/03) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 3 (HID_Page_Button/04) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 4 (HID_Page_Button/05) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 5 (HID_Page_Button/06) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 6 (HID_Page_Button/07) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 7 (HID_Page_Button/08) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 8 (HID_Page_Button/09) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 9 (HID_Page_Button/0a) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 10 (HID_Page_Button/0b) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 11 (HID_Page_Button/0c) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 12 (HID_Page_Button/0d) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 13 (HID_Page_Button/0e) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 14 (HID_Page_Button/0f) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 15 (HID_Page_Button/10) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 16 (HID_Page_Button/11) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 17 (HID_Page_Button/12) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 18 (HID_Page_Button/13) for handle: 000001EE4F956150 0:00:19.461 D/HID: BUTTON Added Index: 19 (HID_Page_Button/14) for handle: 000001EE4F956150 0:00:19.461 D/HID: AXIS Added Index: 0 (HID_Page_GenericDesktop/HID_Usage_GD_X) for handle: 000001EE4F956150. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 1 (HID_Page_GenericDesktop/HID_Usage_GD_Y) for handle: 000001EE4F956150. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 2 (HID_Page_GenericDesktop/HID_Usage_GD_Z) for handle: 000001EE4F956150. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 3 (HID_Page_GenericDesktop/HID_Usage_GD_Rz) for handle: 000001EE4F956150. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 4 (HID_Page_GenericDesktop/HID_Usage_GD_Ry) for handle: 000001EE4F956150. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 5 (HID_Page_GenericDesktop/HID_Usage_GD_Rx) for handle: 000001EE4F956150. Min/Max 1 - 255 0:00:19.461 D/HID: END HARDWARE 0:00:19.461 E/JOY: UNCALIBRATED AXIS DETECTED: Control Manager ID #00 (VID:1678PID:49170) Axis #3, assigned to joy_use_none, has an expected min/max range of [0.000000, 255.000000], but is calibrated for [0.286275, 0.921569]. You should recalibrate the device in the Settings menu. 0:00:19.461 E/JOY: UNCALIBRATED AXIS DETECTED: Control Manager ID #00 (VID:1678PID:49170) Axis #4, assigned to joy_use_none, has an expected min/max range of [0.000000, 255.000000], but is calibrated for [0.254902, 0.933333]. You should recalibrate the device in the Settings menu. 0:00:19.461 D/HID: Hardware ADDED(068e, c010) - Control Manager ID #00 at path \\?\hid#vid_068e&pid_c010#3&1c0f12de&1&0000#{4d1e55b2-f16f-11cf-88cb-001111000030} 0:00:19.461 I/JOY: REGISTER Joystick device: Control Manager ID #00 - VID:1678PID:49168, this is a familiar device. 0:00:19.461 D/HID: BUTTON Added Index: 0 (HID_Page_Button/01) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 1 (HID_Page_Button/02) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 2 (HID_Page_Button/03) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 3 (HID_Page_Button/04) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 4 (HID_Page_Button/05) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 5 (HID_Page_Button/06) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 6 (HID_Page_Button/07) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 7 (HID_Page_Button/08) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 8 (HID_Page_Button/09) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 9 (HID_Page_Button/0a) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 10 (HID_Page_Button/0b) for handle: 000001EE4F956090 0:00:19.461 D/HID: BUTTON Added Index: 11 (HID_Page_Button/0c) for handle: 000001EE4F956090 0:00:19.461 D/HID: AXIS Added Index: 0 (HID_Page_GenericDesktop/HID_Usage_GD_X) for handle: 000001EE4F956090. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 1 (HID_Page_GenericDesktop/HID_Usage_GD_Y) for handle: 000001EE4F956090. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 2 (HID_Page_GenericDesktop/HID_Usage_GD_Z) for handle: 000001EE4F956090. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 3 (HID_Page_GenericDesktop/HID_Usage_GD_Rz) for handle: 000001EE4F956090. Min/Max 0 - 255 0:00:19.461 D/HID: AXIS Added Index: 4 (HID_Page_GenericDesktop/HID_Usage_GD_Ry) for handle: 000001EE4F956090. Min/Max 0 - 255 0:00:19.461 D/HID: END HARDWARE 0:00:19.461 D/STM: Transitioning from state state_InitialLoading to state_QuickFlightWizard 0:00:19.461 I/WIN: Showing subscreen Main Menu 0:00:19.461 I/OVR: User is entering the holodeck. 0:00:19.461 I/WIN: Opened window Primary V11 UI 0:00:23.076 I/WIN: Showing subscreen Freeflight 0:00:37.372 W/JOY: Warning about uncalibrated joystick(s): 0:00:37.372 W/JOY: Control Manager ID #00 0:00:37.372 I/JOY: All known joystick(s): 0:00:37.372 W/JOY: Control Manager ID #00 (000001EE4F956090) 0:00:37.372 W/JOY: Control Manager ID #00 (000001EE4F956150) 0:00:37.372 W/JOY: Control Manager ID #00 (000001EE4F956690) 0:00:40.978 I/WIN: Showing subscreen Settings 0:00:44.696 I/FCG: Applying changes from Flight Configuration screen 0:00:44.696 I/FCG: Starting new flight in Aircraft/Laminar Research/Cessna 172SP/Cessna_172SP_G1000.acf at Ramp 0:00:44.769 W/ATC: Found overlapping airports with overlapping controller freqs at KCEA and KIAB! 0:00:44.769 W/ATC: Found overlapping airports with overlapping controller freqs at WSAC and WSSS! 0:00:44.769 W/ATC: Found overlapping airports with overlapping controller freqs at RJCJ and RJCC! 0:00:44.769 W/ATC: Found overlapping airports with overlapping controller freqs at WASS and WA1A! 0:00:44.769 W/ATC: Found overlapping airports with overlapping controller freqs at KNSE and KNDZ! 0:00:44.769 W/ATC: Rewrote KAHN atc_ControllerRole_Del's freq: 133950 to 118000 for being in conflict! 0:00:44.769 W/ATC: Rewrote KCEA atc_ControllerRole_Twr's freq: 127250 to 118025 for being in conflict! 0:00:44.769 W/ATC: Rewrote WSAC atc_ControllerRole_Twr's freq: 118250 to 118000 for being in conflict! 0:00:44.769 W/ATC: Rewrote RJCJ atc_ControllerRole_Del's freq: 121900 to 118000 for being in conflict! 0:00:44.769 W/ATC: Rewrote WASS atc_ControllerRole_Twr's freq: 122400 to 118000 for being in conflict! 0:00:44.769 W/ATC: Rewrote KNSE atc_ControllerRole_Twr's freq: 121400 to 118000 for being in conflict! 0:00:47.408 D/STM: Transitioning from state state_QuickFlightWizard to state_NormalLoading 0:00:47.408 I/ACF: Loading airplane number 0 with Aircraft/Laminar Research/Cessna 172SP/Cessna_172SP_G1000.acf G64: 43.572: Aircraft File: (Aircraft/Laminar Research/Cessna 172SP/Cessna_172SP_G1000.acf) G64: 43.575: MD5 Hash: (3a501697a7d3af0d9f2d8efdd5dd8057) G64: 43.575: Description: (Cessna 172 SP Skyhawk - 180HP - G1000) G64: 43.575: Author: (Laminar Research - dmax3d.com) G64: 43.576: Plugin Mode: Community Edition G64: 43.582: Boot Count: 2 G64: 43.892: Run: OnKickStart.. Fetching plugins for C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Aircraft/Laminar Research/Cessna 172SP/plugins Loaded: C:\Program Files (x86)/Steam/steamapps/common/X-Plane 11/Aircraft/Laminar Research/Cessna 172SP/plugins/xlua/64/win.xpl (com.x-plane.xlua.1.0.0r1). 0:00:48.445 I/SND: This aircraft is using FMOD in PARTIAL mode. G64: 44.643: Livery Changed:(full_white) Folder:(Aircraft/Laminar Research/Cessna 172SP/liveries/full_white/) =====Resources/shaders/hdr.glsl===== (prefix: #define APL 0 #define IBM 1 #define LIN 0 #define MOBILE 0 #define HAS_SRGB 1 #define HAS_SHADER_LOD 1 #define HAS_UBO 1 #define HAS_ARRAY_TEX 1 #define HAS_TEX_RG 1 #define HAS_TESS 1 #define HAS_AMD_SHADER_VIEWPORT_LAYER_ARRAY 1 #define HAS_GSHADER 0 #define MODE 0 #define TONE_MAP 1 end prefix) 0(183) : error C0000: syntax error, unexpected '(', expecting ';' at token "(" =====Resources/shaders/hdr.glsl+Resources/shaders/hdr.glsl===== (prefix: #define APL 0 #define IBM 1 #define LIN 0 #define MOBILE 0 #define HAS_SRGB 1 #define HAS_SHADER_LOD 1 #define HAS_UBO 1 #define HAS_ARRAY_TEX 1 #define HAS_TEX_RG 1 #define HAS_TESS 1 #define HAS_AMD_SHADER_VIEWPORT_LAYER_ARRAY 1 #define HAS_GSHADER 0 #define MODE 0 #define TONE_MAP 1 end prefix) Fragment info ------------- 0(183) : error C0000: syntax error, unexpected '(', expecting ';' at token "(" (0) : error C2003: incompatible options for link SHADER SOURCE FOR VERTEX SHADER: Resources/shaders/hdr.glsl #version 150 compatibility #extension GL_AMD_vertex_shader_viewport_index : enable uniform int uViewportTarget; uniform mat4 uStereoOffsets[2]; #define VSHADER 1 #define GSHADER 0 #define FSHADER 0 #define TCSHADER 0 #define TESHADER 0 #define APL 0 #define IBM 1 #define LIN 0 #define MOBILE 0 #define HAS_SRGB 1 #define HAS_SHADER_LOD 1 #define HAS_UBO 1 #define HAS_ARRAY_TEX 1 #define HAS_TEX_RG 1 #define HAS_TESS 1 #define HAS_AMD_SHADER_VIEWPORT_LAYER_ARRAY 1 #define HAS_GSHADER 0 #define MODE 0 #define TONE_MAP 1 #define USE_UBOS 1 #ifndef HAS_CORE #define HAS_CORE 0 #endif #if __VERSION__ >= 130 // For everything MODERN we support, use the "new" generic texture function names. texture() has no macro // arguments because the bias parameter is optional. We define the old GPU shader4 specific types to generic // because we can't macro the other way. // This case covers GLSL 130 (various Intel GPUs), GLSL 150 + arb compatibility (modern red and green team // drivers on Win/Lin) and GLSL 150 ("core" profile, still experimental) #define texture2D texture #define texture2DArray texture #define texture2DProj(a,b) textureProj(a,b) #define texture2DGrad(a,b,c,d) textureGrad(a,b,c,d) #define texture2DLod(a,b,c) textureLod(a,b,c) #define texture2DArrayLod(a,b,c) textureLod(a,b,c) #define shadow2D(a,b) vec4(texture(a,b)) #define shadow2DArray(a,b) vec4(texture(a,b)) #define shadow2DGrad(a,b,c,d) vec4(textureGrad(a,b,c,d)) #define shadow2DArrayGrad(a,b,c,d) vec4(textureGrad(a,b,c,d)) #elif defined(GL_EXT_gpu_shader4) // On OS X legacy profile we're stuck at GLSL 120 + a ton of extensions. Grab GPU_shader4 - if we don't have // that we are missing a TON of stuff we need. Then cope with the quirk of uint missing. // Note that for all practical purposes we expect to ONLY hit this case on OS X. #extension GL_EXT_gpu_shader4 : enable #define uint unsigned int #else // Back in the day we supported even OLDER skankier stuff, but no need now - every supported GPU // we support is at least GPU_shader4 (which matches shader model 4.0, which is in DX10) - because // DX10 is a MANDATORY level of hw for XP11. So we can just pumt out. This is convenient because // GPU_shader4 gets us gradients and texture arrays. #error we require GLSL 130 or GLSL 120 + GPU_shader4 #endif #if HAS_CORE #define texture3D texture #define textureCubeLod textureLod #endif // Opt into a bunch of useful stuff we want. #if HAS_AMD_SHADER_VIEWPORT_LAYER_ARRAY && HAS_UBO #if defined(GL_ARB_shader_viewport_layer_array) #extension GL_ARB_shader_viewport_layer_array : enable #define STD_BROADCAST \ gl_ViewportIndex = gl_InstanceID % u_viewport_count; \ gl_Layer = gl_InstanceID % u_layer_count; #define HAS_BROADCAST 1 #else #extension GL_AMD_vertex_shader_layer : enable #extension GL_AMD_vertex_shader_viewport_index : enable #define STD_BROADCAST \ gl_ViewportIndex = gl_InstanceID % u_viewport_count; \ gl_Layer = gl_InstanceID % u_layer_count; #define HAS_BROADCAST 1 #endif #else #define STD_BROADCAST #define HAS_BROADCAST 0 #endif #if defined(GL_ARB_gpu_shader5) #extension GL_ARB_gpu_shader5 : enable #endif #if HAS_UBO && !HAS_CORE #extension GL_ARB_uniform_buffer_object : enable #endif #if HAS_TESS && !HAS_CORE && !defined(GL_ARB_gpu_shader5) #extension GL_ARB_tessellation_shader : enable #endif #if __VERSION__ >= 130 // If we run GLSL >= 130 we should use in/out instead of varying and attribute // These macros wrap the legacy keywords to their modern equivalent #if VSHADER #define VARYING(t,v) out t v; #define ATTRIBUTE(t,v) in t v; #elif FSHADER #define VARYING(t,v) in t v; #endif #if TCSHADER #define VARYING(t,v) in t v[]; out t v##_tess[]; #elif TESHADER #define VARYING(t,v) in t v##_tess[]; out t v; #endif #else #define VARYING(t,v) varying t v; #if VSHADER #define ATTRIBUTE(t,v) attribute t v; #endif #endif // This is a hack-around for a broken mat3 cast in some version of the GLSL // compiler on OS X. Tragically I didn't note WHICH version has the bug. TODO in // v11: confirm that on our new minimum OS (10.9? 10.10?) this isn't needed. mat3 mat3_cast(mat4 m) { return mat3(m[0].xyz,m[1].xyz,m[2].xyz); } #define mat3(X) mat3_cast(X) #ifndef USE_UBOS #define USE_UBOS 0 #endif #if VSHADER #if !HAS_BROADCAST || !HAS_UBO #define u_modelview_matrix modelview_matrix #define u_mvp_matrix mvp_matrix #define u_projection_matrix projection_matrix #define u_modelview_inverse modelview_inverse #define u_eye (0) #else #define u_modelview_matrix modelview_matrix[gl_InstanceID % u_transform_count] #define u_mvp_matrix mvp_matrix[gl_InstanceID % u_transform_count] #define u_projection_matrix projection_matrix[gl_InstanceID % u_transform_count] #define u_modelview_inverse modelview_inverse[gl_InstanceID % u_transform_count] #define u_eye (gl_InstanceID % u_viewport_count) #endif #endif layout(std140) uniform u_environment_data { vec3 u_fog_rgb; float u_fog_scale; vec3 u_sun_position; vec3 u_sun_specular; vec3 u_sun_ambient; vec3 u_sun_diffuse; float u_sky_gain; vec4 u_od_scale; vec4 u_scatter_gain; vec4 u_tonemap_gain; vec2 fade_dist; }; layout(std140) uniform u_hdr_data { vec4 u_kernel1; vec4 u_kernel2; vec4 u_super_dim; vec2 u_offset; float u_time; }; layout(std140) uniform u_transform { int u_viewport_count; int u_layer_count; int u_transform_count; vec3 u_mv_offset; mat4 modelview_matrix[6]; mat4 modelview_inverse[6]; mat4 projection_matrix[6]; mat4 mvp_matrix[6]; }; #define MODE_COPY 0 #define MODE_GRAY 1 #define MODE_BLUR 2 #define MODE_HEAT 3 #define MODE_FXAA 4 #if MODE==MODE_HEAT #define HAS_FULL_TRANSFORM 1 #else #define HAS_FULL_TRANSFORM 0 #endif // Set this to 1 to get light metering HUD in HDR mode. #define WANT_LIGHTMETER 0 #define FXAA_GREEN_AS_LUMA 1 #if __VERSION__ >= 130 #define FXAA_GLSL_130 1 #else #define FXAA_GLSL_120 1 #endif #if FSHADER && MODE == MODE_FXAA /*============================================================================ NVIDIA FXAA 3.11 by TIMOTHY LOTTES ------------------------------------------------------------------------------ COPYRIGHT (C) 2010, 2011 NVIDIA CORPORATION. ALL RIGHTS RESERVED. ------------------------------------------------------------------------------ TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THIS SOFTWARE IS PROVIDED *AS IS* AND NVIDIA AND ITS SUPPLIERS DISCLAIM ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL NVIDIA OR ITS SUPPLIERS BE LIABLE FOR ANY SPECIAL, INCIDENTAL, INDIRECT, OR CONSEQUENTIAL DAMAGES WHATSOEVER (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF BUSINESS INFORMATION, OR ANY OTHER PECUNIARY LOSS) ARISING OUT OF THE USE OF OR INABILITY TO USE THIS SOFTWARE, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. ------------------------------------------------------------------------------ CHANGE LOG ------------------------------------------------------------------------------ 11/25/13 - BAS - removed __ from macros to comply with GLSL 1.20.8 spec. ------------------------------------------------------------------------------ INTEGRATION CHECKLIST ------------------------------------------------------------------------------ (1.) In the shader source, setup defines for the desired configuration. When providing multiple shaders (for different presets), simply setup the defines differently in multiple files. Example, #define FXAA_PC 1 #define FXAA_HLSL_5 1 #define FXAA_QUALITY_PRESET 12 Or, #define FXAA_360 1 Or, #define FXAA_PS3 1 Etc. (2.) Then include this file, # include "Fxaa3_11.h" (3.) Then call the FXAA pixel shader from within your desired shader. Look at the FXAA Quality FxaaPixelShader() for docs on inputs. As for FXAA 3.11 all inputs for all shaders are the same to enable easy porting between platforms. return FxaaPixelShader(...); (4.) Insure pass prior to FXAA outputs RGBL (see next section). Or use, #define FXAA_GREEN_AS_LUMA 1 (5.) Setup engine to provide the following constants which are used in the FxaaPixelShader() inputs, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir Look at the FXAA Quality FxaaPixelShader() for docs on inputs. (6.) Have FXAA vertex shader run as a full screen triangle, and output "pos" and "fxaaConsolePosPos" such that inputs in the pixel shader provide, // {xy} = center of pixel FxaaFloat2 pos, // {xy__} = upper left of pixel // {__zw} = lower right of pixel FxaaFloat4 fxaaConsolePosPos, (7.) Insure the texture sampler(s) used by FXAA are set to bilinear filtering. ------------------------------------------------------------------------------ INTEGRATION - RGBL AND COLORSPACE ------------------------------------------------------------------------------ FXAA3 requires RGBL as input unless the following is set, #define FXAA_GREEN_AS_LUMA 1 In which case the engine uses green in place of luma, and requires RGB input is in a non-linear colorspace. RGB should be LDR (low dynamic range). Specifically do FXAA after tonemapping. RGB data as returned by a texture fetch can be non-linear, or linear when FXAA_GREEN_AS_LUMA is not set. Note an "sRGB format" texture counts as linear, because the result of a texture fetch is linear data. Regular "RGBA8" textures in the sRGB colorspace are non-linear. If FXAA_GREEN_AS_LUMA is not set, luma must be stored in the alpha channel prior to running FXAA. This luma should be in a perceptual space (could be gamma 2.0). Example pass before FXAA where output is gamma 2.0 encoded, color.rgb = ToneMap(color.rgb); // linear color output color.rgb = sqrt(color.rgb); // gamma 2.0 color output return color; To use FXAA, color.rgb = ToneMap(color.rgb); // linear color output color.rgb = sqrt(color.rgb); // gamma 2.0 color output color.a = dot(color.rgb, FxaaFloat3(0.299, 0.587, 0.114)); // compute luma return color; Another example where output is linear encoded, say for instance writing to an sRGB formated render target, where the render target does the conversion back to sRGB after blending, color.rgb = ToneMap(color.rgb); // linear color output return color; To use FXAA, color.rgb = ToneMap(color.rgb); // linear color output color.a = sqrt(dot(color.rgb, FxaaFloat3(0.299, 0.587, 0.114))); // compute luma return color; Getting luma correct is required for the algorithm to work correctly. ------------------------------------------------------------------------------ BEING LINEARLY CORRECT? ------------------------------------------------------------------------------ Applying FXAA to a framebuffer with linear RGB color will look worse. This is very counter intuitive, but happends to be true in this case. The reason is because dithering artifacts will be more visiable in a linear colorspace. ------------------------------------------------------------------------------ COMPLEX INTEGRATION ------------------------------------------------------------------------------ Q. What if the engine is blending into RGB before wanting to run FXAA? A. In the last opaque pass prior to FXAA, have the pass write out luma into alpha. Then blend into RGB only. FXAA should be able to run ok assuming the blending pass did not any add aliasing. This should be the common case for particles and common blending passes. A. Or use FXAA_GREEN_AS_LUMA. ============================================================================*/ /*============================================================================ INTEGRATION KNOBS ============================================================================*/ // // FXAA_PS3 and FXAA_360 choose the console algorithm (FXAA3 CONSOLE). // FXAA_360_OPT is a prototype for the new optimized 360 version. // // 1 = Use API. // 0 = Don't use API. // /*--------------------------------------------------------------------------*/ #ifndef FXAA_PS3 #define FXAA_PS3 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_360 #define FXAA_360 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_360_OPT #define FXAA_360_OPT 0 #endif /*==========================================================================*/ #ifndef FXAA_PC // // FXAA Quality // The high quality PC algorithm. // #define FXAA_PC 1 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_PC_CONSOLE // // The console algorithm for PC is included // for developers targeting really low spec machines. // Likely better to just run FXAA_PC, and use a really low preset. // #define FXAA_PC_CONSOLE 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_GLSL_120 #define FXAA_GLSL_120 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_GLSL_130 #define FXAA_GLSL_130 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_HLSL_3 #define FXAA_HLSL_3 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_HLSL_4 #define FXAA_HLSL_4 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_HLSL_5 #define FXAA_HLSL_5 0 #endif /*==========================================================================*/ #ifndef FXAA_GREEN_AS_LUMA // // For those using non-linear color, // and either not able to get luma in alpha, or not wanting to, // this enables FXAA to run using green as a proxy for luma. // So with this enabled, no need to pack luma in alpha. // // This will turn off AA on anything which lacks some amount of green. // Pure red and blue or combination of only R and B, will get no AA. // // Might want to lower the settings for both, // fxaaConsoleEdgeThresholdMin // fxaaQualityEdgeThresholdMin // In order to insure AA does not get turned off on colors // which contain a minor amount of green. // // 1 = On. // 0 = Off. // #define FXAA_GREEN_AS_LUMA 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_EARLY_EXIT // // Controls algorithm's early exit path. // On PS3 turning this ON adds 2 cycles to the shader. // On 360 turning this OFF adds 10ths of a millisecond to the shader. // Turning this off on console will result in a more blurry image. // So this defaults to on. // // 1 = On. // 0 = Off. // #define FXAA_EARLY_EXIT 1 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_DISCARD // // Only valid for PC OpenGL currently. // Probably will not work when FXAA_GREEN_AS_LUMA = 1. // // 1 = Use discard on pixels which don't need AA. // For APIs which enable concurrent TEX+ROP from same surface. // 0 = Return unchanged color on pixels which don't need AA. // #define FXAA_DISCARD 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_FAST_PIXEL_OFFSET // // Used for GLSL 120 only. // // 1 = GL API supports fast pixel offsets // 0 = do not use fast pixel offsets // #ifdef GL_EXT_gpu_shader4 #define FXAA_FAST_PIXEL_OFFSET 1 #endif #ifdef GL_NV_gpu_shader5 #define FXAA_FAST_PIXEL_OFFSET 1 #endif #ifdef GL_ARB_gpu_shader5 #define FXAA_FAST_PIXEL_OFFSET 1 #endif #ifndef FXAA_FAST_PIXEL_OFFSET #define FXAA_FAST_PIXEL_OFFSET 0 #endif #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_GATHER4_ALPHA // // 1 = API supports gather4 on alpha channel. // 0 = API does not support gather4 on alpha channel. // #if (FXAA_HLSL_5 == 1) #define FXAA_GATHER4_ALPHA 1 #endif #ifdef GL_ARB_gpu_shader5 #define FXAA_GATHER4_ALPHA 1 #endif #ifdef GL_NV_gpu_shader5 #define FXAA_GATHER4_ALPHA 1 #endif #ifndef FXAA_GATHER4_ALPHA #define FXAA_GATHER4_ALPHA 0 #endif #endif /*============================================================================ FXAA CONSOLE PS3 - TUNING KNOBS ============================================================================*/ #ifndef FXAA_CONSOLE_PS3_EDGE_SHARPNESS // // Consoles the sharpness of edges on PS3 only. // Non-PS3 tuning is done with shader input. // // Due to the PS3 being ALU bound, // there are only two safe values here: 4 and 8. // These options use the shaders ability to a free *|/ by 2|4|8. // // 8.0 is sharper // 4.0 is softer // 2.0 is really soft (good for vector graphics inputs) // #if 1 #define FXAA_CONSOLE_PS3_EDGE_SHARPNESS 8.0 #endif #if 0 #define FXAA_CONSOLE_PS3_EDGE_SHARPNESS 4.0 #endif #if 0 #define FXAA_CONSOLE_PS3_EDGE_SHARPNESS 2.0 #endif #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_CONSOLE_PS3_EDGE_THRESHOLD // // Only effects PS3. // Non-PS3 tuning is done with shader input. // // The minimum amount of local contrast required to apply algorithm. // The console setting has a different mapping than the quality setting. // // This only applies when FXAA_EARLY_EXIT is 1. // // Due to the PS3 being ALU bound, // there are only two safe values here: 0.25 and 0.125. // These options use the shaders ability to a free *|/ by 2|4|8. // // 0.125 leaves less aliasing, but is softer // 0.25 leaves more aliasing, and is sharper // #if 1 #define FXAA_CONSOLE_PS3_EDGE_THRESHOLD 0.125 #else #define FXAA_CONSOLE_PS3_EDGE_THRESHOLD 0.25 #endif #endif /*============================================================================ FXAA QUALITY - TUNING KNOBS ------------------------------------------------------------------------------ NOTE the other tuning knobs are now in the shader function inputs! ============================================================================*/ #ifndef FXAA_QUALITY_PRESET // // Choose the quality preset. // This needs to be compiled into the shader as it effects code. // Best option to include multiple presets is to // in each shader define the preset, then include this file. // // OPTIONS // ----------------------------------------------------------------------- // 10 to 15 - default medium dither (10=fastest, 15=highest quality) // 20 to 29 - less dither, more expensive (20=fastest, 29=highest quality) // 39 - no dither, very expensive // // NOTES // ----------------------------------------------------------------------- // 12 = slightly faster then FXAA 3.9 and higher edge quality (default) // 13 = about same speed as FXAA 3.9 and better than 12 // 23 = closest to FXAA 3.9 visually and performance wise // _ = the lowest digit is directly related to performance // _ = the highest digit is directly related to style // #define FXAA_QUALITY_PRESET 12 #endif /*============================================================================ FXAA QUALITY - PRESETS ============================================================================*/ /*============================================================================ FXAA QUALITY - MEDIUM DITHER PRESETS ============================================================================*/ #if (FXAA_QUALITY_PRESET == 10) #define FXAA_QUALITY_PS 3 #define FXAA_QUALITY_P0 1.5 #define FXAA_QUALITY_P1 3.0 #define FXAA_QUALITY_P2 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 11) #define FXAA_QUALITY_PS 4 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 3.0 #define FXAA_QUALITY_P3 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 12) #define FXAA_QUALITY_PS 5 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 4.0 #define FXAA_QUALITY_P4 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 13) #define FXAA_QUALITY_PS 6 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 4.0 #define FXAA_QUALITY_P5 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 14) #define FXAA_QUALITY_PS 7 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 4.0 #define FXAA_QUALITY_P6 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 15) #define FXAA_QUALITY_PS 8 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 4.0 #define FXAA_QUALITY_P7 12.0 #endif /*============================================================================ FXAA QUALITY - LOW DITHER PRESETS ============================================================================*/ #if (FXAA_QUALITY_PRESET == 20) #define FXAA_QUALITY_PS 3 #define FXAA_QUALITY_P0 1.5 #define FXAA_QUALITY_P1 2.0 #define FXAA_QUALITY_P2 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 21) #define FXAA_QUALITY_PS 4 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 22) #define FXAA_QUALITY_PS 5 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 23) #define FXAA_QUALITY_PS 6 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 24) #define FXAA_QUALITY_PS 7 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 3.0 #define FXAA_QUALITY_P6 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 25) #define FXAA_QUALITY_PS 8 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 4.0 #define FXAA_QUALITY_P7 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 26) #define FXAA_QUALITY_PS 9 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 4.0 #define FXAA_QUALITY_P8 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 27) #define FXAA_QUALITY_PS 10 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 4.0 #define FXAA_QUALITY_P9 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 28) #define FXAA_QUALITY_PS 11 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 2.0 #define FXAA_QUALITY_P9 4.0 #define FXAA_QUALITY_P10 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 29) #define FXAA_QUALITY_PS 12 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 2.0 #define FXAA_QUALITY_P9 2.0 #define FXAA_QUALITY_P10 4.0 #define FXAA_QUALITY_P11 8.0 #endif /*============================================================================ FXAA QUALITY - EXTREME QUALITY ============================================================================*/ #if (FXAA_QUALITY_PRESET == 39) #define FXAA_QUALITY_PS 12 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.0 #define FXAA_QUALITY_P2 1.0 #define FXAA_QUALITY_P3 1.0 #define FXAA_QUALITY_P4 1.0 #define FXAA_QUALITY_P5 1.5 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 2.0 #define FXAA_QUALITY_P9 2.0 #define FXAA_QUALITY_P10 4.0 #define FXAA_QUALITY_P11 8.0 #endif /*============================================================================ API PORTING ============================================================================*/ #if (FXAA_GLSL_120 == 1) || (FXAA_GLSL_130 == 1) #define FxaaBool bool #define FxaaDiscard discard #define FxaaFloat float #define FxaaFloat2 vec2 #define FxaaFloat3 vec3 #define FxaaFloat4 vec4 #define FxaaHalf float #define FxaaHalf2 vec2 #define FxaaHalf3 vec3 #define FxaaHalf4 vec4 #define FxaaInt2 ivec2 #define FxaaSat(x) clamp(x, 0.0, 1.0) #define FxaaTex sampler2D #else #define FxaaBool bool #define FxaaDiscard clip(-1) #define FxaaFloat float #define FxaaFloat2 float2 #define FxaaFloat3 float3 #define FxaaFloat4 float4 #define FxaaHalf half #define FxaaHalf2 half2 #define FxaaHalf3 half3 #define FxaaHalf4 half4 #define FxaaSat(x) saturate(x) #endif /*--------------------------------------------------------------------------*/ #if (FXAA_GLSL_120 == 1) // Requires, // #version 120 // And at least, // #extension GL_EXT_gpu_shader4 : enable // (or set FXAA_FAST_PIXEL_OFFSET 1 to work like DX9) #define FxaaTexTop(t, p) texture2DLod(t, p, 0.0) #if (FXAA_FAST_PIXEL_OFFSET == 1) #define FxaaTexOff(t, p, o, r) texture2DLodOffset(t, p, 0.0, o) #else #define FxaaTexOff(t, p, o, r) texture2DLod(t, p + (o * r), 0.0) #endif #if (FXAA_GATHER4_ALPHA == 1) // use #extension GL_ARB_gpu_shader5 : enable #define FxaaTexAlpha4(t, p) textureGather(t, p, 3) #define FxaaTexOffAlpha4(t, p, o) textureGatherOffset(t, p, o, 3) #define FxaaTexGreen4(t, p) textureGather(t, p, 1) #define FxaaTexOffGreen4(t, p, o) textureGatherOffset(t, p, o, 1) #endif #endif /*--------------------------------------------------------------------------*/ #if (FXAA_GLSL_130 == 1) // Requires "#version 130" or better #define FxaaTexTop(t, p) textureLod(t, p, 0.0) #define FxaaTexOff(t, p, o, r) textureLodOffset(t, p, 0.0, o) #if (FXAA_GATHER4_ALPHA == 1) // use #extension GL_ARB_gpu_shader5 : enable #define FxaaTexAlpha4(t, p) textureGather(t, p, 3) #define FxaaTexOffAlpha4(t, p, o) textureGatherOffset(t, p, o, 3) #define FxaaTexGreen4(t, p) textureGather(t, p, 1) #define FxaaTexOffGreen4(t, p, o) textureGatherOffset(t, p, o, 1) #endif #endif /*--------------------------------------------------------------------------*/ #if (FXAA_HLSL_3 == 1) || (FXAA_360 == 1) || (FXAA_PS3 == 1) #define FxaaInt2 float2 #define FxaaTex sampler2D #define FxaaTexTop(t, p) tex2Dlod(t, float4(p, 0.0, 0.0)) #define FxaaTexOff(t, p, o, r) tex2Dlod(t, float4(p + (o * r), 0, 0)) #endif /*--------------------------------------------------------------------------*/ #if (FXAA_HLSL_4 == 1) #define FxaaInt2 int2 struct FxaaTex { SamplerState smpl; Texture2D tex; }; #define FxaaTexTop(t, p) t.tex.SampleLevel(t.smpl, p, 0.0) #define FxaaTexOff(t, p, o, r) t.tex.SampleLevel(t.smpl, p, 0.0, o) #endif /*--------------------------------------------------------------------------*/ #if (FXAA_HLSL_5 == 1) #define FxaaInt2 int2 struct FxaaTex { SamplerState smpl; Texture2D tex; }; #define FxaaTexTop(t, p) t.tex.SampleLevel(t.smpl, p, 0.0) #define FxaaTexOff(t, p, o, r) t.tex.SampleLevel(t.smpl, p, 0.0, o) #define FxaaTexAlpha4(t, p) t.tex.GatherAlpha(t.smpl, p) #define FxaaTexOffAlpha4(t, p, o) t.tex.GatherAlpha(t.smpl, p, o) #define FxaaTexGreen4(t, p) t.tex.GatherGreen(t.smpl, p) #define FxaaTexOffGreen4(t, p, o) t.tex.GatherGreen(t.smpl, p, o) #endif /*============================================================================ GREEN AS LUMA OPTION SUPPORT FUNCTION ============================================================================*/ #if (FXAA_GREEN_AS_LUMA == 0) FxaaFloat FxaaLuma(FxaaFloat4 rgba) { return rgba.w; } #else FxaaFloat FxaaLuma(FxaaFloat4 rgba) { return rgba.y; } #endif /*============================================================================ FXAA3 QUALITY - PC ============================================================================*/ #if (FXAA_PC == 1) /*--------------------------------------------------------------------------*/ FxaaFloat4 FxaaPixelShader( // // Use noperspective interpolation here (turn off perspective interpolation). // {xy} = center of pixel FxaaFloat2 pos, // // Used only for FXAA Console, and not used on the 360 version. // Use noperspective interpolation here (turn off perspective interpolation). // {xy__} = upper left of pixel // {__zw} = lower right of pixel FxaaFloat4 fxaaConsolePosPos, // // Input color texture. // {rgb_} = color in linear or perceptual color space // if (FXAA_GREEN_AS_LUMA == 0) // {___a} = luma in perceptual color space (not linear) FxaaTex tex, // // Only used on the optimized 360 version of FXAA Console. // For everything but 360, just use the same input here as for "tex". // For 360, same texture, just alias with a 2nd sampler. // This sampler needs to have an exponent bias of -1. FxaaTex fxaaConsole360TexExpBiasNegOne, // // Only used on the optimized 360 version of FXAA Console. // For everything but 360, just use the same input here as for "tex". // For 360, same texture, just alias with a 3nd sampler. // This sampler needs to have an exponent bias of -2. FxaaTex fxaaConsole360TexExpBiasNegTwo, // // Only used on FXAA Quality. // This must be from a constant/uniform. // {x_} = 1.0/screenWidthInPixels // {_y} = 1.0/screenHeightInPixels FxaaFloat2 fxaaQualityRcpFrame, // // Only used on FXAA Console. // This must be from a constant/uniform. // This effects sub-pixel AA quality and inversely sharpness. // Where N ranges between, // N = 0.50 (default) // N = 0.33 (sharper) // {x___} = -N/screenWidthInPixels // {_y__} = -N/screenHeightInPixels // {__z_} = N/screenWidthInPixels // {___w} = N/screenHeightInPixels FxaaFloat4 fxaaConsoleRcpFrameOpt, // // Only used on FXAA Console. // Not used on 360, but used on PS3 and PC. // This must be from a constant/uniform. // {x___} = -2.0/screenWidthInPixels // {_y__} = -2.0/screenHeightInPixels // {__z_} = 2.0/screenWidthInPixels // {___w} = 2.0/screenHeightInPixels FxaaFloat4 fxaaConsoleRcpFrameOpt2, // // Only used on FXAA Console. // Only used on 360 in place of fxaaConsoleRcpFrameOpt2. // This must be from a constant/uniform. // {x___} = 8.0/screenWidthInPixels // {_y__} = 8.0/screenHeightInPixels // {__z_} = -4.0/screenWidthInPixels // {___w} = -4.0/screenHeightInPixels FxaaFloat4 fxaaConsole360RcpFrameOpt2, // // Only used on FXAA Quality. // This used to be the FXAA_QUALITY_SUBPIX define. // It is here now to allow easier tuning. // Choose the amount of sub-pixel aliasing removal. // This can effect sharpness. // 1.00 - upper limit (softer) // 0.75 - default amount of filtering // 0.50 - lower limit (sharper, less sub-pixel aliasing removal) // 0.25 - almost off // 0.00 - completely off FxaaFloat fxaaQualitySubpix, // // Only used on FXAA Quality. // This used to be the FXAA_QUALITY_EDGE_THRESHOLD define. // It is here now to allow easier tuning. // The minimum amount of local contrast required to apply algorithm. // 0.333 - too little (faster) // 0.250 - low quality // 0.166 - default // 0.125 - high quality // 0.063 - overkill (slower) FxaaFloat fxaaQualityEdgeThreshold, // // Only used on FXAA Quality. // This used to be the FXAA_QUALITY_EDGE_THRESHOLD_MIN define. // It is here now to allow easier tuning. // Trims the algorithm from processing darks. // 0.0833 - upper limit (default, the start of visible unfiltered edges) // 0.0625 - high quality (faster) // 0.0312 - visible limit (slower) // Special notes when using FXAA_GREEN_AS_LUMA, // Likely want to set this to zero. // As colors that are mostly not-green // will appear very dark in the green channel! // Tune by looking at mostly non-green content, // then start at zero and increase until aliasing is a problem. FxaaFloat fxaaQualityEdgeThresholdMin, // // Only used on FXAA Console. // This used to be the FXAA_CONSOLE_EDGE_SHARPNESS define. // It is here now to allow easier tuning. // This does not effect PS3, as this needs to be compiled in. // Use FXAA_CONSOLE_PS3_EDGE_SHARPNESS for PS3. // Due to the PS3 being ALU bound, // there are only three safe values here: 2 and 4 and 8. // These options use the shaders ability to a free *|/ by 2|4|8. // For all other platforms can be a non-power of two. // 8.0 is sharper (default!!!) // 4.0 is softer // 2.0 is really soft (good only for vector graphics inputs) FxaaFloat fxaaConsoleEdgeSharpness, // // Only used on FXAA Console. // This used to be the FXAA_CONSOLE_EDGE_THRESHOLD define. // It is here now to allow easier tuning. // This does not effect PS3, as this needs to be compiled in. // Use FXAA_CONSOLE_PS3_EDGE_THRESHOLD for PS3. // Due to the PS3 being ALU bound, // there are only two safe values here: 1/4 and 1/8. // These options use the shaders ability to a free *|/ by 2|4|8. // The console setting has a different mapping than the quality setting. // Other platforms can use other values. // 0.125 leaves less aliasing, but is softer (default!!!) // 0.25 leaves more aliasing, and is sharper FxaaFloat fxaaConsoleEdgeThreshold, // // Only used on FXAA Console. // This used to be the FXAA_CONSOLE_EDGE_THRESHOLD_MIN define. // It is here now to allow easier tuning. // Trims the algorithm from processing darks. // The console setting has a different mapping than the quality setting. // This only applies when FXAA_EARLY_EXIT is 1. // This does not apply to PS3, // PS3 was simplified to avoid more shader instructions. // 0.06 - faster but more aliasing in darks // 0.05 - default // 0.04 - slower and less aliasing in darks // Special notes when using FXAA_GREEN_AS_LUMA, // Likely want to set this to zero. // As colors that are mostly not-green // will appear very dark in the green channel! // Tune by looking at mostly non-green content, // then start at zero and increase until aliasing is a problem. FxaaFloat fxaaConsoleEdgeThresholdMin, // // Extra constants for 360 FXAA Console only. // Use zeros or anything else for other platforms. // These must be in physical constant registers and NOT immedates. // Immedates will result in compiler un-optimizing. // {xyzw} = float4(1.0, -1.0, 0.25, -0.25) FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ FxaaFloat2 posM; posM.x = pos.x; posM.y = pos.y; #if (FXAA_GATHER4_ALPHA == 1) #if (FXAA_DISCARD == 0) FxaaFloat4 rgbyM = FxaaTexTop(tex, posM); #if (FXAA_GREEN_AS_LUMA == 0) #define lumaM rgbyM.w #else #define lumaM rgbyM.y #endif #endif #if (FXAA_GREEN_AS_LUMA == 0) FxaaFloat4 luma4A = FxaaTexAlpha4(tex, posM); FxaaFloat4 luma4B = FxaaTexOffAlpha4(tex, posM, FxaaInt2(-1, -1)); #else FxaaFloat4 luma4A = FxaaTexGreen4(tex, posM); FxaaFloat4 luma4B = FxaaTexOffGreen4(tex, posM, FxaaInt2(-1, -1)); #endif #if (FXAA_DISCARD == 1) #define lumaM luma4A.w #endif #define lumaE luma4A.z #define lumaS luma4A.x #define lumaSE luma4A.y #define lumaNW luma4B.w #define lumaN luma4B.z #define lumaW luma4B.x #else FxaaFloat4 rgbyM = FxaaTexTop(tex, posM); #if (FXAA_GREEN_AS_LUMA == 0) #define lumaM rgbyM.w #else #define lumaM rgbyM.y #endif FxaaFloat lumaS = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 0, 1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 1, 0), fxaaQualityRcpFrame.xy)); FxaaFloat lumaN = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 0,-1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1, 0), fxaaQualityRcpFrame.xy)); #endif /*--------------------------------------------------------------------------*/ FxaaFloat maxSM = max(lumaS, lumaM); FxaaFloat minSM = min(lumaS, lumaM); FxaaFloat maxESM = max(lumaE, maxSM); FxaaFloat minESM = min(lumaE, minSM); FxaaFloat maxWN = max(lumaN, lumaW); FxaaFloat minWN = min(lumaN, lumaW); FxaaFloat rangeMax = max(maxWN, maxESM); FxaaFloat rangeMin = min(minWN, minESM); FxaaFloat rangeMaxScaled = rangeMax * fxaaQualityEdgeThreshold; FxaaFloat range = rangeMax - rangeMin; FxaaFloat rangeMaxClamped = max(fxaaQualityEdgeThresholdMin, rangeMaxScaled); FxaaBool earlyExit = range < rangeMaxClamped; /*--------------------------------------------------------------------------*/ if(earlyExit) #if (FXAA_DISCARD == 1) FxaaDiscard; #else return rgbyM; #endif /*--------------------------------------------------------------------------*/ #if (FXAA_GATHER4_ALPHA == 0) FxaaFloat lumaNW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1,-1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaSE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 1, 1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaNE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 1,-1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaSW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1, 1), fxaaQualityRcpFrame.xy)); #else FxaaFloat lumaNE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(1, -1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaSW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1, 1), fxaaQualityRcpFrame.xy)); #endif /*--------------------------------------------------------------------------*/ FxaaFloat lumaNS = lumaN + lumaS; FxaaFloat lumaWE = lumaW + lumaE; FxaaFloat subpixRcpRange = 1.0/range; FxaaFloat subpixNSWE = lumaNS + lumaWE; FxaaFloat edgeHorz1 = (-2.0 * lumaM) + lumaNS; FxaaFloat edgeVert1 = (-2.0 * lumaM) + lumaWE; /*--------------------------------------------------------------------------*/ FxaaFloat lumaNESE = lumaNE + lumaSE; FxaaFloat lumaNWNE = lumaNW + lumaNE; FxaaFloat edgeHorz2 = (-2.0 * lumaE) + lumaNESE; FxaaFloat edgeVert2 = (-2.0 * lumaN) + lumaNWNE; /*--------------------------------------------------------------------------*/ FxaaFloat lumaNWSW = lumaNW + lumaSW; FxaaFloat lumaSWSE = lumaSW + lumaSE; FxaaFloat edgeHorz4 = (abs(edgeHorz1) * 2.0) + abs(edgeHorz2); FxaaFloat edgeVert4 = (abs(edgeVert1) * 2.0) + abs(edgeVert2); FxaaFloat edgeHorz3 = (-2.0 * lumaW) + lumaNWSW; FxaaFloat edgeVert3 = (-2.0 * lumaS) + lumaSWSE; FxaaFloat edgeHorz = abs(edgeHorz3) + edgeHorz4; FxaaFloat edgeVert = abs(edgeVert3) + edgeVert4; /*--------------------------------------------------------------------------*/ FxaaFloat subpixNWSWNESE = lumaNWSW + lumaNESE; FxaaFloat lengthSign = fxaaQualityRcpFrame.x; FxaaBool horzSpan = edgeHorz >= edgeVert; FxaaFloat subpixA = subpixNSWE * 2.0 + subpixNWSWNESE; /*--------------------------------------------------------------------------*/ if(!horzSpan) lumaN = lumaW; if(!horzSpan) lumaS = lumaE; if(horzSpan) lengthSign = fxaaQualityRcpFrame.y; FxaaFloat subpixB = (subpixA * (1.0/12.0)) - lumaM; /*--------------------------------------------------------------------------*/ FxaaFloat gradientN = lumaN - lumaM; FxaaFloat gradientS = lumaS - lumaM; FxaaFloat lumaNN = lumaN + lumaM; FxaaFloat lumaSS = lumaS + lumaM; FxaaBool pairN = abs(gradientN) >= abs(gradientS); FxaaFloat gradient = max(abs(gradientN), abs(gradientS)); if(pairN) lengthSign = -lengthSign; FxaaFloat subpixC = FxaaSat(abs(subpixB) * subpixRcpRange); /*--------------------------------------------------------------------------*/ FxaaFloat2 posB; posB.x = posM.x; posB.y = posM.y; FxaaFloat2 offNP; offNP.x = (!horzSpan) ? 0.0 : fxaaQualityRcpFrame.x; offNP.y = ( horzSpan) ? 0.0 : fxaaQualityRcpFrame.y; if(!horzSpan) posB.x += lengthSign * 0.5; if( horzSpan) posB.y += lengthSign * 0.5; /*--------------------------------------------------------------------------*/ FxaaFloat2 posN; posN.x = posB.x - offNP.x * FXAA_QUALITY_P0; posN.y = posB.y - offNP.y * FXAA_QUALITY_P0; FxaaFloat2 posP; posP.x = posB.x + offNP.x * FXAA_QUALITY_P0; posP.y = posB.y + offNP.y * FXAA_QUALITY_P0; FxaaFloat subpixD = ((-2.0)*subpixC) + 3.0; FxaaFloat lumaEndN = FxaaLuma(FxaaTexTop(tex, posN)); FxaaFloat subpixE = subpixC * subpixC; FxaaFloat lumaEndP = FxaaLuma(FxaaTexTop(tex, posP)); /*--------------------------------------------------------------------------*/ if(!pairN) lumaNN = lumaSS; FxaaFloat gradientScaled = gradient * 1.0/4.0; FxaaFloat lumaMM = lumaM - lumaNN * 0.5; FxaaFloat subpixF = subpixD * subpixE; FxaaBool lumaMLTZero = lumaMM < 0.0; /*--------------------------------------------------------------------------*/ lumaEndN -= lumaNN * 0.5; lumaEndP -= lumaNN * 0.5; FxaaBool doneN = abs(lumaEndN) >= gradientScaled; FxaaBool doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P1; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P1; FxaaBool doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P1; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P1; /*--------------------------------------------------------------------------*/ if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P2; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P2; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P2; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P2; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 3) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P3; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P3; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P3; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P3; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 4) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P4; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P4; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P4; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P4; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 5) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P5; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P5; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P5; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P5; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 6) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P6; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P6; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P6; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P6; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 7) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P7; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P7; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P7; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P7; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 8) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P8; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P8; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P8; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P8; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 9) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P9; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P9; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P9; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P9; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 10) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P10; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P10; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P10; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P10; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 11) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P11; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P11; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P11; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P11; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 12) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P12; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P12; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P12; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P12; /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } /*--------------------------------------------------------------------------*/ FxaaFloat dstN = posM.x - posN.x; FxaaFloat dstP = posP.x - posM.x; if(!horzSpan) dstN = posM.y - posN.y; if(!horzSpan) dstP = posP.y - posM.y; /*--------------------------------------------------------------------------*/ FxaaBool goodSpanN = (lumaEndN < 0.0) != lumaMLTZero; FxaaFloat spanLength = (dstP + dstN); FxaaBool goodSpanP = (lumaEndP < 0.0) != lumaMLTZero; FxaaFloat spanLengthRcp = 1.0/spanLength; /*--------------------------------------------------------------------------*/ FxaaBool directionN = dstN < dstP; FxaaFloat dst = min(dstN, dstP); FxaaBool goodSpan = directionN ? goodSpanN : goodSpanP; FxaaFloat subpixG = subpixF * subpixF; FxaaFloat pixelOffset = (dst * (-spanLengthRcp)) + 0.5; FxaaFloat subpixH = subpixG * fxaaQualitySubpix; /*--------------------------------------------------------------------------*/ FxaaFloat pixelOffsetGood = goodSpan ? pixelOffset : 0.0; FxaaFloat pixelOffsetSubpix = max(pixelOffsetGood, subpixH); if(!horzSpan) posM.x += pixelOffsetSubpix * lengthSign; if( horzSpan) posM.y += pixelOffsetSubpix * lengthSign; #if (FXAA_DISCARD == 1) return FxaaTexTop(tex, posM); #else return FxaaFloat4(FxaaTexTop(tex, posM).xyz, lumaM); #endif } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - PC VERSION ------------------------------------------------------------------------------ Instead of using this on PC, I'd suggest just using FXAA Quality with #define FXAA_QUALITY_PRESET 10 Or #define FXAA_QUALITY_PRESET 20 Either are higher qualilty and almost as fast as this on modern PC GPUs. ============================================================================*/ #if (FXAA_PC_CONSOLE == 1) /*--------------------------------------------------------------------------*/ FxaaFloat4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ FxaaFloat lumaNw = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.xy)); FxaaFloat lumaSw = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.xw)); FxaaFloat lumaNe = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.zy)); FxaaFloat lumaSe = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.zw)); /*--------------------------------------------------------------------------*/ FxaaFloat4 rgbyM = FxaaTexTop(tex, pos.xy); #if (FXAA_GREEN_AS_LUMA == 0) FxaaFloat lumaM = rgbyM.w; #else FxaaFloat lumaM = rgbyM.y; #endif /*--------------------------------------------------------------------------*/ FxaaFloat lumaMaxNwSw = max(lumaNw, lumaSw); lumaNe += 1.0/384.0; FxaaFloat lumaMinNwSw = min(lumaNw, lumaSw); /*--------------------------------------------------------------------------*/ FxaaFloat lumaMaxNeSe = max(lumaNe, lumaSe); FxaaFloat lumaMinNeSe = min(lumaNe, lumaSe); /*--------------------------------------------------------------------------*/ FxaaFloat lumaMax = max(lumaMaxNeSe, lumaMaxNwSw); FxaaFloat lumaMin = min(lumaMinNeSe, lumaMinNwSw); /*--------------------------------------------------------------------------*/ FxaaFloat lumaMaxScaled = lumaMax * fxaaConsoleEdgeThreshold; /*--------------------------------------------------------------------------*/ FxaaFloat lumaMinM = min(lumaMin, lumaM); FxaaFloat lumaMaxScaledClamped = max(fxaaConsoleEdgeThresholdMin, lumaMaxScaled); FxaaFloat lumaMaxM = max(lumaMax, lumaM); FxaaFloat dirSwMinusNe = lumaSw - lumaNe; FxaaFloat lumaMaxSubMinM = lumaMaxM - lumaMinM; FxaaFloat dirSeMinusNw = lumaSe - lumaNw; if(lumaMaxSubMinM < lumaMaxScaledClamped) return rgbyM; /*--------------------------------------------------------------------------*/ FxaaFloat2 dir; dir.x = dirSwMinusNe + dirSeMinusNw; dir.y = dirSwMinusNe - dirSeMinusNw; /*--------------------------------------------------------------------------*/ FxaaFloat2 dir1 = normalize(dir.xy); FxaaFloat4 rgbyN1 = FxaaTexTop(tex, pos.xy - dir1 * fxaaConsoleRcpFrameOpt.zw); FxaaFloat4 rgbyP1 = FxaaTexTop(tex, pos.xy + dir1 * fxaaConsoleRcpFrameOpt.zw); /*--------------------------------------------------------------------------*/ FxaaFloat dirAbsMinTimesC = min(abs(dir1.x), abs(dir1.y)) * fxaaConsoleEdgeSharpness; FxaaFloat2 dir2 = clamp(dir1.xy / dirAbsMinTimesC, -2.0, 2.0); /*--------------------------------------------------------------------------*/ FxaaFloat4 rgbyN2 = FxaaTexTop(tex, pos.xy - dir2 * fxaaConsoleRcpFrameOpt2.zw); FxaaFloat4 rgbyP2 = FxaaTexTop(tex, pos.xy + dir2 * fxaaConsoleRcpFrameOpt2.zw); /*--------------------------------------------------------------------------*/ FxaaFloat4 rgbyA = rgbyN1 + rgbyP1; FxaaFloat4 rgbyB = ((rgbyN2 + rgbyP2) * 0.25) + (rgbyA * 0.25); /*--------------------------------------------------------------------------*/ #if (FXAA_GREEN_AS_LUMA == 0) FxaaBool twoTap = (rgbyB.w < lumaMin) || (rgbyB.w > lumaMax); #else FxaaBool twoTap = (rgbyB.y < lumaMin) || (rgbyB.y > lumaMax); #endif if(twoTap) rgbyB.xyz = rgbyA.xyz * 0.5; return rgbyB; } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - 360 PIXEL SHADER ------------------------------------------------------------------------------ This optimized version thanks to suggestions from Andy Luedke. Should be fully tex bound in all cases. As of the FXAA 3.11 release, I have still not tested this code, however I fixed a bug which was in both FXAA 3.9 and FXAA 3.10. And note this is replacing the old unoptimized version. If it does not work, please let me know so I can fix it. ============================================================================*/ #if (FXAA_360 == 1) /*--------------------------------------------------------------------------*/ [reduceTempRegUsage(4)] float4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ float4 lumaNwNeSwSe; #if (FXAA_GREEN_AS_LUMA == 0) asm { tfetch2D lumaNwNeSwSe.w___, tex, pos.xy, OffsetX = -0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe._w__, tex, pos.xy, OffsetX = 0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.__w_, tex, pos.xy, OffsetX = -0.5, OffsetY = 0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.___w, tex, pos.xy, OffsetX = 0.5, OffsetY = 0.5, UseComputedLOD=false }; #else asm { tfetch2D lumaNwNeSwSe.y___, tex, pos.xy, OffsetX = -0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe._y__, tex, pos.xy, OffsetX = 0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.__y_, tex, pos.xy, OffsetX = -0.5, OffsetY = 0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.___y, tex, pos.xy, OffsetX = 0.5, OffsetY = 0.5, UseComputedLOD=false }; #endif /*--------------------------------------------------------------------------*/ lumaNwNeSwSe.y += 1.0/384.0; float2 lumaMinTemp = min(lumaNwNeSwSe.xy, lumaNwNeSwSe.zw); float2 lumaMaxTemp = max(lumaNwNeSwSe.xy, lumaNwNeSwSe.zw); float lumaMin = min(lumaMinTemp.x, lumaMinTemp.y); float lumaMax = max(lumaMaxTemp.x, lumaMaxTemp.y); /*--------------------------------------------------------------------------*/ float4 rgbyM = tex2Dlod(tex, float4(pos.xy, 0.0, 0.0)); #if (FXAA_GREEN_AS_LUMA == 0) float lumaMinM = min(lumaMin, rgbyM.w); float lumaMaxM = max(lumaMax, rgbyM.w); #else float lumaMinM = min(lumaMin, rgbyM.y); float lumaMaxM = max(lumaMax, rgbyM.y); #endif if((lumaMaxM - lumaMinM) < max(fxaaConsoleEdgeThresholdMin, lumaMax * fxaaConsoleEdgeThreshold)) return rgbyM; /*--------------------------------------------------------------------------*/ float2 dir; dir.x = dot(lumaNwNeSwSe, fxaaConsole360ConstDir.yyxx); dir.y = dot(lumaNwNeSwSe, fxaaConsole360ConstDir.xyxy); dir = normalize(dir); /*--------------------------------------------------------------------------*/ float4 dir1 = dir.xyxy * fxaaConsoleRcpFrameOpt.xyzw; /*--------------------------------------------------------------------------*/ float4 dir2; float dirAbsMinTimesC = min(abs(dir.x), abs(dir.y)) * fxaaConsoleEdgeSharpness; dir2 = saturate(fxaaConsole360ConstDir.zzww * dir.xyxy / dirAbsMinTimesC + 0.5); dir2 = dir2 * fxaaConsole360RcpFrameOpt2.xyxy + fxaaConsole360RcpFrameOpt2.zwzw; /*--------------------------------------------------------------------------*/ float4 rgbyN1 = tex2Dlod(fxaaConsole360TexExpBiasNegOne, float4(pos.xy + dir1.xy, 0.0, 0.0)); float4 rgbyP1 = tex2Dlod(fxaaConsole360TexExpBiasNegOne, float4(pos.xy + dir1.zw, 0.0, 0.0)); float4 rgbyN2 = tex2Dlod(fxaaConsole360TexExpBiasNegTwo, float4(pos.xy + dir2.xy, 0.0, 0.0)); float4 rgbyP2 = tex2Dlod(fxaaConsole360TexExpBiasNegTwo, float4(pos.xy + dir2.zw, 0.0, 0.0)); /*--------------------------------------------------------------------------*/ float4 rgbyA = rgbyN1 + rgbyP1; float4 rgbyB = rgbyN2 + rgbyP2 + rgbyA * 0.5; /*--------------------------------------------------------------------------*/ float4 rgbyR = ((FxaaLuma(rgbyB) - lumaMax) > 0.0) ? rgbyA : rgbyB; rgbyR = ((FxaaLuma(rgbyB) - lumaMin) > 0.0) ? rgbyR : rgbyA; return rgbyR; } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - OPTIMIZED PS3 PIXEL SHADER (NO EARLY EXIT) ============================================================================== The code below does not exactly match the assembly. I have a feeling that 12 cycles is possible, but was not able to get there. Might have to increase register count to get full performance. Note this shader does not use perspective interpolation. Use the following cgc options, --fenable-bx2 --fastmath --fastprecision --nofloatbindings ------------------------------------------------------------------------------ NVSHADERPERF OUTPUT ------------------------------------------------------------------------------ For reference and to aid in debug, output of NVShaderPerf should match this, Shader to schedule: 0: texpkb h0.w(TRUE), v5.zyxx, #0 2: addh h2.z(TRUE), h0.w, constant(0.001953, 0.000000, 0.000000, 0.000000).x 4: texpkb h0.w(TRUE), v5.xwxx, #0 6: addh h0.z(TRUE), -h2, h0.w 7: texpkb h1.w(TRUE), v5, #0 9: addh h0.x(TRUE), h0.z, -h1.w 10: addh h3.w(TRUE), h0.z, h1 11: texpkb h2.w(TRUE), v5.zwzz, #0 13: addh h0.z(TRUE), h3.w, -h2.w 14: addh h0.x(TRUE), h2.w, h0 15: nrmh h1.xz(TRUE), h0_n 16: minh_m8 h0.x(TRUE), |h1|, |h1.z| 17: maxh h4.w(TRUE), h0, h1 18: divx h2.xy(TRUE), h1_n.xzzw, h0_n 19: movr r1.zw(TRUE), v4.xxxy 20: madr r2.xz(TRUE), -h1, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).zzww, r1.zzww 22: minh h5.w(TRUE), h0, h1 23: texpkb h0(TRUE), r2.xzxx, #0 25: madr r0.zw(TRUE), h1.xzxz, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w), r1 27: maxh h4.x(TRUE), h2.z, h2.w 28: texpkb h1(TRUE), r0.zwzz, #0 30: addh_d2 h1(TRUE), h0, h1 31: madr r0.xy(TRUE), -h2, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 33: texpkb h0(TRUE), r0, #0 35: minh h4.z(TRUE), h2, h2.w 36: fenct TRUE 37: madr r1.xy(TRUE), h2, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 39: texpkb h2(TRUE), r1, #0 41: addh_d2 h0(TRUE), h0, h2 42: maxh h2.w(TRUE), h4, h4.x 43: minh h2.x(TRUE), h5.w, h4.z 44: addh_d2 h0(TRUE), h0, h1 45: slth h2.x(TRUE), h0.w, h2 46: sgth h2.w(TRUE), h0, h2 47: movh h0(TRUE), h0 48: addx.c0 rc(TRUE), h2, h2.w 49: movh h0(c0.NE.x), h1 IPU0 ------ Simplified schedule: -------- Pass | Unit | uOp | PC: Op -----+--------+------+------------------------- 1 | SCT0/1 | mov | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | TEX | txl | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | SCB1 | add | 2: ADDh h2.z, h0.--w-, const.--x-; | | | 2 | SCT0/1 | mov | 4: TXLr h0.w, g[TEX1].xwxx, const.xxxx, TEX0; | TEX | txl | 4: TXLr h0.w, g[TEX1].xwxx, const.xxxx, TEX0; | SCB1 | add | 6: ADDh h0.z,-h2, h0.--w-; | | | 3 | SCT0/1 | mov | 7: TXLr h1.w, g[TEX1], const.xxxx, TEX0; | TEX | txl | 7: TXLr h1.w, g[TEX1], const.xxxx, TEX0; | SCB0 | add | 9: ADDh h0.x, h0.z---,-h1.w---; | SCB1 | add | 10: ADDh h3.w, h0.---z, h1; | | | 4 | SCT0/1 | mov | 11: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | TEX | txl | 11: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | SCB0 | add | 14: ADDh h0.x, h2.w---, h0; | SCB1 | add | 13: ADDh h0.z, h3.--w-,-h2.--w-; | | | 5 | SCT1 | mov | 15: NRMh h1.xz, h0; | SRB | nrm | 15: NRMh h1.xz, h0; | SCB0 | min | 16: MINh*8 h0.x, |h1|, |h1.z---|; | SCB1 | max | 17: MAXh h4.w, h0, h1; | | | 6 | SCT0 | div | 18: DIVx h2.xy, h1.xz--, h0; | SCT1 | mov | 19: MOVr r1.zw, g[TEX0].--xy; | SCB0 | mad | 20: MADr r2.xz,-h1, const.z-w-, r1.z-w-; | SCB1 | min | 22: MINh h5.w, h0, h1; | | | 7 | SCT0/1 | mov | 23: TXLr h0, r2.xzxx, const.xxxx, TEX0; | TEX | txl | 23: TXLr h0, r2.xzxx, const.xxxx, TEX0; | SCB0 | max | 27: MAXh h4.x, h2.z---, h2.w---; | SCB1 | mad | 25: MADr r0.zw, h1.--xz, const, r1; | | | 8 | SCT0/1 | mov | 28: TXLr h1, r0.zwzz, const.xxxx, TEX0; | TEX | txl | 28: TXLr h1, r0.zwzz, const.xxxx, TEX0; | SCB0/1 | add | 30: ADDh/2 h1, h0, h1; | | | 9 | SCT0 | mad | 31: MADr r0.xy,-h2, const.xy--, r1.zw--; | SCT1 | mov | 33: TXLr h0, r0, const.zzzz, TEX0; | TEX | txl | 33: TXLr h0, r0, const.zzzz, TEX0; | SCB1 | min | 35: MINh h4.z, h2, h2.--w-; | | | 10 | SCT0 | mad | 37: MADr r1.xy, h2, const.xy--, r1.zw--; | SCT1 | mov | 39: TXLr h2, r1, const.zzzz, TEX0; | TEX | txl | 39: TXLr h2, r1, const.zzzz, TEX0; | SCB0/1 | add | 41: ADDh/2 h0, h0, h2; | | | 11 | SCT0 | min | 43: MINh h2.x, h5.w---, h4.z---; | SCT1 | max | 42: MAXh h2.w, h4, h4.---x; | SCB0/1 | add | 44: ADDh/2 h0, h0, h1; | | | 12 | SCT0 | set | 45: SLTh h2.x, h0.w---, h2; | SCT1 | set | 46: SGTh h2.w, h0, h2; | SCB0/1 | mul | 47: MOVh h0, h0; | | | 13 | SCT0 | mad | 48: ADDxc0_s rc, h2, h2.w---; | SCB0/1 | mul | 49: MOVh h0(NE0.xxxx), h1; Pass SCT TEX SCB 1: 0% 100% 25% 2: 0% 100% 25% 3: 0% 100% 50% 4: 0% 100% 50% 5: 0% 0% 50% 6: 100% 0% 75% 7: 0% 100% 75% 8: 0% 100% 100% 9: 0% 100% 25% 10: 0% 100% 100% 11: 50% 0% 100% 12: 50% 0% 100% 13: 25% 0% 100% MEAN: 17% 61% 67% Pass SCT0 SCT1 TEX SCB0 SCB1 1: 0% 0% 100% 0% 100% 2: 0% 0% 100% 0% 100% 3: 0% 0% 100% 100% 100% 4: 0% 0% 100% 100% 100% 5: 0% 0% 0% 100% 100% 6: 100% 100% 0% 100% 100% 7: 0% 0% 100% 100% 100% 8: 0% 0% 100% 100% 100% 9: 0% 0% 100% 0% 100% 10: 0% 0% 100% 100% 100% 11: 100% 100% 0% 100% 100% 12: 100% 100% 0% 100% 100% 13: 100% 0% 0% 100% 100% MEAN: 30% 23% 61% 76% 100% Fragment Performance Setup: Driver RSX Compiler, GPU RSX, Flags 0x5 Results 13 cycles, 3 r regs, 923,076,923 pixels/s ============================================================================*/ #if (FXAA_PS3 == 1) && (FXAA_EARLY_EXIT == 0) /*--------------------------------------------------------------------------*/ #pragma regcount 7 #pragma disablepc all #pragma option O3 #pragma option OutColorPrec=fp16 #pragma texformat default RGBA8 /*==========================================================================*/ half4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ // (1) half4 dir; half4 lumaNe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) lumaNe.w += half(1.0/512.0); dir.x = -lumaNe.w; dir.z = -lumaNe.w; #else lumaNe.y += half(1.0/512.0); dir.x = -lumaNe.y; dir.z = -lumaNe.y; #endif /*--------------------------------------------------------------------------*/ // (2) half4 lumaSw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) dir.x += lumaSw.w; dir.z += lumaSw.w; #else dir.x += lumaSw.y; dir.z += lumaSw.y; #endif /*--------------------------------------------------------------------------*/ // (3) half4 lumaNw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) dir.x -= lumaNw.w; dir.z += lumaNw.w; #else dir.x -= lumaNw.y; dir.z += lumaNw.y; #endif /*--------------------------------------------------------------------------*/ // (4) half4 lumaSe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) dir.x += lumaSe.w; dir.z -= lumaSe.w; #else dir.x += lumaSe.y; dir.z -= lumaSe.y; #endif /*--------------------------------------------------------------------------*/ // (5) half4 dir1_pos; dir1_pos.xy = normalize(dir.xyz).xz; half dirAbsMinTimesC = min(abs(dir1_pos.x), abs(dir1_pos.y)) * half(FXAA_CONSOLE_PS3_EDGE_SHARPNESS); /*--------------------------------------------------------------------------*/ // (6) half4 dir2_pos; dir2_pos.xy = clamp(dir1_pos.xy / dirAbsMinTimesC, half(-2.0), half(2.0)); dir1_pos.zw = pos.xy; dir2_pos.zw = pos.xy; half4 temp1N; temp1N.xy = dir1_pos.zw - dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; /*--------------------------------------------------------------------------*/ // (7) temp1N = h4tex2Dlod(tex, half4(temp1N.xy, 0.0, 0.0)); half4 rgby1; rgby1.xy = dir1_pos.zw + dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; /*--------------------------------------------------------------------------*/ // (8) rgby1 = h4tex2Dlod(tex, half4(rgby1.xy, 0.0, 0.0)); rgby1 = (temp1N + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (9) half4 temp2N; temp2N.xy = dir2_pos.zw - dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; temp2N = h4tex2Dlod(tex, half4(temp2N.xy, 0.0, 0.0)); /*--------------------------------------------------------------------------*/ // (10) half4 rgby2; rgby2.xy = dir2_pos.zw + dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; rgby2 = h4tex2Dlod(tex, half4(rgby2.xy, 0.0, 0.0)); rgby2 = (temp2N + rgby2) * 0.5; /*--------------------------------------------------------------------------*/ // (11) // compilier moves these scalar ops up to other cycles #if (FXAA_GREEN_AS_LUMA == 0) half lumaMin = min(min(lumaNw.w, lumaSw.w), min(lumaNe.w, lumaSe.w)); half lumaMax = max(max(lumaNw.w, lumaSw.w), max(lumaNe.w, lumaSe.w)); #else half lumaMin = min(min(lumaNw.y, lumaSw.y), min(lumaNe.y, lumaSe.y)); half lumaMax = max(max(lumaNw.y, lumaSw.y), max(lumaNe.y, lumaSe.y)); #endif rgby2 = (rgby2 + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (12) #if (FXAA_GREEN_AS_LUMA == 0) bool twoTapLt = rgby2.w < lumaMin; bool twoTapGt = rgby2.w > lumaMax; #else bool twoTapLt = rgby2.y < lumaMin; bool twoTapGt = rgby2.y > lumaMax; #endif /*--------------------------------------------------------------------------*/ // (13) if(twoTapLt || twoTapGt) rgby2 = rgby1; /*--------------------------------------------------------------------------*/ return rgby2; } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - OPTIMIZED PS3 PIXEL SHADER (WITH EARLY EXIT) ============================================================================== The code mostly matches the assembly. I have a feeling that 14 cycles is possible, but was not able to get there. Might have to increase register count to get full performance. Note this shader does not use perspective interpolation. Use the following cgc options, --fenable-bx2 --fastmath --fastprecision --nofloatbindings Use of FXAA_GREEN_AS_LUMA currently adds a cycle (16 clks). Will look at fixing this for FXAA 3.12. ------------------------------------------------------------------------------ NVSHADERPERF OUTPUT ------------------------------------------------------------------------------ For reference and to aid in debug, output of NVShaderPerf should match this, Shader to schedule: 0: texpkb h0.w(TRUE), v5.zyxx, #0 2: addh h2.y(TRUE), h0.w, constant(0.001953, 0.000000, 0.000000, 0.000000).x 4: texpkb h1.w(TRUE), v5.xwxx, #0 6: addh h0.x(TRUE), h1.w, -h2.y 7: texpkb h2.w(TRUE), v5.zwzz, #0 9: minh h4.w(TRUE), h2.y, h2 10: maxh h5.x(TRUE), h2.y, h2.w 11: texpkb h0.w(TRUE), v5, #0 13: addh h3.w(TRUE), -h0, h0.x 14: addh h0.x(TRUE), h0.w, h0 15: addh h0.z(TRUE), -h2.w, h0.x 16: addh h0.x(TRUE), h2.w, h3.w 17: minh h5.y(TRUE), h0.w, h1.w 18: nrmh h2.xz(TRUE), h0_n 19: minh_m8 h2.w(TRUE), |h2.x|, |h2.z| 20: divx h4.xy(TRUE), h2_n.xzzw, h2_n.w 21: movr r1.zw(TRUE), v4.xxxy 22: maxh h2.w(TRUE), h0, h1 23: fenct TRUE 24: madr r0.xy(TRUE), -h2.xzzw, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).zwzz, r1.zwzz 26: texpkb h0(TRUE), r0, #0 28: maxh h5.x(TRUE), h2.w, h5 29: minh h5.w(TRUE), h5.y, h4 30: madr r1.xy(TRUE), h2.xzzw, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).zwzz, r1.zwzz 32: texpkb h2(TRUE), r1, #0 34: addh_d2 h2(TRUE), h0, h2 35: texpkb h1(TRUE), v4, #0 37: maxh h5.y(TRUE), h5.x, h1.w 38: minh h4.w(TRUE), h1, h5 39: madr r0.xy(TRUE), -h4, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 41: texpkb h0(TRUE), r0, #0 43: addh_m8 h5.z(TRUE), h5.y, -h4.w 44: madr r2.xy(TRUE), h4, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 46: texpkb h3(TRUE), r2, #0 48: addh_d2 h0(TRUE), h0, h3 49: addh_d2 h3(TRUE), h0, h2 50: movh h0(TRUE), h3 51: slth h3.x(TRUE), h3.w, h5.w 52: sgth h3.w(TRUE), h3, h5.x 53: addx.c0 rc(TRUE), h3.x, h3 54: slth.c0 rc(TRUE), h5.z, h5 55: movh h0(c0.NE.w), h2 56: movh h0(c0.NE.x), h1 IPU0 ------ Simplified schedule: -------- Pass | Unit | uOp | PC: Op -----+--------+------+------------------------- 1 | SCT0/1 | mov | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | TEX | txl | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | SCB0 | add | 2: ADDh h2.y, h0.-w--, const.-x--; | | | 2 | SCT0/1 | mov | 4: TXLr h1.w, g[TEX1].xwxx, const.xxxx, TEX0; | TEX | txl | 4: TXLr h1.w, g[TEX1].xwxx, const.xxxx, TEX0; | SCB0 | add | 6: ADDh h0.x, h1.w---,-h2.y---; | | | 3 | SCT0/1 | mov | 7: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | TEX | txl | 7: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | SCB0 | max | 10: MAXh h5.x, h2.y---, h2.w---; | SCB1 | min | 9: MINh h4.w, h2.---y, h2; | | | 4 | SCT0/1 | mov | 11: TXLr h0.w, g[TEX1], const.xxxx, TEX0; | TEX | txl | 11: TXLr h0.w, g[TEX1], const.xxxx, TEX0; | SCB0 | add | 14: ADDh h0.x, h0.w---, h0; | SCB1 | add | 13: ADDh h3.w,-h0, h0.---x; | | | 5 | SCT0 | mad | 16: ADDh h0.x, h2.w---, h3.w---; | SCT1 | mad | 15: ADDh h0.z,-h2.--w-, h0.--x-; | SCB0 | min | 17: MINh h5.y, h0.-w--, h1.-w--; | | | 6 | SCT1 | mov | 18: NRMh h2.xz, h0; | SRB | nrm | 18: NRMh h2.xz, h0; | SCB1 | min | 19: MINh*8 h2.w, |h2.---x|, |h2.---z|; | | | 7 | SCT0 | div | 20: DIVx h4.xy, h2.xz--, h2.ww--; | SCT1 | mov | 21: MOVr r1.zw, g[TEX0].--xy; | SCB1 | max | 22: MAXh h2.w, h0, h1; | | | 8 | SCT0 | mad | 24: MADr r0.xy,-h2.xz--, const.zw--, r1.zw--; | SCT1 | mov | 26: TXLr h0, r0, const.xxxx, TEX0; | TEX | txl | 26: TXLr h0, r0, const.xxxx, TEX0; | SCB0 | max | 28: MAXh h5.x, h2.w---, h5; | SCB1 | min | 29: MINh h5.w, h5.---y, h4; | | | 9 | SCT0 | mad | 30: MADr r1.xy, h2.xz--, const.zw--, r1.zw--; | SCT1 | mov | 32: TXLr h2, r1, const.xxxx, TEX0; | TEX | txl | 32: TXLr h2, r1, const.xxxx, TEX0; | SCB0/1 | add | 34: ADDh/2 h2, h0, h2; | | | 10 | SCT0/1 | mov | 35: TXLr h1, g[TEX0], const.xxxx, TEX0; | TEX | txl | 35: TXLr h1, g[TEX0], const.xxxx, TEX0; | SCB0 | max | 37: MAXh h5.y, h5.-x--, h1.-w--; | SCB1 | min | 38: MINh h4.w, h1, h5; | | | 11 | SCT0 | mad | 39: MADr r0.xy,-h4, const.xy--, r1.zw--; | SCT1 | mov | 41: TXLr h0, r0, const.zzzz, TEX0; | TEX | txl | 41: TXLr h0, r0, const.zzzz, TEX0; | SCB0 | mad | 44: MADr r2.xy, h4, const.xy--, r1.zw--; | SCB1 | add | 43: ADDh*8 h5.z, h5.--y-,-h4.--w-; | | | 12 | SCT0/1 | mov | 46: TXLr h3, r2, const.xxxx, TEX0; | TEX | txl | 46: TXLr h3, r2, const.xxxx, TEX0; | SCB0/1 | add | 48: ADDh/2 h0, h0, h3; | | | 13 | SCT0/1 | mad | 49: ADDh/2 h3, h0, h2; | SCB0/1 | mul | 50: MOVh h0, h3; | | | 14 | SCT0 | set | 51: SLTh h3.x, h3.w---, h5.w---; | SCT1 | set | 52: SGTh h3.w, h3, h5.---x; | SCB0 | set | 54: SLThc0 rc, h5.z---, h5; | SCB1 | add | 53: ADDxc0_s rc, h3.---x, h3; | | | 15 | SCT0/1 | mul | 55: MOVh h0(NE0.wwww), h2; | SCB0/1 | mul | 56: MOVh h0(NE0.xxxx), h1; Pass SCT TEX SCB 1: 0% 100% 25% 2: 0% 100% 25% 3: 0% 100% 50% 4: 0% 100% 50% 5: 50% 0% 25% 6: 0% 0% 25% 7: 100% 0% 25% 8: 0% 100% 50% 9: 0% 100% 100% 10: 0% 100% 50% 11: 0% 100% 75% 12: 0% 100% 100% 13: 100% 0% 100% 14: 50% 0% 50% 15: 100% 0% 100% MEAN: 26% 60% 56% Pass SCT0 SCT1 TEX SCB0 SCB1 1: 0% 0% 100% 100% 0% 2: 0% 0% 100% 100% 0% 3: 0% 0% 100% 100% 100% 4: 0% 0% 100% 100% 100% 5: 100% 100% 0% 100% 0% 6: 0% 0% 0% 0% 100% 7: 100% 100% 0% 0% 100% 8: 0% 0% 100% 100% 100% 9: 0% 0% 100% 100% 100% 10: 0% 0% 100% 100% 100% 11: 0% 0% 100% 100% 100% 12: 0% 0% 100% 100% 100% 13: 100% 100% 0% 100% 100% 14: 100% 100% 0% 100% 100% 15: 100% 100% 0% 100% 100% MEAN: 33% 33% 60% 86% 80% Fragment Performance Setup: Driver RSX Compiler, GPU RSX, Flags 0x5 Results 15 cycles, 3 r regs, 800,000,000 pixels/s ============================================================================*/ #if (FXAA_PS3 == 1) && (FXAA_EARLY_EXIT == 1) /*--------------------------------------------------------------------------*/ #pragma regcount 7 #pragma disablepc all #pragma option O2 #pragma option OutColorPrec=fp16 #pragma texformat default RGBA8 /*==========================================================================*/ half4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ // (1) half4 rgbyNe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaNe = rgbyNe.w + half(1.0/512.0); #else half lumaNe = rgbyNe.y + half(1.0/512.0); #endif /*--------------------------------------------------------------------------*/ // (2) half4 lumaSw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaSwNegNe = lumaSw.w - lumaNe; #else half lumaSwNegNe = lumaSw.y - lumaNe; #endif /*--------------------------------------------------------------------------*/ // (3) half4 lumaNw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaMaxNwSw = max(lumaNw.w, lumaSw.w); half lumaMinNwSw = min(lumaNw.w, lumaSw.w); #else half lumaMaxNwSw = max(lumaNw.y, lumaSw.y); half lumaMinNwSw = min(lumaNw.y, lumaSw.y); #endif /*--------------------------------------------------------------------------*/ // (4) half4 lumaSe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half dirZ = lumaNw.w + lumaSwNegNe; half dirX = -lumaNw.w + lumaSwNegNe; #else half dirZ = lumaNw.y + lumaSwNegNe; half dirX = -lumaNw.y + lumaSwNegNe; #endif /*--------------------------------------------------------------------------*/ // (5) half3 dir; dir.y = 0.0; #if (FXAA_GREEN_AS_LUMA == 0) dir.x = lumaSe.w + dirX; dir.z = -lumaSe.w + dirZ; half lumaMinNeSe = min(lumaNe, lumaSe.w); #else dir.x = lumaSe.y + dirX; dir.z = -lumaSe.y + dirZ; half lumaMinNeSe = min(lumaNe, lumaSe.y); #endif /*--------------------------------------------------------------------------*/ // (6) half4 dir1_pos; dir1_pos.xy = normalize(dir).xz; half dirAbsMinTimes8 = min(abs(dir1_pos.x), abs(dir1_pos.y)) * half(FXAA_CONSOLE_PS3_EDGE_SHARPNESS); /*--------------------------------------------------------------------------*/ // (7) half4 dir2_pos; dir2_pos.xy = clamp(dir1_pos.xy / dirAbsMinTimes8, half(-2.0), half(2.0)); dir1_pos.zw = pos.xy; dir2_pos.zw = pos.xy; #if (FXAA_GREEN_AS_LUMA == 0) half lumaMaxNeSe = max(lumaNe, lumaSe.w); #else half lumaMaxNeSe = max(lumaNe, lumaSe.y); #endif /*--------------------------------------------------------------------------*/ // (8) half4 temp1N; temp1N.xy = dir1_pos.zw - dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; temp1N = h4tex2Dlod(tex, half4(temp1N.xy, 0.0, 0.0)); half lumaMax = max(lumaMaxNwSw, lumaMaxNeSe); half lumaMin = min(lumaMinNwSw, lumaMinNeSe); /*--------------------------------------------------------------------------*/ // (9) half4 rgby1; rgby1.xy = dir1_pos.zw + dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; rgby1 = h4tex2Dlod(tex, half4(rgby1.xy, 0.0, 0.0)); rgby1 = (temp1N + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (10) half4 rgbyM = h4tex2Dlod(tex, half4(pos.xy, 0.0, 0.0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaMaxM = max(lumaMax, rgbyM.w); half lumaMinM = min(lumaMin, rgbyM.w); #else half lumaMaxM = max(lumaMax, rgbyM.y); half lumaMinM = min(lumaMin, rgbyM.y); #endif /*--------------------------------------------------------------------------*/ // (11) half4 temp2N; temp2N.xy = dir2_pos.zw - dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; temp2N = h4tex2Dlod(tex, half4(temp2N.xy, 0.0, 0.0)); half4 rgby2; rgby2.xy = dir2_pos.zw + dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; half lumaRangeM = (lumaMaxM - lumaMinM) / FXAA_CONSOLE_PS3_EDGE_THRESHOLD; /*--------------------------------------------------------------------------*/ // (12) rgby2 = h4tex2Dlod(tex, half4(rgby2.xy, 0.0, 0.0)); rgby2 = (temp2N + rgby2) * 0.5; /*--------------------------------------------------------------------------*/ // (13) rgby2 = (rgby2 + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (14) #if (FXAA_GREEN_AS_LUMA == 0) bool twoTapLt = rgby2.w < lumaMin; bool twoTapGt = rgby2.w > lumaMax; #else bool twoTapLt = rgby2.y < lumaMin; bool twoTapGt = rgby2.y > lumaMax; #endif bool earlyExit = lumaRangeM < lumaMax; bool twoTap = twoTapLt || twoTapGt; /*--------------------------------------------------------------------------*/ // (15) if(twoTap) rgby2 = rgby1; if(earlyExit) rgby2 = rgbyM; /*--------------------------------------------------------------------------*/ return rgby2; } /*==========================================================================*/ #endif #endif ////#line 1 #ifndef SRGB_DEFINED #define SRGB_DEFINED #if HAS_SRGB vec3 to_srgb(vec3 lin) { return mix( 1.055 * pow(lin,vec3(0.41666)) - 0.055, 12.92 * lin, vec3(lessThanEqual(lin,vec3(0.0031308)))); } vec3 to_linear(vec3 srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),vec3(2.4)), srgb * (1.0 / 12.92), vec3(lessThanEqual(srgb,vec3(0.04045)))); } float to_srgb(float lin) { return mix( 1.055 * pow(lin,0.41666) - 0.055, 12.92 * lin, lin <= 0.0031308 ? 1.0 : 0.0); } float to_linear(float srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),2.4), srgb * (1.0 / 12.92), srgb <= 0.04045 ? 1.0 : 0.0); } #else vec3 to_linear(vec3 srgb) { return pow(srgb,vec3(2.2)); } vec3 to_srgb(vec3 lin) { return pow(lin,vec3(0.454545)); } float to_linear(float srgb) { return pow(srgb,2.2); } float to_srgb(float lin) { return pow(lin,0.454545); } #endif #endif /* SRGB_DEFINED */ //#line 27 VARYING(vec4, v_screen_pos) VARYING(vec4, v_color) #if TONE_MAP ////#line 1 #ifndef WANT_LIGHTMETER #define WANT_LIGHTMETER 0 #endif //////#line 1 #ifndef SRGB_DEFINED #define SRGB_DEFINED #if HAS_SRGB vec3 to_srgb(vec3 lin) { return mix( 1.055 * pow(lin,vec3(0.41666)) - 0.055, 12.92 * lin, vec3(lessThanEqual(lin,vec3(0.0031308)))); } vec3 to_linear(vec3 srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),vec3(2.4)), srgb * (1.0 / 12.92), vec3(lessThanEqual(srgb,vec3(0.04045)))); } float to_srgb(float lin) { return mix( 1.055 * pow(lin,0.41666) - 0.055, 12.92 * lin, lin <= 0.0031308 ? 1.0 : 0.0); } float to_linear(float srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),2.4), srgb * (1.0 / 12.92), srgb <= 0.04045 ? 1.0 : 0.0); } #else vec3 to_linear(vec3 srgb) { return pow(srgb,vec3(2.2)); } vec3 to_srgb(vec3 lin) { return pow(lin,vec3(0.454545)); } float to_linear(float srgb) { return pow(srgb,2.2); } float to_srgb(float lin) { return pow(lin,0.454545); } #endif #endif /* SRGB_DEFINED */ ////#line 8 #if !defined(VSHADER_TONE_ONLY) VARYING(float, v_tonemap_gain) #endif uniform sampler2D u_tex_tone; #define bias 0.75 #if WANT_LIGHTMETER VARYING(vec2, amp_meter) #endif #ifndef MODE_TONE_DYNAMIC #define MODE_TONE_DYNAMIC 2 #endif #ifndef TONE_MODE #define TONE_MODE 2 #endif #if TONE_MODE == MODE_TONE_DYNAMIC #define tonemap_gain u_tonemap_gain #else #define tonemap_gain vec4(0, 0, 0, 3.2) #endif float A = 0.15; float B = 0.50; float C = 0.10; float D = 0.20; float E = 0.02; float F = 0.30; vec3 W = vec3(3.2); vec3 Uncharted2Tonemap(vec3 x) { return ((x*(A*x+C*B)+D*E)/(x*(A*x+B)+D*F))-E/F; } float R_white = 3.0; vec3 Reinhard(vec3 L) { vec3 num = L * (L/(R_white*R_white) + 1.0); vec3 den = L + 1.0; return num/den; } // If we are going to tone map using Alex's epic equation, // go to sRGB-ish space first; his stuff is tuned to sRGB, // not linear. void tone_map_uncharted(inout vec3 rgb, in float gain) { // dumb clamping // rgb = clamp(rgb,0.0,1.0); // naive coolor Reinhardt - needs srgb // rgb = Reinhard(rgb); // Dawson - no srgb // vec3 x = max(rgb-0.004,0.0); // rgb = (x*(6.2*x+.5))/(x*(6.2*x+1.7)+0.06); // uncharted 2 float ExposureBias = gain; vec3 curr = Uncharted2Tonemap(ExposureBias*rgb); vec3 whiteScale = 1.0f/Uncharted2Tonemap(vec3(tonemap_gain.w)); rgb = curr*whiteScale; } void tone_map_gifford(inout vec3 rgb, in float gain) { // rgb = to_srgb(rgb); vec3 blend_rat = pow(clamp(rgb,0.0,1.0),vec3(bias)); rgb = mix(rgb,1.0 - pow(clamp(1.0 - rgb,0.0,1.0), vec3(gain)),blend_rat); } #if VSHADER #if !defined(VSHADER_TONE_ONLY) void calc_light_meter() { // Tone mapping light meter. // This samples the dynamic exposure texture (a 1x1 result of much copying) // and figures out how much gain to apply. Hoisted up into the vertex shader // to cut ops out of fragment processing - the tone map levels do not vary // by frame position!! // The light meter texture contains the _immediate_ brightness in "r" and // a blended value in "g" - the result is that g assymptotically r. // By sampling the max of both, we take the brighter of what we see now or // what we did see, for a fast adjustment when things get bright but slow // adjustment when things get dim. vec2 light_meter_raw = texture2DLod(u_tex_tone,vec2(0.5,0.5),0).rg; float overall_color = max(light_meter_raw.r,light_meter_raw.g); float amp = (overall_color); float amp_lim = clamp(amp,tonemap_gain.y,tonemap_gain.z); float total_gain = tonemap_gain.x / amp_lim; v_tonemap_gain = total_gain; #if WANT_LIGHTMETER amp_meter = vec2(amp,amp_lim); #endif } #endif void vshader_tone_map(inout vec3 rgb) { vec2 light_meter_raw = texture2DLod(u_tex_tone,vec2(0.5,0.5),0).rg; float overall_color = max(light_meter_raw.r,light_meter_raw.g); float amp = (overall_color); float amp_lim = clamp(amp,tonemap_gain.y,tonemap_gain.z); float total_gain = tonemap_gain.x / amp_lim; tone_map_uncharted(rgb, total_gain); } void vshader_tone_map_srgb(inout vec3 rgb) { rgb = to_linear(rgb); vshader_tone_map(rgb); rgb = to_srgb(rgb); } #endif #if FSHADER && !defined(VSHADER_TONE_ONLY) void tone_map(inout vec3 rgb) { tone_map_uncharted(rgb, v_tonemap_gain); // And now...Alex's epic equation. v_tonemap_gain is a gain // ratio (e.g. 1.0 is unity) and bias is a fixed constant // for where in the brightness curve we start to squash. // Generally lower biases squash lower brightness levels and // as a result make tings more, um, "crispy". // Light meter code - just for debugging - not particularly // well optimized. // How to read the light meter! // The yellow interval shows the range of clamping for the light // meter's input. This effectively limits how much it can increase // or decrease the gain (since the exposure is lowered by the // brightness of the input image). A wider range means dynamic // exposure can do more, and more space to the left means darks are // brought up more (because the system will go "whoa, the input image // is rully, rully dark) and crank it. // // The top bar is the light meter - left is darker, right is brighter. // The bar is green when it is within the unclamped range and is thus // affecting exposure; it is red when it is out of range and thus // small changes to brightness don't change exposure. // The bottom bar is the total gain applied to the frame, taking into // account both the fixed gain adjustment from skyc.cpp and the tone // mapping adjustment from the level meter. This bar is green when // the image is being made brighter and red when it is being made // darker. #if WANT_LIGHTMETER if(gl_FragCoord.y < 20) { float total_gain_log = log2(v_tonemap_gain); float amp_percent = (total_gain_log / 16.0) + 0.5; float pixel_pos = gl_FragCoord.x / 512.0; if(amp_percent < 0.5) { if(amp_percent > pixel_pos) rgb = vec3(1,0,0); } else { if(amp_percent > pixel_pos) rgb = vec3(0,1,0); } } else if (gl_FragCoord.y < 40) { float amp_log = log2(amp_meter.x); float amp_lim_log = log2(amp_meter.y); float meter_percent = (amp_log / 16.0) + 0.5; float pixel_pos = gl_FragCoord.x / 512.0; if(amp_lim_log != amp_log) { if(meter_percent > pixel_pos) rgb = vec3(1,0,0); } else { if(meter_percent > pixel_pos) rgb = vec3(0,1,0); } } else if(gl_FragCoord.y < 48) { vec2 bounds = log2(tonemap_gain.yz); vec2 range = vec2(bounds / 16.0) + 0.5; float pixel_pos = gl_FragCoord.x / 512.0; if(pixel_pos > range.x && pixel_pos < range.y) { rgb = vec3(1,1,0); } } #endif } void tone_map_srgb(inout vec3 rgb) { rgb = to_linear(rgb); tone_map(rgb); rgb = to_srgb(rgb); } void tone_map_srgb_to_linear(inout vec3 rgb) { rgb = to_linear(rgb); tone_map(rgb); } void tone_map_linear_to_srgb(inout vec3 rgb) { tone_map(rgb); rgb = to_srgb(rgb); } #endif #endif //#line 48 #if VSHADER ATTRIBUTE(vec4, a_color) ATTRIBUTE(vec4, a_vertex) void main (void) { STD_BROADCAST #if HAS_FULL_TRANSFORM gl_Position = (uStereoOffsets[uViewportTarget] * u_mvp_matrix) * (a_vertex - vec4(u_mv_offset,0.0)); gl_ViewportIndex = uViewportTarget; #else gl_Position = a_vertex; gl_ViewportIndex = uViewportTarget; #endif v_color = a_color; #if TONE_MAP calc_light_meter(); #endif } #endif #if FSHADER uniform sampler2D u_tex1; uniform sampler2D u_tex2; #define K 1.0 vec4 final_color_correct(vec4 in_color) { vec3 rgb = in_color.rgb; #if TONE_MAP tone_map(rgb); #endif return vec4(rgb,in_color.a); } void main (void) { #if MODE==MODE_BLUR gl_FragColor = u_kernel2.w * texture2D(u_tex1, v_texcoord0.st - 7.0 * u_offset * K) + u_kernel2.z * texture2D(u_tex1, v_texcoord0.st - 6.0 * u_offset * K) + u_kernel2.y * texture2D(u_tex1, v_texcoord0.st - 5.0 * u_offset * K) + u_kernel2.x * texture2D(u_tex1, v_texcoord0.st - 4.0 * u_offset * K) + u_kernel1.w * texture2D(u_tex1, v_texcoord0.st - 3.0 * u_offset * K) + u_kernel1.z * texture2D(u_tex1, v_texcoord0.st - 2.0 * u_offset * K) + u_kernel1.y * texture2D(u_tex1, v_texcoord0.st - u_offset * K) + u_kernel1.x * texture2D(u_tex1, v_texcoord0.st ) + u_kernel1.y * texture2D(u_tex1, v_texcoord0.st + u_offset * K) + u_kernel1.z * texture2D(u_tex1, v_texcoord0.st + 2.0 * u_offset * K) + u_kernel1.w * texture2D(u_tex1, v_texcoord0.st + 3.0 * u_offset * K) + u_kernel2.x * texture2D(u_tex1, v_texcoord0.st + 4.0 * u_offset * K) + u_kernel2.y * texture2D(u_tex1, v_texcoord0.st + 5.0 * u_offset * K) + u_kernel2.z * texture2D(u_tex1, v_texcoord0.st + 6.0 * u_offset * K) + u_kernel2.w * texture2D(u_tex1, v_texcoord0.st + 7.0 * u_offset * K); #endif #if MODE==MODE_COPY gl_FragColor = v_color * final_color_correct(texture2D(u_tex1, gl_FragCoord.xy * u_offset)); #endif #if MODE==MODE_GRAY // Grayscale shader. v_color's rgb is the weight ratios to convert to gray - usually // mostly green since that's how our eyes work. vec4 rgba = texture2D(u_tex1, gl_FragCoord.xy * u_offset); rgba.rgb = vec3(dot(v_color.rgb,rgba.rgb)); rgba = final_color_correct(rgba); rgba.a *= v_color.a; gl_FragColor = rgba; #endif #if MODE==MODE_HEAT vec2 screen_rat = gl_FragCoord.xy * u_offset; float t1 = fract(u_time); t1 = (t1 > 0.5) ? (1.0 - t1) : t1; t1 *= 2.0; float t2 = 1.0 - t1; vec2 noise_st1 = texture2D(u_tex2, screen_rat*0.5 +vec2(0,u_time)).rg * 2.0 - vec2(1.0); vec2 noise_st2 = texture2D(u_tex2, screen_rat*0.5 + vec2(0.5)+vec2(u_time,0)).rg * 2.0 - vec2(1.0); vec2 uv = screen_rat + noise_st1 * 0.005 * t1 + noise_st2 * 0.005 * t2; vec3 blur = texture2D(u_tex1, uv).rgb; gl_FragColor = vec4(blur,v_color.a); gl_FragColor = final_color_correct(gl_FragColor); #endif #if MODE==MODE_FXAA gl_FragColor = vec4(0.0); vec2 tex_ctr = gl_FragCoord.xy * (u_offset * u_super_dim.xy); for(int x = 0; x < u_super_dim.x; ++x) for(int y = 0; y < u_super_dim.y; ++y) { vec2 sub_pos = tex_ctr + vec2(x,y) * u_offset - u_super_dim.zw; gl_FragColor += FxaaPixelShader( sub_pos, vec4(0.0), // console pos u_tex1, // tex u_tex1, // console tex-1 u_tex1, // console tex-2 u_offset, // 1/screen size vec4(0.0), // console screen offsets vec4(0.0), // console screen offsets vec4(0.0), // console screen offsets 0.75, // sub px qual 0.166, // edge thesh 0.0833, // min thresh 0.0, // sharpness (console) 0.05, // console edge 0.0, // console edge min vec4(0.0)); // console directional constants } gl_FragColor *= (1.0 / (u_super_dim.x * u_super_dim.y)); gl_FragColor = final_color_correct(gl_FragColor); gl_FragColor.a = 1.0; #endif } #endif SHADER SOURCE FOR FRAGMENT SHADER: Resources/shaders/hdr.glsl #version 150 compatibility uniform vec2 uFiScreenSize; uniform vec2 uFiScreenInv; uniform int uViewportTarget; uniform float uEyeShift; #define VSHADER 0 #define GSHADER 0 #define FSHADER 1 #define TCSHADER 0 #define TESHADER 0 #define APL 0 #define IBM 1 #define LIN 0 #define MOBILE 0 #define HAS_SRGB 1 #define HAS_SHADER_LOD 1 #define HAS_UBO 1 #define HAS_ARRAY_TEX 1 #define HAS_TEX_RG 1 #define HAS_TESS 1 #define HAS_AMD_SHADER_VIEWPORT_LAYER_ARRAY 1 #define HAS_GSHADER 0 #define MODE 0 #define TONE_MAP 1 #define USE_UBOS 1 #ifndef HAS_CORE #define HAS_CORE 0 #endif #if __VERSION__ >= 130 // For everything MODERN we support, use the "new" generic texture function names. texture() has no macro // arguments because the bias parameter is optional. We define the old GPU shader4 specific types to generic // because we can't macro the other way. // This case covers GLSL 130 (various Intel GPUs), GLSL 150 + arb compatibility (modern red and green team // drivers on Win/Lin) and GLSL 150 ("core" profile, still experimental) #define texture2D texture #define texture2DArray texture #define texture2DProj(a,b) textureProj(a,b) #define texture2DGrad(a,b,c,d) textureGrad(a,b,c,d) #define texture2DLod(a,b,c) textureLod(a,b,c) #define texture2DArrayLod(a,b,c) textureLod(a,b,c) #define shadow2D(a,b) vec4(texture(a,b)) #define shadow2DArray(a,b) vec4(texture(a,b)) #define shadow2DGrad(a,b,c,d) vec4(textureGrad(a,b,c,d)) #define shadow2DArrayGrad(a,b,c,d) vec4(textureGrad(a,b,c,d)) #elif defined(GL_EXT_gpu_shader4) // On OS X legacy profile we're stuck at GLSL 120 + a ton of extensions. Grab GPU_shader4 - if we don't have // that we are missing a TON of stuff we need. Then cope with the quirk of uint missing. // Note that for all practical purposes we expect to ONLY hit this case on OS X. #extension GL_EXT_gpu_shader4 : enable #define uint unsigned int #else // Back in the day we supported even OLDER skankier stuff, but no need now - every supported GPU // we support is at least GPU_shader4 (which matches shader model 4.0, which is in DX10) - because // DX10 is a MANDATORY level of hw for XP11. So we can just pumt out. This is convenient because // GPU_shader4 gets us gradients and texture arrays. #error we require GLSL 130 or GLSL 120 + GPU_shader4 #endif #if HAS_CORE #define texture3D texture #define textureCubeLod textureLod #endif // Opt into a bunch of useful stuff we want. #if HAS_AMD_SHADER_VIEWPORT_LAYER_ARRAY && HAS_UBO #if defined(GL_ARB_shader_viewport_layer_array) #extension GL_ARB_shader_viewport_layer_array : enable #define STD_BROADCAST \ gl_ViewportIndex = gl_InstanceID % u_viewport_count; \ gl_Layer = gl_InstanceID % u_layer_count; #define HAS_BROADCAST 1 #else #extension GL_AMD_vertex_shader_layer : enable #extension GL_AMD_vertex_shader_viewport_index : enable #define STD_BROADCAST \ gl_ViewportIndex = gl_InstanceID % u_viewport_count; \ gl_Layer = gl_InstanceID % u_layer_count; #define HAS_BROADCAST 1 #endif #else #define STD_BROADCAST #define HAS_BROADCAST 0 #endif #if defined(GL_ARB_gpu_shader5) #extension GL_ARB_gpu_shader5 : enable #endif #if HAS_UBO && !HAS_CORE #extension GL_ARB_uniform_buffer_object : enable #endif #if HAS_TESS && !HAS_CORE && !defined(GL_ARB_gpu_shader5) #extension GL_ARB_tessellation_shader : enable #endif #if __VERSION__ >= 130 // If we run GLSL >= 130 we should use in/out instead of varying and attribute // These macros wrap the legacy keywords to their modern equivalent #if VSHADER #define VARYING(t,v) out t v; #define ATTRIBUTE(t,v) in t v; #elif FSHADER #define VARYING(t,v) in t v; #endif #if TCSHADER #define VARYING(t,v) in t v[]; out t v##_tess[]; #elif TESHADER #define VARYING(t,v) in t v##_tess[]; out t v; #endif #else #define VARYING(t,v) varying t v; #if VSHADER #define ATTRIBUTE(t,v) attribute t v; #endif #endif // This is a hack-around for a broken mat3 cast in some version of the GLSL // compiler on OS X. Tragically I didn't note WHICH version has the bug. TODO in // v11: confirm that on our new minimum OS (10.9? 10.10?) this isn't needed. mat3 mat3_cast(mat4 m) { return mat3(m[0].xyz,m[1].xyz,m[2].xyz); } #define mat3(X) mat3_cast(X) #ifndef USE_UBOS #define USE_UBOS 0 #endif #if VSHADER #if !HAS_BROADCAST || !HAS_UBO #define u_modelview_matrix modelview_matrix #define u_mvp_matrix mvp_matrix #define u_projection_matrix projection_matrix #define u_modelview_inverse modelview_inverse #define u_eye (0) #else #define u_modelview_matrix modelview_matrix[gl_InstanceID % u_transform_count] #define u_mvp_matrix mvp_matrix[gl_InstanceID % u_transform_count] #define u_projection_matrix projection_matrix[gl_InstanceID % u_transform_count] #define u_modelview_inverse modelview_inverse[gl_InstanceID % u_transform_count] #define u_eye (gl_InstanceID % u_viewport_count) #endif #endif layout(std140) uniform u_environment_data { vec3 u_fog_rgb; float u_fog_scale; vec3 u_sun_position; vec3 u_sun_specular; vec3 u_sun_ambient; vec3 u_sun_diffuse; float u_sky_gain; vec4 u_od_scale; vec4 u_scatter_gain; vec4 u_tonemap_gain; vec2 fade_dist; }; layout(std140) uniform u_hdr_data { vec4 u_kernel1; vec4 u_kernel2; vec4 u_super_dim; vec2 (uFiScreenInv * vec2(0.5, 1.0)); float u_time; }; layout(std140) uniform u_transform { int u_viewport_count; int u_layer_count; int u_transform_count; vec3 u_mv_offset; mat4 modelview_matrix[6]; mat4 modelview_inverse[6]; mat4 projection_matrix[6]; mat4 mvp_matrix[6]; }; #define MODE_COPY 0 #define MODE_GRAY 1 #define MODE_BLUR 2 #define MODE_HEAT 3 #define MODE_FXAA 4 #if MODE==MODE_HEAT #define HAS_FULL_TRANSFORM 1 #else #define HAS_FULL_TRANSFORM 0 #endif // Set this to 1 to get light metering HUD in HDR mode. #define WANT_LIGHTMETER 0 #define FXAA_GREEN_AS_LUMA 1 #if __VERSION__ >= 130 #define FXAA_GLSL_130 1 #else #define FXAA_GLSL_120 1 #endif #if FSHADER && MODE == MODE_FXAA /*============================================================================ NVIDIA FXAA 3.11 by TIMOTHY LOTTES ------------------------------------------------------------------------------ COPYRIGHT (C) 2010, 2011 NVIDIA CORPORATION. ALL RIGHTS RESERVED. ------------------------------------------------------------------------------ TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THIS SOFTWARE IS PROVIDED *AS IS* AND NVIDIA AND ITS SUPPLIERS DISCLAIM ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL NVIDIA OR ITS SUPPLIERS BE LIABLE FOR ANY SPECIAL, INCIDENTAL, INDIRECT, OR CONSEQUENTIAL DAMAGES WHATSOEVER (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF BUSINESS INFORMATION, OR ANY OTHER PECUNIARY LOSS) ARISING OUT OF THE USE OF OR INABILITY TO USE THIS SOFTWARE, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. ------------------------------------------------------------------------------ CHANGE LOG ------------------------------------------------------------------------------ 11/25/13 - BAS - removed __ from macros to comply with GLSL 1.20.8 spec. ------------------------------------------------------------------------------ INTEGRATION CHECKLIST ------------------------------------------------------------------------------ (1.) In the shader source, setup defines for the desired configuration. When providing multiple shaders (for different presets), simply setup the defines differently in multiple files. Example, #define FXAA_PC 1 #define FXAA_HLSL_5 1 #define FXAA_QUALITY_PRESET 12 Or, #define FXAA_360 1 Or, #define FXAA_PS3 1 Etc. (2.) Then include this file, # include "Fxaa3_11.h" (3.) Then call the FXAA pixel shader from within your desired shader. Look at the FXAA Quality FxaaPixelShader() for docs on inputs. As for FXAA 3.11 all inputs for all shaders are the same to enable easy porting between platforms. return FxaaPixelShader(...); (4.) Insure pass prior to FXAA outputs RGBL (see next section). Or use, #define FXAA_GREEN_AS_LUMA 1 (5.) Setup engine to provide the following constants which are used in the FxaaPixelShader() inputs, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir Look at the FXAA Quality FxaaPixelShader() for docs on inputs. (6.) Have FXAA vertex shader run as a full screen triangle, and output "pos" and "fxaaConsolePosPos" such that inputs in the pixel shader provide, // {xy} = center of pixel FxaaFloat2 pos, // {xy__} = upper left of pixel // {__zw} = lower right of pixel FxaaFloat4 fxaaConsolePosPos, (7.) Insure the texture sampler(s) used by FXAA are set to bilinear filtering. ------------------------------------------------------------------------------ INTEGRATION - RGBL AND COLORSPACE ------------------------------------------------------------------------------ FXAA3 requires RGBL as input unless the following is set, #define FXAA_GREEN_AS_LUMA 1 In which case the engine uses green in place of luma, and requires RGB input is in a non-linear colorspace. RGB should be LDR (low dynamic range). Specifically do FXAA after tonemapping. RGB data as returned by a texture fetch can be non-linear, or linear when FXAA_GREEN_AS_LUMA is not set. Note an "sRGB format" texture counts as linear, because the result of a texture fetch is linear data. Regular "RGBA8" textures in the sRGB colorspace are non-linear. If FXAA_GREEN_AS_LUMA is not set, luma must be stored in the alpha channel prior to running FXAA. This luma should be in a perceptual space (could be gamma 2.0). Example pass before FXAA where output is gamma 2.0 encoded, color.rgb = ToneMap(color.rgb); // linear color output color.rgb = sqrt(color.rgb); // gamma 2.0 color output return color; To use FXAA, color.rgb = ToneMap(color.rgb); // linear color output color.rgb = sqrt(color.rgb); // gamma 2.0 color output color.a = dot(color.rgb, FxaaFloat3(0.299, 0.587, 0.114)); // compute luma return color; Another example where output is linear encoded, say for instance writing to an sRGB formated render target, where the render target does the conversion back to sRGB after blending, color.rgb = ToneMap(color.rgb); // linear color output return color; To use FXAA, color.rgb = ToneMap(color.rgb); // linear color output color.a = sqrt(dot(color.rgb, FxaaFloat3(0.299, 0.587, 0.114))); // compute luma return color; Getting luma correct is required for the algorithm to work correctly. ------------------------------------------------------------------------------ BEING LINEARLY CORRECT? ------------------------------------------------------------------------------ Applying FXAA to a framebuffer with linear RGB color will look worse. This is very counter intuitive, but happends to be true in this case. The reason is because dithering artifacts will be more visiable in a linear colorspace. ------------------------------------------------------------------------------ COMPLEX INTEGRATION ------------------------------------------------------------------------------ Q. What if the engine is blending into RGB before wanting to run FXAA? A. In the last opaque pass prior to FXAA, have the pass write out luma into alpha. Then blend into RGB only. FXAA should be able to run ok assuming the blending pass did not any add aliasing. This should be the common case for particles and common blending passes. A. Or use FXAA_GREEN_AS_LUMA. ============================================================================*/ /*============================================================================ INTEGRATION KNOBS ============================================================================*/ // // FXAA_PS3 and FXAA_360 choose the console algorithm (FXAA3 CONSOLE). // FXAA_360_OPT is a prototype for the new optimized 360 version. // // 1 = Use API. // 0 = Don't use API. // /*--------------------------------------------------------------------------*/ #ifndef FXAA_PS3 #define FXAA_PS3 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_360 #define FXAA_360 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_360_OPT #define FXAA_360_OPT 0 #endif /*==========================================================================*/ #ifndef FXAA_PC // // FXAA Quality // The high quality PC algorithm. // #define FXAA_PC 1 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_PC_CONSOLE // // The console algorithm for PC is included // for developers targeting really low spec machines. // Likely better to just run FXAA_PC, and use a really low preset. // #define FXAA_PC_CONSOLE 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_GLSL_120 #define FXAA_GLSL_120 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_GLSL_130 #define FXAA_GLSL_130 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_HLSL_3 #define FXAA_HLSL_3 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_HLSL_4 #define FXAA_HLSL_4 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_HLSL_5 #define FXAA_HLSL_5 0 #endif /*==========================================================================*/ #ifndef FXAA_GREEN_AS_LUMA // // For those using non-linear color, // and either not able to get luma in alpha, or not wanting to, // this enables FXAA to run using green as a proxy for luma. // So with this enabled, no need to pack luma in alpha. // // This will turn off AA on anything which lacks some amount of green. // Pure red and blue or combination of only R and B, will get no AA. // // Might want to lower the settings for both, // fxaaConsoleEdgeThresholdMin // fxaaQualityEdgeThresholdMin // In order to insure AA does not get turned off on colors // which contain a minor amount of green. // // 1 = On. // 0 = Off. // #define FXAA_GREEN_AS_LUMA 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_EARLY_EXIT // // Controls algorithm's early exit path. // On PS3 turning this ON adds 2 cycles to the shader. // On 360 turning this OFF adds 10ths of a millisecond to the shader. // Turning this off on console will result in a more blurry image. // So this defaults to on. // // 1 = On. // 0 = Off. // #define FXAA_EARLY_EXIT 1 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_DISCARD // // Only valid for PC OpenGL currently. // Probably will not work when FXAA_GREEN_AS_LUMA = 1. // // 1 = Use discard on pixels which don't need AA. // For APIs which enable concurrent TEX+ROP from same surface. // 0 = Return unchanged color on pixels which don't need AA. // #define FXAA_DISCARD 0 #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_FAST_PIXEL_OFFSET // // Used for GLSL 120 only. // // 1 = GL API supports fast pixel offsets // 0 = do not use fast pixel offsets // #ifdef GL_EXT_gpu_shader4 #define FXAA_FAST_PIXEL_OFFSET 1 #endif #ifdef GL_NV_gpu_shader5 #define FXAA_FAST_PIXEL_OFFSET 1 #endif #ifdef GL_ARB_gpu_shader5 #define FXAA_FAST_PIXEL_OFFSET 1 #endif #ifndef FXAA_FAST_PIXEL_OFFSET #define FXAA_FAST_PIXEL_OFFSET 0 #endif #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_GATHER4_ALPHA // // 1 = API supports gather4 on alpha channel. // 0 = API does not support gather4 on alpha channel. // #if (FXAA_HLSL_5 == 1) #define FXAA_GATHER4_ALPHA 1 #endif #ifdef GL_ARB_gpu_shader5 #define FXAA_GATHER4_ALPHA 1 #endif #ifdef GL_NV_gpu_shader5 #define FXAA_GATHER4_ALPHA 1 #endif #ifndef FXAA_GATHER4_ALPHA #define FXAA_GATHER4_ALPHA 0 #endif #endif /*============================================================================ FXAA CONSOLE PS3 - TUNING KNOBS ============================================================================*/ #ifndef FXAA_CONSOLE_PS3_EDGE_SHARPNESS // // Consoles the sharpness of edges on PS3 only. // Non-PS3 tuning is done with shader input. // // Due to the PS3 being ALU bound, // there are only two safe values here: 4 and 8. // These options use the shaders ability to a free *|/ by 2|4|8. // // 8.0 is sharper // 4.0 is softer // 2.0 is really soft (good for vector graphics inputs) // #if 1 #define FXAA_CONSOLE_PS3_EDGE_SHARPNESS 8.0 #endif #if 0 #define FXAA_CONSOLE_PS3_EDGE_SHARPNESS 4.0 #endif #if 0 #define FXAA_CONSOLE_PS3_EDGE_SHARPNESS 2.0 #endif #endif /*--------------------------------------------------------------------------*/ #ifndef FXAA_CONSOLE_PS3_EDGE_THRESHOLD // // Only effects PS3. // Non-PS3 tuning is done with shader input. // // The minimum amount of local contrast required to apply algorithm. // The console setting has a different mapping than the quality setting. // // This only applies when FXAA_EARLY_EXIT is 1. // // Due to the PS3 being ALU bound, // there are only two safe values here: 0.25 and 0.125. // These options use the shaders ability to a free *|/ by 2|4|8. // // 0.125 leaves less aliasing, but is softer // 0.25 leaves more aliasing, and is sharper // #if 1 #define FXAA_CONSOLE_PS3_EDGE_THRESHOLD 0.125 #else #define FXAA_CONSOLE_PS3_EDGE_THRESHOLD 0.25 #endif #endif /*============================================================================ FXAA QUALITY - TUNING KNOBS ------------------------------------------------------------------------------ NOTE the other tuning knobs are now in the shader function inputs! ============================================================================*/ #ifndef FXAA_QUALITY_PRESET // // Choose the quality preset. // This needs to be compiled into the shader as it effects code. // Best option to include multiple presets is to // in each shader define the preset, then include this file. // // OPTIONS // ----------------------------------------------------------------------- // 10 to 15 - default medium dither (10=fastest, 15=highest quality) // 20 to 29 - less dither, more expensive (20=fastest, 29=highest quality) // 39 - no dither, very expensive // // NOTES // ----------------------------------------------------------------------- // 12 = slightly faster then FXAA 3.9 and higher edge quality (default) // 13 = about same speed as FXAA 3.9 and better than 12 // 23 = closest to FXAA 3.9 visually and performance wise // _ = the lowest digit is directly related to performance // _ = the highest digit is directly related to style // #define FXAA_QUALITY_PRESET 12 #endif /*============================================================================ FXAA QUALITY - PRESETS ============================================================================*/ /*============================================================================ FXAA QUALITY - MEDIUM DITHER PRESETS ============================================================================*/ #if (FXAA_QUALITY_PRESET == 10) #define FXAA_QUALITY_PS 3 #define FXAA_QUALITY_P0 1.5 #define FXAA_QUALITY_P1 3.0 #define FXAA_QUALITY_P2 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 11) #define FXAA_QUALITY_PS 4 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 3.0 #define FXAA_QUALITY_P3 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 12) #define FXAA_QUALITY_PS 5 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 4.0 #define FXAA_QUALITY_P4 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 13) #define FXAA_QUALITY_PS 6 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 4.0 #define FXAA_QUALITY_P5 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 14) #define FXAA_QUALITY_PS 7 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 4.0 #define FXAA_QUALITY_P6 12.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 15) #define FXAA_QUALITY_PS 8 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 4.0 #define FXAA_QUALITY_P7 12.0 #endif /*============================================================================ FXAA QUALITY - LOW DITHER PRESETS ============================================================================*/ #if (FXAA_QUALITY_PRESET == 20) #define FXAA_QUALITY_PS 3 #define FXAA_QUALITY_P0 1.5 #define FXAA_QUALITY_P1 2.0 #define FXAA_QUALITY_P2 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 21) #define FXAA_QUALITY_PS 4 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 22) #define FXAA_QUALITY_PS 5 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 23) #define FXAA_QUALITY_PS 6 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 24) #define FXAA_QUALITY_PS 7 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 3.0 #define FXAA_QUALITY_P6 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 25) #define FXAA_QUALITY_PS 8 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 4.0 #define FXAA_QUALITY_P7 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 26) #define FXAA_QUALITY_PS 9 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 4.0 #define FXAA_QUALITY_P8 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 27) #define FXAA_QUALITY_PS 10 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 4.0 #define FXAA_QUALITY_P9 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 28) #define FXAA_QUALITY_PS 11 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 2.0 #define FXAA_QUALITY_P9 4.0 #define FXAA_QUALITY_P10 8.0 #endif /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PRESET == 29) #define FXAA_QUALITY_PS 12 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.5 #define FXAA_QUALITY_P2 2.0 #define FXAA_QUALITY_P3 2.0 #define FXAA_QUALITY_P4 2.0 #define FXAA_QUALITY_P5 2.0 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 2.0 #define FXAA_QUALITY_P9 2.0 #define FXAA_QUALITY_P10 4.0 #define FXAA_QUALITY_P11 8.0 #endif /*============================================================================ FXAA QUALITY - EXTREME QUALITY ============================================================================*/ #if (FXAA_QUALITY_PRESET == 39) #define FXAA_QUALITY_PS 12 #define FXAA_QUALITY_P0 1.0 #define FXAA_QUALITY_P1 1.0 #define FXAA_QUALITY_P2 1.0 #define FXAA_QUALITY_P3 1.0 #define FXAA_QUALITY_P4 1.0 #define FXAA_QUALITY_P5 1.5 #define FXAA_QUALITY_P6 2.0 #define FXAA_QUALITY_P7 2.0 #define FXAA_QUALITY_P8 2.0 #define FXAA_QUALITY_P9 2.0 #define FXAA_QUALITY_P10 4.0 #define FXAA_QUALITY_P11 8.0 #endif /*============================================================================ API PORTING ============================================================================*/ #if (FXAA_GLSL_120 == 1) || (FXAA_GLSL_130 == 1) #define FxaaBool bool #define FxaaDiscard discard #define FxaaFloat float #define FxaaFloat2 vec2 #define FxaaFloat3 vec3 #define FxaaFloat4 vec4 #define FxaaHalf float #define FxaaHalf2 vec2 #define FxaaHalf3 vec3 #define FxaaHalf4 vec4 #define FxaaInt2 ivec2 #define FxaaSat(x) clamp(x, 0.0, 1.0) #define FxaaTex sampler2D #else #define FxaaBool bool #define FxaaDiscard clip(-1) #define FxaaFloat float #define FxaaFloat2 float2 #define FxaaFloat3 float3 #define FxaaFloat4 float4 #define FxaaHalf half #define FxaaHalf2 half2 #define FxaaHalf3 half3 #define FxaaHalf4 half4 #define FxaaSat(x) saturate(x) #endif /*--------------------------------------------------------------------------*/ #if (FXAA_GLSL_120 == 1) // Requires, // #version 120 // And at least, // #extension GL_EXT_gpu_shader4 : enable // (or set FXAA_FAST_PIXEL_OFFSET 1 to work like DX9) #define FxaaTexTop(t, p) texture2DLod(t, p, 0.0) #if (FXAA_FAST_PIXEL_OFFSET == 1) #define FxaaTexOff(t, p, o, r) texture2DLodOffset(t, p, 0.0, o) #else #define FxaaTexOff(t, p, o, r) texture2DLod(t, p + (o * r), 0.0) #endif #if (FXAA_GATHER4_ALPHA == 1) // use #extension GL_ARB_gpu_shader5 : enable #define FxaaTexAlpha4(t, p) textureGather(t, p, 3) #define FxaaTexOffAlpha4(t, p, o) textureGatherOffset(t, p, o, 3) #define FxaaTexGreen4(t, p) textureGather(t, p, 1) #define FxaaTexOffGreen4(t, p, o) textureGatherOffset(t, p, o, 1) #endif #endif /*--------------------------------------------------------------------------*/ #if (FXAA_GLSL_130 == 1) // Requires "#version 130" or better #define FxaaTexTop(t, p) textureLod(t, p, 0.0) #define FxaaTexOff(t, p, o, r) textureLodOffset(t, p, 0.0, o) #if (FXAA_GATHER4_ALPHA == 1) // use #extension GL_ARB_gpu_shader5 : enable #define FxaaTexAlpha4(t, p) textureGather(t, p, 3) #define FxaaTexOffAlpha4(t, p, o) textureGatherOffset(t, p, o, 3) #define FxaaTexGreen4(t, p) textureGather(t, p, 1) #define FxaaTexOffGreen4(t, p, o) textureGatherOffset(t, p, o, 1) #endif #endif /*--------------------------------------------------------------------------*/ #if (FXAA_HLSL_3 == 1) || (FXAA_360 == 1) || (FXAA_PS3 == 1) #define FxaaInt2 float2 #define FxaaTex sampler2D #define FxaaTexTop(t, p) tex2Dlod(t, float4(p, 0.0, 0.0)) #define FxaaTexOff(t, p, o, r) tex2Dlod(t, float4(p + (o * r), 0, 0)) #endif /*--------------------------------------------------------------------------*/ #if (FXAA_HLSL_4 == 1) #define FxaaInt2 int2 struct FxaaTex { SamplerState smpl; Texture2D tex; }; #define FxaaTexTop(t, p) t.tex.SampleLevel(t.smpl, p, 0.0) #define FxaaTexOff(t, p, o, r) t.tex.SampleLevel(t.smpl, p, 0.0, o) #endif /*--------------------------------------------------------------------------*/ #if (FXAA_HLSL_5 == 1) #define FxaaInt2 int2 struct FxaaTex { SamplerState smpl; Texture2D tex; }; #define FxaaTexTop(t, p) t.tex.SampleLevel(t.smpl, p, 0.0) #define FxaaTexOff(t, p, o, r) t.tex.SampleLevel(t.smpl, p, 0.0, o) #define FxaaTexAlpha4(t, p) t.tex.GatherAlpha(t.smpl, p) #define FxaaTexOffAlpha4(t, p, o) t.tex.GatherAlpha(t.smpl, p, o) #define FxaaTexGreen4(t, p) t.tex.GatherGreen(t.smpl, p) #define FxaaTexOffGreen4(t, p, o) t.tex.GatherGreen(t.smpl, p, o) #endif /*============================================================================ GREEN AS LUMA OPTION SUPPORT FUNCTION ============================================================================*/ #if (FXAA_GREEN_AS_LUMA == 0) FxaaFloat FxaaLuma(FxaaFloat4 rgba) { return rgba.w; } #else FxaaFloat FxaaLuma(FxaaFloat4 rgba) { return rgba.y; } #endif /*============================================================================ FXAA3 QUALITY - PC ============================================================================*/ #if (FXAA_PC == 1) /*--------------------------------------------------------------------------*/ FxaaFloat4 FxaaPixelShader( // // Use noperspective interpolation here (turn off perspective interpolation). // {xy} = center of pixel FxaaFloat2 pos, // // Used only for FXAA Console, and not used on the 360 version. // Use noperspective interpolation here (turn off perspective interpolation). // {xy__} = upper left of pixel // {__zw} = lower right of pixel FxaaFloat4 fxaaConsolePosPos, // // Input color texture. // {rgb_} = color in linear or perceptual color space // if (FXAA_GREEN_AS_LUMA == 0) // {___a} = luma in perceptual color space (not linear) FxaaTex tex, // // Only used on the optimized 360 version of FXAA Console. // For everything but 360, just use the same input here as for "tex". // For 360, same texture, just alias with a 2nd sampler. // This sampler needs to have an exponent bias of -1. FxaaTex fxaaConsole360TexExpBiasNegOne, // // Only used on the optimized 360 version of FXAA Console. // For everything but 360, just use the same input here as for "tex". // For 360, same texture, just alias with a 3nd sampler. // This sampler needs to have an exponent bias of -2. FxaaTex fxaaConsole360TexExpBiasNegTwo, // // Only used on FXAA Quality. // This must be from a constant/uniform. // {x_} = 1.0/screenWidthInPixels // {_y} = 1.0/screenHeightInPixels FxaaFloat2 fxaaQualityRcpFrame, // // Only used on FXAA Console. // This must be from a constant/uniform. // This effects sub-pixel AA quality and inversely sharpness. // Where N ranges between, // N = 0.50 (default) // N = 0.33 (sharper) // {x___} = -N/screenWidthInPixels // {_y__} = -N/screenHeightInPixels // {__z_} = N/screenWidthInPixels // {___w} = N/screenHeightInPixels FxaaFloat4 fxaaConsoleRcpFrameOpt, // // Only used on FXAA Console. // Not used on 360, but used on PS3 and PC. // This must be from a constant/uniform. // {x___} = -2.0/screenWidthInPixels // {_y__} = -2.0/screenHeightInPixels // {__z_} = 2.0/screenWidthInPixels // {___w} = 2.0/screenHeightInPixels FxaaFloat4 fxaaConsoleRcpFrameOpt2, // // Only used on FXAA Console. // Only used on 360 in place of fxaaConsoleRcpFrameOpt2. // This must be from a constant/uniform. // {x___} = 8.0/screenWidthInPixels // {_y__} = 8.0/screenHeightInPixels // {__z_} = -4.0/screenWidthInPixels // {___w} = -4.0/screenHeightInPixels FxaaFloat4 fxaaConsole360RcpFrameOpt2, // // Only used on FXAA Quality. // This used to be the FXAA_QUALITY_SUBPIX define. // It is here now to allow easier tuning. // Choose the amount of sub-pixel aliasing removal. // This can effect sharpness. // 1.00 - upper limit (softer) // 0.75 - default amount of filtering // 0.50 - lower limit (sharper, less sub-pixel aliasing removal) // 0.25 - almost off // 0.00 - completely off FxaaFloat fxaaQualitySubpix, // // Only used on FXAA Quality. // This used to be the FXAA_QUALITY_EDGE_THRESHOLD define. // It is here now to allow easier tuning. // The minimum amount of local contrast required to apply algorithm. // 0.333 - too little (faster) // 0.250 - low quality // 0.166 - default // 0.125 - high quality // 0.063 - overkill (slower) FxaaFloat fxaaQualityEdgeThreshold, // // Only used on FXAA Quality. // This used to be the FXAA_QUALITY_EDGE_THRESHOLD_MIN define. // It is here now to allow easier tuning. // Trims the algorithm from processing darks. // 0.0833 - upper limit (default, the start of visible unfiltered edges) // 0.0625 - high quality (faster) // 0.0312 - visible limit (slower) // Special notes when using FXAA_GREEN_AS_LUMA, // Likely want to set this to zero. // As colors that are mostly not-green // will appear very dark in the green channel! // Tune by looking at mostly non-green content, // then start at zero and increase until aliasing is a problem. FxaaFloat fxaaQualityEdgeThresholdMin, // // Only used on FXAA Console. // This used to be the FXAA_CONSOLE_EDGE_SHARPNESS define. // It is here now to allow easier tuning. // This does not effect PS3, as this needs to be compiled in. // Use FXAA_CONSOLE_PS3_EDGE_SHARPNESS for PS3. // Due to the PS3 being ALU bound, // there are only three safe values here: 2 and 4 and 8. // These options use the shaders ability to a free *|/ by 2|4|8. // For all other platforms can be a non-power of two. // 8.0 is sharper (default!!!) // 4.0 is softer // 2.0 is really soft (good only for vector graphics inputs) FxaaFloat fxaaConsoleEdgeSharpness, // // Only used on FXAA Console. // This used to be the FXAA_CONSOLE_EDGE_THRESHOLD define. // It is here now to allow easier tuning. // This does not effect PS3, as this needs to be compiled in. // Use FXAA_CONSOLE_PS3_EDGE_THRESHOLD for PS3. // Due to the PS3 being ALU bound, // there are only two safe values here: 1/4 and 1/8. // These options use the shaders ability to a free *|/ by 2|4|8. // The console setting has a different mapping than the quality setting. // Other platforms can use other values. // 0.125 leaves less aliasing, but is softer (default!!!) // 0.25 leaves more aliasing, and is sharper FxaaFloat fxaaConsoleEdgeThreshold, // // Only used on FXAA Console. // This used to be the FXAA_CONSOLE_EDGE_THRESHOLD_MIN define. // It is here now to allow easier tuning. // Trims the algorithm from processing darks. // The console setting has a different mapping than the quality setting. // This only applies when FXAA_EARLY_EXIT is 1. // This does not apply to PS3, // PS3 was simplified to avoid more shader instructions. // 0.06 - faster but more aliasing in darks // 0.05 - default // 0.04 - slower and less aliasing in darks // Special notes when using FXAA_GREEN_AS_LUMA, // Likely want to set this to zero. // As colors that are mostly not-green // will appear very dark in the green channel! // Tune by looking at mostly non-green content, // then start at zero and increase until aliasing is a problem. FxaaFloat fxaaConsoleEdgeThresholdMin, // // Extra constants for 360 FXAA Console only. // Use zeros or anything else for other platforms. // These must be in physical constant registers and NOT immedates. // Immedates will result in compiler un-optimizing. // {xyzw} = float4(1.0, -1.0, 0.25, -0.25) FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ FxaaFloat2 posM; posM.x = pos.x; posM.y = pos.y; #if (FXAA_GATHER4_ALPHA == 1) #if (FXAA_DISCARD == 0) FxaaFloat4 rgbyM = FxaaTexTop(tex, posM); #if (FXAA_GREEN_AS_LUMA == 0) #define lumaM rgbyM.w #else #define lumaM rgbyM.y #endif #endif #if (FXAA_GREEN_AS_LUMA == 0) FxaaFloat4 luma4A = FxaaTexAlpha4(tex, posM); FxaaFloat4 luma4B = FxaaTexOffAlpha4(tex, posM, FxaaInt2(-1, -1)); #else FxaaFloat4 luma4A = FxaaTexGreen4(tex, posM); FxaaFloat4 luma4B = FxaaTexOffGreen4(tex, posM, FxaaInt2(-1, -1)); #endif #if (FXAA_DISCARD == 1) #define lumaM luma4A.w #endif #define lumaE luma4A.z #define lumaS luma4A.x #define lumaSE luma4A.y #define lumaNW luma4B.w #define lumaN luma4B.z #define lumaW luma4B.x #else FxaaFloat4 rgbyM = FxaaTexTop(tex, posM); #if (FXAA_GREEN_AS_LUMA == 0) #define lumaM rgbyM.w #else #define lumaM rgbyM.y #endif FxaaFloat lumaS = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 0, 1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 1, 0), fxaaQualityRcpFrame.xy)); FxaaFloat lumaN = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 0,-1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1, 0), fxaaQualityRcpFrame.xy)); #endif /*--------------------------------------------------------------------------*/ FxaaFloat maxSM = max(lumaS, lumaM); FxaaFloat minSM = min(lumaS, lumaM); FxaaFloat maxESM = max(lumaE, maxSM); FxaaFloat minESM = min(lumaE, minSM); FxaaFloat maxWN = max(lumaN, lumaW); FxaaFloat minWN = min(lumaN, lumaW); FxaaFloat rangeMax = max(maxWN, maxESM); FxaaFloat rangeMin = min(minWN, minESM); FxaaFloat rangeMaxScaled = rangeMax * fxaaQualityEdgeThreshold; FxaaFloat range = rangeMax - rangeMin; FxaaFloat rangeMaxClamped = max(fxaaQualityEdgeThresholdMin, rangeMaxScaled); FxaaBool earlyExit = range < rangeMaxClamped; /*--------------------------------------------------------------------------*/ if(earlyExit) #if (FXAA_DISCARD == 1) FxaaDiscard; #else return rgbyM; #endif /*--------------------------------------------------------------------------*/ #if (FXAA_GATHER4_ALPHA == 0) FxaaFloat lumaNW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1,-1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaSE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 1, 1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaNE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2( 1,-1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaSW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1, 1), fxaaQualityRcpFrame.xy)); #else FxaaFloat lumaNE = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(1, -1), fxaaQualityRcpFrame.xy)); FxaaFloat lumaSW = FxaaLuma(FxaaTexOff(tex, posM, FxaaInt2(-1, 1), fxaaQualityRcpFrame.xy)); #endif /*--------------------------------------------------------------------------*/ FxaaFloat lumaNS = lumaN + lumaS; FxaaFloat lumaWE = lumaW + lumaE; FxaaFloat subpixRcpRange = 1.0/range; FxaaFloat subpixNSWE = lumaNS + lumaWE; FxaaFloat edgeHorz1 = (-2.0 * lumaM) + lumaNS; FxaaFloat edgeVert1 = (-2.0 * lumaM) + lumaWE; /*--------------------------------------------------------------------------*/ FxaaFloat lumaNESE = lumaNE + lumaSE; FxaaFloat lumaNWNE = lumaNW + lumaNE; FxaaFloat edgeHorz2 = (-2.0 * lumaE) + lumaNESE; FxaaFloat edgeVert2 = (-2.0 * lumaN) + lumaNWNE; /*--------------------------------------------------------------------------*/ FxaaFloat lumaNWSW = lumaNW + lumaSW; FxaaFloat lumaSWSE = lumaSW + lumaSE; FxaaFloat edgeHorz4 = (abs(edgeHorz1) * 2.0) + abs(edgeHorz2); FxaaFloat edgeVert4 = (abs(edgeVert1) * 2.0) + abs(edgeVert2); FxaaFloat edgeHorz3 = (-2.0 * lumaW) + lumaNWSW; FxaaFloat edgeVert3 = (-2.0 * lumaS) + lumaSWSE; FxaaFloat edgeHorz = abs(edgeHorz3) + edgeHorz4; FxaaFloat edgeVert = abs(edgeVert3) + edgeVert4; /*--------------------------------------------------------------------------*/ FxaaFloat subpixNWSWNESE = lumaNWSW + lumaNESE; FxaaFloat lengthSign = fxaaQualityRcpFrame.x; FxaaBool horzSpan = edgeHorz >= edgeVert; FxaaFloat subpixA = subpixNSWE * 2.0 + subpixNWSWNESE; /*--------------------------------------------------------------------------*/ if(!horzSpan) lumaN = lumaW; if(!horzSpan) lumaS = lumaE; if(horzSpan) lengthSign = fxaaQualityRcpFrame.y; FxaaFloat subpixB = (subpixA * (1.0/12.0)) - lumaM; /*--------------------------------------------------------------------------*/ FxaaFloat gradientN = lumaN - lumaM; FxaaFloat gradientS = lumaS - lumaM; FxaaFloat lumaNN = lumaN + lumaM; FxaaFloat lumaSS = lumaS + lumaM; FxaaBool pairN = abs(gradientN) >= abs(gradientS); FxaaFloat gradient = max(abs(gradientN), abs(gradientS)); if(pairN) lengthSign = -lengthSign; FxaaFloat subpixC = FxaaSat(abs(subpixB) * subpixRcpRange); /*--------------------------------------------------------------------------*/ FxaaFloat2 posB; posB.x = posM.x; posB.y = posM.y; FxaaFloat2 offNP; offNP.x = (!horzSpan) ? 0.0 : fxaaQualityRcpFrame.x; offNP.y = ( horzSpan) ? 0.0 : fxaaQualityRcpFrame.y; if(!horzSpan) posB.x += lengthSign * 0.5; if( horzSpan) posB.y += lengthSign * 0.5; /*--------------------------------------------------------------------------*/ FxaaFloat2 posN; posN.x = posB.x - offNP.x * FXAA_QUALITY_P0; posN.y = posB.y - offNP.y * FXAA_QUALITY_P0; FxaaFloat2 posP; posP.x = posB.x + offNP.x * FXAA_QUALITY_P0; posP.y = posB.y + offNP.y * FXAA_QUALITY_P0; FxaaFloat subpixD = ((-2.0)*subpixC) + 3.0; FxaaFloat lumaEndN = FxaaLuma(FxaaTexTop(tex, posN)); FxaaFloat subpixE = subpixC * subpixC; FxaaFloat lumaEndP = FxaaLuma(FxaaTexTop(tex, posP)); /*--------------------------------------------------------------------------*/ if(!pairN) lumaNN = lumaSS; FxaaFloat gradientScaled = gradient * 1.0/4.0; FxaaFloat lumaMM = lumaM - lumaNN * 0.5; FxaaFloat subpixF = subpixD * subpixE; FxaaBool lumaMLTZero = lumaMM < 0.0; /*--------------------------------------------------------------------------*/ lumaEndN -= lumaNN * 0.5; lumaEndP -= lumaNN * 0.5; FxaaBool doneN = abs(lumaEndN) >= gradientScaled; FxaaBool doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P1; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P1; FxaaBool doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P1; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P1; /*--------------------------------------------------------------------------*/ if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P2; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P2; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P2; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P2; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 3) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P3; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P3; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P3; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P3; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 4) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P4; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P4; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P4; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P4; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 5) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P5; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P5; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P5; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P5; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 6) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P6; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P6; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P6; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P6; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 7) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P7; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P7; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P7; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P7; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 8) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P8; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P8; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P8; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P8; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 9) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P9; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P9; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P9; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P9; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 10) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P10; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P10; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P10; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P10; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 11) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P11; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P11; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P11; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P11; /*--------------------------------------------------------------------------*/ #if (FXAA_QUALITY_PS > 12) if(doneNP) { if(!doneN) lumaEndN = FxaaLuma(FxaaTexTop(tex, posN.xy)); if(!doneP) lumaEndP = FxaaLuma(FxaaTexTop(tex, posP.xy)); if(!doneN) lumaEndN = lumaEndN - lumaNN * 0.5; if(!doneP) lumaEndP = lumaEndP - lumaNN * 0.5; doneN = abs(lumaEndN) >= gradientScaled; doneP = abs(lumaEndP) >= gradientScaled; if(!doneN) posN.x -= offNP.x * FXAA_QUALITY_P12; if(!doneN) posN.y -= offNP.y * FXAA_QUALITY_P12; doneNP = (!doneN) || (!doneP); if(!doneP) posP.x += offNP.x * FXAA_QUALITY_P12; if(!doneP) posP.y += offNP.y * FXAA_QUALITY_P12; /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } #endif /*--------------------------------------------------------------------------*/ } /*--------------------------------------------------------------------------*/ FxaaFloat dstN = posM.x - posN.x; FxaaFloat dstP = posP.x - posM.x; if(!horzSpan) dstN = posM.y - posN.y; if(!horzSpan) dstP = posP.y - posM.y; /*--------------------------------------------------------------------------*/ FxaaBool goodSpanN = (lumaEndN < 0.0) != lumaMLTZero; FxaaFloat spanLength = (dstP + dstN); FxaaBool goodSpanP = (lumaEndP < 0.0) != lumaMLTZero; FxaaFloat spanLengthRcp = 1.0/spanLength; /*--------------------------------------------------------------------------*/ FxaaBool directionN = dstN < dstP; FxaaFloat dst = min(dstN, dstP); FxaaBool goodSpan = directionN ? goodSpanN : goodSpanP; FxaaFloat subpixG = subpixF * subpixF; FxaaFloat pixelOffset = (dst * (-spanLengthRcp)) + 0.5; FxaaFloat subpixH = subpixG * fxaaQualitySubpix; /*--------------------------------------------------------------------------*/ FxaaFloat pixelOffsetGood = goodSpan ? pixelOffset : 0.0; FxaaFloat pixelOffsetSubpix = max(pixelOffsetGood, subpixH); if(!horzSpan) posM.x += pixelOffsetSubpix * lengthSign; if( horzSpan) posM.y += pixelOffsetSubpix * lengthSign; #if (FXAA_DISCARD == 1) return FxaaTexTop(tex, posM); #else return FxaaFloat4(FxaaTexTop(tex, posM).xyz, lumaM); #endif } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - PC VERSION ------------------------------------------------------------------------------ Instead of using this on PC, I'd suggest just using FXAA Quality with #define FXAA_QUALITY_PRESET 10 Or #define FXAA_QUALITY_PRESET 20 Either are higher qualilty and almost as fast as this on modern PC GPUs. ============================================================================*/ #if (FXAA_PC_CONSOLE == 1) /*--------------------------------------------------------------------------*/ FxaaFloat4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ FxaaFloat lumaNw = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.xy)); FxaaFloat lumaSw = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.xw)); FxaaFloat lumaNe = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.zy)); FxaaFloat lumaSe = FxaaLuma(FxaaTexTop(tex, fxaaConsolePosPos.zw)); /*--------------------------------------------------------------------------*/ FxaaFloat4 rgbyM = FxaaTexTop(tex, pos.xy); #if (FXAA_GREEN_AS_LUMA == 0) FxaaFloat lumaM = rgbyM.w; #else FxaaFloat lumaM = rgbyM.y; #endif /*--------------------------------------------------------------------------*/ FxaaFloat lumaMaxNwSw = max(lumaNw, lumaSw); lumaNe += 1.0/384.0; FxaaFloat lumaMinNwSw = min(lumaNw, lumaSw); /*--------------------------------------------------------------------------*/ FxaaFloat lumaMaxNeSe = max(lumaNe, lumaSe); FxaaFloat lumaMinNeSe = min(lumaNe, lumaSe); /*--------------------------------------------------------------------------*/ FxaaFloat lumaMax = max(lumaMaxNeSe, lumaMaxNwSw); FxaaFloat lumaMin = min(lumaMinNeSe, lumaMinNwSw); /*--------------------------------------------------------------------------*/ FxaaFloat lumaMaxScaled = lumaMax * fxaaConsoleEdgeThreshold; /*--------------------------------------------------------------------------*/ FxaaFloat lumaMinM = min(lumaMin, lumaM); FxaaFloat lumaMaxScaledClamped = max(fxaaConsoleEdgeThresholdMin, lumaMaxScaled); FxaaFloat lumaMaxM = max(lumaMax, lumaM); FxaaFloat dirSwMinusNe = lumaSw - lumaNe; FxaaFloat lumaMaxSubMinM = lumaMaxM - lumaMinM; FxaaFloat dirSeMinusNw = lumaSe - lumaNw; if(lumaMaxSubMinM < lumaMaxScaledClamped) return rgbyM; /*--------------------------------------------------------------------------*/ FxaaFloat2 dir; dir.x = dirSwMinusNe + dirSeMinusNw; dir.y = dirSwMinusNe - dirSeMinusNw; /*--------------------------------------------------------------------------*/ FxaaFloat2 dir1 = normalize(dir.xy); FxaaFloat4 rgbyN1 = FxaaTexTop(tex, pos.xy - dir1 * fxaaConsoleRcpFrameOpt.zw); FxaaFloat4 rgbyP1 = FxaaTexTop(tex, pos.xy + dir1 * fxaaConsoleRcpFrameOpt.zw); /*--------------------------------------------------------------------------*/ FxaaFloat dirAbsMinTimesC = min(abs(dir1.x), abs(dir1.y)) * fxaaConsoleEdgeSharpness; FxaaFloat2 dir2 = clamp(dir1.xy / dirAbsMinTimesC, -2.0, 2.0); /*--------------------------------------------------------------------------*/ FxaaFloat4 rgbyN2 = FxaaTexTop(tex, pos.xy - dir2 * fxaaConsoleRcpFrameOpt2.zw); FxaaFloat4 rgbyP2 = FxaaTexTop(tex, pos.xy + dir2 * fxaaConsoleRcpFrameOpt2.zw); /*--------------------------------------------------------------------------*/ FxaaFloat4 rgbyA = rgbyN1 + rgbyP1; FxaaFloat4 rgbyB = ((rgbyN2 + rgbyP2) * 0.25) + (rgbyA * 0.25); /*--------------------------------------------------------------------------*/ #if (FXAA_GREEN_AS_LUMA == 0) FxaaBool twoTap = (rgbyB.w < lumaMin) || (rgbyB.w > lumaMax); #else FxaaBool twoTap = (rgbyB.y < lumaMin) || (rgbyB.y > lumaMax); #endif if(twoTap) rgbyB.xyz = rgbyA.xyz * 0.5; return rgbyB; } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - 360 PIXEL SHADER ------------------------------------------------------------------------------ This optimized version thanks to suggestions from Andy Luedke. Should be fully tex bound in all cases. As of the FXAA 3.11 release, I have still not tested this code, however I fixed a bug which was in both FXAA 3.9 and FXAA 3.10. And note this is replacing the old unoptimized version. If it does not work, please let me know so I can fix it. ============================================================================*/ #if (FXAA_360 == 1) /*--------------------------------------------------------------------------*/ [reduceTempRegUsage(4)] float4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ float4 lumaNwNeSwSe; #if (FXAA_GREEN_AS_LUMA == 0) asm { tfetch2D lumaNwNeSwSe.w___, tex, pos.xy, OffsetX = -0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe._w__, tex, pos.xy, OffsetX = 0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.__w_, tex, pos.xy, OffsetX = -0.5, OffsetY = 0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.___w, tex, pos.xy, OffsetX = 0.5, OffsetY = 0.5, UseComputedLOD=false }; #else asm { tfetch2D lumaNwNeSwSe.y___, tex, pos.xy, OffsetX = -0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe._y__, tex, pos.xy, OffsetX = 0.5, OffsetY = -0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.__y_, tex, pos.xy, OffsetX = -0.5, OffsetY = 0.5, UseComputedLOD=false tfetch2D lumaNwNeSwSe.___y, tex, pos.xy, OffsetX = 0.5, OffsetY = 0.5, UseComputedLOD=false }; #endif /*--------------------------------------------------------------------------*/ lumaNwNeSwSe.y += 1.0/384.0; float2 lumaMinTemp = min(lumaNwNeSwSe.xy, lumaNwNeSwSe.zw); float2 lumaMaxTemp = max(lumaNwNeSwSe.xy, lumaNwNeSwSe.zw); float lumaMin = min(lumaMinTemp.x, lumaMinTemp.y); float lumaMax = max(lumaMaxTemp.x, lumaMaxTemp.y); /*--------------------------------------------------------------------------*/ float4 rgbyM = tex2Dlod(tex, float4(pos.xy, 0.0, 0.0)); #if (FXAA_GREEN_AS_LUMA == 0) float lumaMinM = min(lumaMin, rgbyM.w); float lumaMaxM = max(lumaMax, rgbyM.w); #else float lumaMinM = min(lumaMin, rgbyM.y); float lumaMaxM = max(lumaMax, rgbyM.y); #endif if((lumaMaxM - lumaMinM) < max(fxaaConsoleEdgeThresholdMin, lumaMax * fxaaConsoleEdgeThreshold)) return rgbyM; /*--------------------------------------------------------------------------*/ float2 dir; dir.x = dot(lumaNwNeSwSe, fxaaConsole360ConstDir.yyxx); dir.y = dot(lumaNwNeSwSe, fxaaConsole360ConstDir.xyxy); dir = normalize(dir); /*--------------------------------------------------------------------------*/ float4 dir1 = dir.xyxy * fxaaConsoleRcpFrameOpt.xyzw; /*--------------------------------------------------------------------------*/ float4 dir2; float dirAbsMinTimesC = min(abs(dir.x), abs(dir.y)) * fxaaConsoleEdgeSharpness; dir2 = saturate(fxaaConsole360ConstDir.zzww * dir.xyxy / dirAbsMinTimesC + 0.5); dir2 = dir2 * fxaaConsole360RcpFrameOpt2.xyxy + fxaaConsole360RcpFrameOpt2.zwzw; /*--------------------------------------------------------------------------*/ float4 rgbyN1 = tex2Dlod(fxaaConsole360TexExpBiasNegOne, float4(pos.xy + dir1.xy, 0.0, 0.0)); float4 rgbyP1 = tex2Dlod(fxaaConsole360TexExpBiasNegOne, float4(pos.xy + dir1.zw, 0.0, 0.0)); float4 rgbyN2 = tex2Dlod(fxaaConsole360TexExpBiasNegTwo, float4(pos.xy + dir2.xy, 0.0, 0.0)); float4 rgbyP2 = tex2Dlod(fxaaConsole360TexExpBiasNegTwo, float4(pos.xy + dir2.zw, 0.0, 0.0)); /*--------------------------------------------------------------------------*/ float4 rgbyA = rgbyN1 + rgbyP1; float4 rgbyB = rgbyN2 + rgbyP2 + rgbyA * 0.5; /*--------------------------------------------------------------------------*/ float4 rgbyR = ((FxaaLuma(rgbyB) - lumaMax) > 0.0) ? rgbyA : rgbyB; rgbyR = ((FxaaLuma(rgbyB) - lumaMin) > 0.0) ? rgbyR : rgbyA; return rgbyR; } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - OPTIMIZED PS3 PIXEL SHADER (NO EARLY EXIT) ============================================================================== The code below does not exactly match the assembly. I have a feeling that 12 cycles is possible, but was not able to get there. Might have to increase register count to get full performance. Note this shader does not use perspective interpolation. Use the following cgc options, --fenable-bx2 --fastmath --fastprecision --nofloatbindings ------------------------------------------------------------------------------ NVSHADERPERF OUTPUT ------------------------------------------------------------------------------ For reference and to aid in debug, output of NVShaderPerf should match this, Shader to schedule: 0: texpkb h0.w(TRUE), v5.zyxx, #0 2: addh h2.z(TRUE), h0.w, constant(0.001953, 0.000000, 0.000000, 0.000000).x 4: texpkb h0.w(TRUE), v5.xwxx, #0 6: addh h0.z(TRUE), -h2, h0.w 7: texpkb h1.w(TRUE), v5, #0 9: addh h0.x(TRUE), h0.z, -h1.w 10: addh h3.w(TRUE), h0.z, h1 11: texpkb h2.w(TRUE), v5.zwzz, #0 13: addh h0.z(TRUE), h3.w, -h2.w 14: addh h0.x(TRUE), h2.w, h0 15: nrmh h1.xz(TRUE), h0_n 16: minh_m8 h0.x(TRUE), |h1|, |h1.z| 17: maxh h4.w(TRUE), h0, h1 18: divx h2.xy(TRUE), h1_n.xzzw, h0_n 19: movr r1.zw(TRUE), v4.xxxy 20: madr r2.xz(TRUE), -h1, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).zzww, r1.zzww 22: minh h5.w(TRUE), h0, h1 23: texpkb h0(TRUE), r2.xzxx, #0 25: madr r0.zw(TRUE), h1.xzxz, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w), r1 27: maxh h4.x(TRUE), h2.z, h2.w 28: texpkb h1(TRUE), r0.zwzz, #0 30: addh_d2 h1(TRUE), h0, h1 31: madr r0.xy(TRUE), -h2, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 33: texpkb h0(TRUE), r0, #0 35: minh h4.z(TRUE), h2, h2.w 36: fenct TRUE 37: madr r1.xy(TRUE), h2, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 39: texpkb h2(TRUE), r1, #0 41: addh_d2 h0(TRUE), h0, h2 42: maxh h2.w(TRUE), h4, h4.x 43: minh h2.x(TRUE), h5.w, h4.z 44: addh_d2 h0(TRUE), h0, h1 45: slth h2.x(TRUE), h0.w, h2 46: sgth h2.w(TRUE), h0, h2 47: movh h0(TRUE), h0 48: addx.c0 rc(TRUE), h2, h2.w 49: movh h0(c0.NE.x), h1 IPU0 ------ Simplified schedule: -------- Pass | Unit | uOp | PC: Op -----+--------+------+------------------------- 1 | SCT0/1 | mov | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | TEX | txl | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | SCB1 | add | 2: ADDh h2.z, h0.--w-, const.--x-; | | | 2 | SCT0/1 | mov | 4: TXLr h0.w, g[TEX1].xwxx, const.xxxx, TEX0; | TEX | txl | 4: TXLr h0.w, g[TEX1].xwxx, const.xxxx, TEX0; | SCB1 | add | 6: ADDh h0.z,-h2, h0.--w-; | | | 3 | SCT0/1 | mov | 7: TXLr h1.w, g[TEX1], const.xxxx, TEX0; | TEX | txl | 7: TXLr h1.w, g[TEX1], const.xxxx, TEX0; | SCB0 | add | 9: ADDh h0.x, h0.z---,-h1.w---; | SCB1 | add | 10: ADDh h3.w, h0.---z, h1; | | | 4 | SCT0/1 | mov | 11: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | TEX | txl | 11: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | SCB0 | add | 14: ADDh h0.x, h2.w---, h0; | SCB1 | add | 13: ADDh h0.z, h3.--w-,-h2.--w-; | | | 5 | SCT1 | mov | 15: NRMh h1.xz, h0; | SRB | nrm | 15: NRMh h1.xz, h0; | SCB0 | min | 16: MINh*8 h0.x, |h1|, |h1.z---|; | SCB1 | max | 17: MAXh h4.w, h0, h1; | | | 6 | SCT0 | div | 18: DIVx h2.xy, h1.xz--, h0; | SCT1 | mov | 19: MOVr r1.zw, g[TEX0].--xy; | SCB0 | mad | 20: MADr r2.xz,-h1, const.z-w-, r1.z-w-; | SCB1 | min | 22: MINh h5.w, h0, h1; | | | 7 | SCT0/1 | mov | 23: TXLr h0, r2.xzxx, const.xxxx, TEX0; | TEX | txl | 23: TXLr h0, r2.xzxx, const.xxxx, TEX0; | SCB0 | max | 27: MAXh h4.x, h2.z---, h2.w---; | SCB1 | mad | 25: MADr r0.zw, h1.--xz, const, r1; | | | 8 | SCT0/1 | mov | 28: TXLr h1, r0.zwzz, const.xxxx, TEX0; | TEX | txl | 28: TXLr h1, r0.zwzz, const.xxxx, TEX0; | SCB0/1 | add | 30: ADDh/2 h1, h0, h1; | | | 9 | SCT0 | mad | 31: MADr r0.xy,-h2, const.xy--, r1.zw--; | SCT1 | mov | 33: TXLr h0, r0, const.zzzz, TEX0; | TEX | txl | 33: TXLr h0, r0, const.zzzz, TEX0; | SCB1 | min | 35: MINh h4.z, h2, h2.--w-; | | | 10 | SCT0 | mad | 37: MADr r1.xy, h2, const.xy--, r1.zw--; | SCT1 | mov | 39: TXLr h2, r1, const.zzzz, TEX0; | TEX | txl | 39: TXLr h2, r1, const.zzzz, TEX0; | SCB0/1 | add | 41: ADDh/2 h0, h0, h2; | | | 11 | SCT0 | min | 43: MINh h2.x, h5.w---, h4.z---; | SCT1 | max | 42: MAXh h2.w, h4, h4.---x; | SCB0/1 | add | 44: ADDh/2 h0, h0, h1; | | | 12 | SCT0 | set | 45: SLTh h2.x, h0.w---, h2; | SCT1 | set | 46: SGTh h2.w, h0, h2; | SCB0/1 | mul | 47: MOVh h0, h0; | | | 13 | SCT0 | mad | 48: ADDxc0_s rc, h2, h2.w---; | SCB0/1 | mul | 49: MOVh h0(NE0.xxxx), h1; Pass SCT TEX SCB 1: 0% 100% 25% 2: 0% 100% 25% 3: 0% 100% 50% 4: 0% 100% 50% 5: 0% 0% 50% 6: 100% 0% 75% 7: 0% 100% 75% 8: 0% 100% 100% 9: 0% 100% 25% 10: 0% 100% 100% 11: 50% 0% 100% 12: 50% 0% 100% 13: 25% 0% 100% MEAN: 17% 61% 67% Pass SCT0 SCT1 TEX SCB0 SCB1 1: 0% 0% 100% 0% 100% 2: 0% 0% 100% 0% 100% 3: 0% 0% 100% 100% 100% 4: 0% 0% 100% 100% 100% 5: 0% 0% 0% 100% 100% 6: 100% 100% 0% 100% 100% 7: 0% 0% 100% 100% 100% 8: 0% 0% 100% 100% 100% 9: 0% 0% 100% 0% 100% 10: 0% 0% 100% 100% 100% 11: 100% 100% 0% 100% 100% 12: 100% 100% 0% 100% 100% 13: 100% 0% 0% 100% 100% MEAN: 30% 23% 61% 76% 100% Fragment Performance Setup: Driver RSX Compiler, GPU RSX, Flags 0x5 Results 13 cycles, 3 r regs, 923,076,923 pixels/s ============================================================================*/ #if (FXAA_PS3 == 1) && (FXAA_EARLY_EXIT == 0) /*--------------------------------------------------------------------------*/ #pragma regcount 7 #pragma disablepc all #pragma option O3 #pragma option OutColorPrec=fp16 #pragma texformat default RGBA8 /*==========================================================================*/ half4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ // (1) half4 dir; half4 lumaNe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) lumaNe.w += half(1.0/512.0); dir.x = -lumaNe.w; dir.z = -lumaNe.w; #else lumaNe.y += half(1.0/512.0); dir.x = -lumaNe.y; dir.z = -lumaNe.y; #endif /*--------------------------------------------------------------------------*/ // (2) half4 lumaSw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) dir.x += lumaSw.w; dir.z += lumaSw.w; #else dir.x += lumaSw.y; dir.z += lumaSw.y; #endif /*--------------------------------------------------------------------------*/ // (3) half4 lumaNw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) dir.x -= lumaNw.w; dir.z += lumaNw.w; #else dir.x -= lumaNw.y; dir.z += lumaNw.y; #endif /*--------------------------------------------------------------------------*/ // (4) half4 lumaSe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) dir.x += lumaSe.w; dir.z -= lumaSe.w; #else dir.x += lumaSe.y; dir.z -= lumaSe.y; #endif /*--------------------------------------------------------------------------*/ // (5) half4 dir1_pos; dir1_pos.xy = normalize(dir.xyz).xz; half dirAbsMinTimesC = min(abs(dir1_pos.x), abs(dir1_pos.y)) * half(FXAA_CONSOLE_PS3_EDGE_SHARPNESS); /*--------------------------------------------------------------------------*/ // (6) half4 dir2_pos; dir2_pos.xy = clamp(dir1_pos.xy / dirAbsMinTimesC, half(-2.0), half(2.0)); dir1_pos.zw = pos.xy; dir2_pos.zw = pos.xy; half4 temp1N; temp1N.xy = dir1_pos.zw - dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; /*--------------------------------------------------------------------------*/ // (7) temp1N = h4tex2Dlod(tex, half4(temp1N.xy, 0.0, 0.0)); half4 rgby1; rgby1.xy = dir1_pos.zw + dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; /*--------------------------------------------------------------------------*/ // (8) rgby1 = h4tex2Dlod(tex, half4(rgby1.xy, 0.0, 0.0)); rgby1 = (temp1N + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (9) half4 temp2N; temp2N.xy = dir2_pos.zw - dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; temp2N = h4tex2Dlod(tex, half4(temp2N.xy, 0.0, 0.0)); /*--------------------------------------------------------------------------*/ // (10) half4 rgby2; rgby2.xy = dir2_pos.zw + dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; rgby2 = h4tex2Dlod(tex, half4(rgby2.xy, 0.0, 0.0)); rgby2 = (temp2N + rgby2) * 0.5; /*--------------------------------------------------------------------------*/ // (11) // compilier moves these scalar ops up to other cycles #if (FXAA_GREEN_AS_LUMA == 0) half lumaMin = min(min(lumaNw.w, lumaSw.w), min(lumaNe.w, lumaSe.w)); half lumaMax = max(max(lumaNw.w, lumaSw.w), max(lumaNe.w, lumaSe.w)); #else half lumaMin = min(min(lumaNw.y, lumaSw.y), min(lumaNe.y, lumaSe.y)); half lumaMax = max(max(lumaNw.y, lumaSw.y), max(lumaNe.y, lumaSe.y)); #endif rgby2 = (rgby2 + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (12) #if (FXAA_GREEN_AS_LUMA == 0) bool twoTapLt = rgby2.w < lumaMin; bool twoTapGt = rgby2.w > lumaMax; #else bool twoTapLt = rgby2.y < lumaMin; bool twoTapGt = rgby2.y > lumaMax; #endif /*--------------------------------------------------------------------------*/ // (13) if(twoTapLt || twoTapGt) rgby2 = rgby1; /*--------------------------------------------------------------------------*/ return rgby2; } /*==========================================================================*/ #endif /*============================================================================ FXAA3 CONSOLE - OPTIMIZED PS3 PIXEL SHADER (WITH EARLY EXIT) ============================================================================== The code mostly matches the assembly. I have a feeling that 14 cycles is possible, but was not able to get there. Might have to increase register count to get full performance. Note this shader does not use perspective interpolation. Use the following cgc options, --fenable-bx2 --fastmath --fastprecision --nofloatbindings Use of FXAA_GREEN_AS_LUMA currently adds a cycle (16 clks). Will look at fixing this for FXAA 3.12. ------------------------------------------------------------------------------ NVSHADERPERF OUTPUT ------------------------------------------------------------------------------ For reference and to aid in debug, output of NVShaderPerf should match this, Shader to schedule: 0: texpkb h0.w(TRUE), v5.zyxx, #0 2: addh h2.y(TRUE), h0.w, constant(0.001953, 0.000000, 0.000000, 0.000000).x 4: texpkb h1.w(TRUE), v5.xwxx, #0 6: addh h0.x(TRUE), h1.w, -h2.y 7: texpkb h2.w(TRUE), v5.zwzz, #0 9: minh h4.w(TRUE), h2.y, h2 10: maxh h5.x(TRUE), h2.y, h2.w 11: texpkb h0.w(TRUE), v5, #0 13: addh h3.w(TRUE), -h0, h0.x 14: addh h0.x(TRUE), h0.w, h0 15: addh h0.z(TRUE), -h2.w, h0.x 16: addh h0.x(TRUE), h2.w, h3.w 17: minh h5.y(TRUE), h0.w, h1.w 18: nrmh h2.xz(TRUE), h0_n 19: minh_m8 h2.w(TRUE), |h2.x|, |h2.z| 20: divx h4.xy(TRUE), h2_n.xzzw, h2_n.w 21: movr r1.zw(TRUE), v4.xxxy 22: maxh h2.w(TRUE), h0, h1 23: fenct TRUE 24: madr r0.xy(TRUE), -h2.xzzw, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).zwzz, r1.zwzz 26: texpkb h0(TRUE), r0, #0 28: maxh h5.x(TRUE), h2.w, h5 29: minh h5.w(TRUE), h5.y, h4 30: madr r1.xy(TRUE), h2.xzzw, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).zwzz, r1.zwzz 32: texpkb h2(TRUE), r1, #0 34: addh_d2 h2(TRUE), h0, h2 35: texpkb h1(TRUE), v4, #0 37: maxh h5.y(TRUE), h5.x, h1.w 38: minh h4.w(TRUE), h1, h5 39: madr r0.xy(TRUE), -h4, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 41: texpkb h0(TRUE), r0, #0 43: addh_m8 h5.z(TRUE), h5.y, -h4.w 44: madr r2.xy(TRUE), h4, constant(cConst5.x, cConst5.y, cConst5.z, cConst5.w).xyxx, r1.zwzz 46: texpkb h3(TRUE), r2, #0 48: addh_d2 h0(TRUE), h0, h3 49: addh_d2 h3(TRUE), h0, h2 50: movh h0(TRUE), h3 51: slth h3.x(TRUE), h3.w, h5.w 52: sgth h3.w(TRUE), h3, h5.x 53: addx.c0 rc(TRUE), h3.x, h3 54: slth.c0 rc(TRUE), h5.z, h5 55: movh h0(c0.NE.w), h2 56: movh h0(c0.NE.x), h1 IPU0 ------ Simplified schedule: -------- Pass | Unit | uOp | PC: Op -----+--------+------+------------------------- 1 | SCT0/1 | mov | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | TEX | txl | 0: TXLr h0.w, g[TEX1].zyxx, const.xxxx, TEX0; | SCB0 | add | 2: ADDh h2.y, h0.-w--, const.-x--; | | | 2 | SCT0/1 | mov | 4: TXLr h1.w, g[TEX1].xwxx, const.xxxx, TEX0; | TEX | txl | 4: TXLr h1.w, g[TEX1].xwxx, const.xxxx, TEX0; | SCB0 | add | 6: ADDh h0.x, h1.w---,-h2.y---; | | | 3 | SCT0/1 | mov | 7: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | TEX | txl | 7: TXLr h2.w, g[TEX1].zwzz, const.xxxx, TEX0; | SCB0 | max | 10: MAXh h5.x, h2.y---, h2.w---; | SCB1 | min | 9: MINh h4.w, h2.---y, h2; | | | 4 | SCT0/1 | mov | 11: TXLr h0.w, g[TEX1], const.xxxx, TEX0; | TEX | txl | 11: TXLr h0.w, g[TEX1], const.xxxx, TEX0; | SCB0 | add | 14: ADDh h0.x, h0.w---, h0; | SCB1 | add | 13: ADDh h3.w,-h0, h0.---x; | | | 5 | SCT0 | mad | 16: ADDh h0.x, h2.w---, h3.w---; | SCT1 | mad | 15: ADDh h0.z,-h2.--w-, h0.--x-; | SCB0 | min | 17: MINh h5.y, h0.-w--, h1.-w--; | | | 6 | SCT1 | mov | 18: NRMh h2.xz, h0; | SRB | nrm | 18: NRMh h2.xz, h0; | SCB1 | min | 19: MINh*8 h2.w, |h2.---x|, |h2.---z|; | | | 7 | SCT0 | div | 20: DIVx h4.xy, h2.xz--, h2.ww--; | SCT1 | mov | 21: MOVr r1.zw, g[TEX0].--xy; | SCB1 | max | 22: MAXh h2.w, h0, h1; | | | 8 | SCT0 | mad | 24: MADr r0.xy,-h2.xz--, const.zw--, r1.zw--; | SCT1 | mov | 26: TXLr h0, r0, const.xxxx, TEX0; | TEX | txl | 26: TXLr h0, r0, const.xxxx, TEX0; | SCB0 | max | 28: MAXh h5.x, h2.w---, h5; | SCB1 | min | 29: MINh h5.w, h5.---y, h4; | | | 9 | SCT0 | mad | 30: MADr r1.xy, h2.xz--, const.zw--, r1.zw--; | SCT1 | mov | 32: TXLr h2, r1, const.xxxx, TEX0; | TEX | txl | 32: TXLr h2, r1, const.xxxx, TEX0; | SCB0/1 | add | 34: ADDh/2 h2, h0, h2; | | | 10 | SCT0/1 | mov | 35: TXLr h1, g[TEX0], const.xxxx, TEX0; | TEX | txl | 35: TXLr h1, g[TEX0], const.xxxx, TEX0; | SCB0 | max | 37: MAXh h5.y, h5.-x--, h1.-w--; | SCB1 | min | 38: MINh h4.w, h1, h5; | | | 11 | SCT0 | mad | 39: MADr r0.xy,-h4, const.xy--, r1.zw--; | SCT1 | mov | 41: TXLr h0, r0, const.zzzz, TEX0; | TEX | txl | 41: TXLr h0, r0, const.zzzz, TEX0; | SCB0 | mad | 44: MADr r2.xy, h4, const.xy--, r1.zw--; | SCB1 | add | 43: ADDh*8 h5.z, h5.--y-,-h4.--w-; | | | 12 | SCT0/1 | mov | 46: TXLr h3, r2, const.xxxx, TEX0; | TEX | txl | 46: TXLr h3, r2, const.xxxx, TEX0; | SCB0/1 | add | 48: ADDh/2 h0, h0, h3; | | | 13 | SCT0/1 | mad | 49: ADDh/2 h3, h0, h2; | SCB0/1 | mul | 50: MOVh h0, h3; | | | 14 | SCT0 | set | 51: SLTh h3.x, h3.w---, h5.w---; | SCT1 | set | 52: SGTh h3.w, h3, h5.---x; | SCB0 | set | 54: SLThc0 rc, h5.z---, h5; | SCB1 | add | 53: ADDxc0_s rc, h3.---x, h3; | | | 15 | SCT0/1 | mul | 55: MOVh h0(NE0.wwww), h2; | SCB0/1 | mul | 56: MOVh h0(NE0.xxxx), h1; Pass SCT TEX SCB 1: 0% 100% 25% 2: 0% 100% 25% 3: 0% 100% 50% 4: 0% 100% 50% 5: 50% 0% 25% 6: 0% 0% 25% 7: 100% 0% 25% 8: 0% 100% 50% 9: 0% 100% 100% 10: 0% 100% 50% 11: 0% 100% 75% 12: 0% 100% 100% 13: 100% 0% 100% 14: 50% 0% 50% 15: 100% 0% 100% MEAN: 26% 60% 56% Pass SCT0 SCT1 TEX SCB0 SCB1 1: 0% 0% 100% 100% 0% 2: 0% 0% 100% 100% 0% 3: 0% 0% 100% 100% 100% 4: 0% 0% 100% 100% 100% 5: 100% 100% 0% 100% 0% 6: 0% 0% 0% 0% 100% 7: 100% 100% 0% 0% 100% 8: 0% 0% 100% 100% 100% 9: 0% 0% 100% 100% 100% 10: 0% 0% 100% 100% 100% 11: 0% 0% 100% 100% 100% 12: 0% 0% 100% 100% 100% 13: 100% 100% 0% 100% 100% 14: 100% 100% 0% 100% 100% 15: 100% 100% 0% 100% 100% MEAN: 33% 33% 60% 86% 80% Fragment Performance Setup: Driver RSX Compiler, GPU RSX, Flags 0x5 Results 15 cycles, 3 r regs, 800,000,000 pixels/s ============================================================================*/ #if (FXAA_PS3 == 1) && (FXAA_EARLY_EXIT == 1) /*--------------------------------------------------------------------------*/ #pragma regcount 7 #pragma disablepc all #pragma option O2 #pragma option OutColorPrec=fp16 #pragma texformat default RGBA8 /*==========================================================================*/ half4 FxaaPixelShader( // See FXAA Quality FxaaPixelShader() source for docs on Inputs! FxaaFloat2 pos, FxaaFloat4 fxaaConsolePosPos, FxaaTex tex, FxaaTex fxaaConsole360TexExpBiasNegOne, FxaaTex fxaaConsole360TexExpBiasNegTwo, FxaaFloat2 fxaaQualityRcpFrame, FxaaFloat4 fxaaConsoleRcpFrameOpt, FxaaFloat4 fxaaConsoleRcpFrameOpt2, FxaaFloat4 fxaaConsole360RcpFrameOpt2, FxaaFloat fxaaQualitySubpix, FxaaFloat fxaaQualityEdgeThreshold, FxaaFloat fxaaQualityEdgeThresholdMin, FxaaFloat fxaaConsoleEdgeSharpness, FxaaFloat fxaaConsoleEdgeThreshold, FxaaFloat fxaaConsoleEdgeThresholdMin, FxaaFloat4 fxaaConsole360ConstDir ) { /*--------------------------------------------------------------------------*/ // (1) half4 rgbyNe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaNe = rgbyNe.w + half(1.0/512.0); #else half lumaNe = rgbyNe.y + half(1.0/512.0); #endif /*--------------------------------------------------------------------------*/ // (2) half4 lumaSw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaSwNegNe = lumaSw.w - lumaNe; #else half lumaSwNegNe = lumaSw.y - lumaNe; #endif /*--------------------------------------------------------------------------*/ // (3) half4 lumaNw = h4tex2Dlod(tex, half4(fxaaConsolePosPos.xy, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaMaxNwSw = max(lumaNw.w, lumaSw.w); half lumaMinNwSw = min(lumaNw.w, lumaSw.w); #else half lumaMaxNwSw = max(lumaNw.y, lumaSw.y); half lumaMinNwSw = min(lumaNw.y, lumaSw.y); #endif /*--------------------------------------------------------------------------*/ // (4) half4 lumaSe = h4tex2Dlod(tex, half4(fxaaConsolePosPos.zw, 0, 0)); #if (FXAA_GREEN_AS_LUMA == 0) half dirZ = lumaNw.w + lumaSwNegNe; half dirX = -lumaNw.w + lumaSwNegNe; #else half dirZ = lumaNw.y + lumaSwNegNe; half dirX = -lumaNw.y + lumaSwNegNe; #endif /*--------------------------------------------------------------------------*/ // (5) half3 dir; dir.y = 0.0; #if (FXAA_GREEN_AS_LUMA == 0) dir.x = lumaSe.w + dirX; dir.z = -lumaSe.w + dirZ; half lumaMinNeSe = min(lumaNe, lumaSe.w); #else dir.x = lumaSe.y + dirX; dir.z = -lumaSe.y + dirZ; half lumaMinNeSe = min(lumaNe, lumaSe.y); #endif /*--------------------------------------------------------------------------*/ // (6) half4 dir1_pos; dir1_pos.xy = normalize(dir).xz; half dirAbsMinTimes8 = min(abs(dir1_pos.x), abs(dir1_pos.y)) * half(FXAA_CONSOLE_PS3_EDGE_SHARPNESS); /*--------------------------------------------------------------------------*/ // (7) half4 dir2_pos; dir2_pos.xy = clamp(dir1_pos.xy / dirAbsMinTimes8, half(-2.0), half(2.0)); dir1_pos.zw = pos.xy; dir2_pos.zw = pos.xy; #if (FXAA_GREEN_AS_LUMA == 0) half lumaMaxNeSe = max(lumaNe, lumaSe.w); #else half lumaMaxNeSe = max(lumaNe, lumaSe.y); #endif /*--------------------------------------------------------------------------*/ // (8) half4 temp1N; temp1N.xy = dir1_pos.zw - dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; temp1N = h4tex2Dlod(tex, half4(temp1N.xy, 0.0, 0.0)); half lumaMax = max(lumaMaxNwSw, lumaMaxNeSe); half lumaMin = min(lumaMinNwSw, lumaMinNeSe); /*--------------------------------------------------------------------------*/ // (9) half4 rgby1; rgby1.xy = dir1_pos.zw + dir1_pos.xy * fxaaConsoleRcpFrameOpt.zw; rgby1 = h4tex2Dlod(tex, half4(rgby1.xy, 0.0, 0.0)); rgby1 = (temp1N + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (10) half4 rgbyM = h4tex2Dlod(tex, half4(pos.xy, 0.0, 0.0)); #if (FXAA_GREEN_AS_LUMA == 0) half lumaMaxM = max(lumaMax, rgbyM.w); half lumaMinM = min(lumaMin, rgbyM.w); #else half lumaMaxM = max(lumaMax, rgbyM.y); half lumaMinM = min(lumaMin, rgbyM.y); #endif /*--------------------------------------------------------------------------*/ // (11) half4 temp2N; temp2N.xy = dir2_pos.zw - dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; temp2N = h4tex2Dlod(tex, half4(temp2N.xy, 0.0, 0.0)); half4 rgby2; rgby2.xy = dir2_pos.zw + dir2_pos.xy * fxaaConsoleRcpFrameOpt2.zw; half lumaRangeM = (lumaMaxM - lumaMinM) / FXAA_CONSOLE_PS3_EDGE_THRESHOLD; /*--------------------------------------------------------------------------*/ // (12) rgby2 = h4tex2Dlod(tex, half4(rgby2.xy, 0.0, 0.0)); rgby2 = (temp2N + rgby2) * 0.5; /*--------------------------------------------------------------------------*/ // (13) rgby2 = (rgby2 + rgby1) * 0.5; /*--------------------------------------------------------------------------*/ // (14) #if (FXAA_GREEN_AS_LUMA == 0) bool twoTapLt = rgby2.w < lumaMin; bool twoTapGt = rgby2.w > lumaMax; #else bool twoTapLt = rgby2.y < lumaMin; bool twoTapGt = rgby2.y > lumaMax; #endif bool earlyExit = lumaRangeM < lumaMax; bool twoTap = twoTapLt || twoTapGt; /*--------------------------------------------------------------------------*/ // (15) if(twoTap) rgby2 = rgby1; if(earlyExit) rgby2 = rgbyM; /*--------------------------------------------------------------------------*/ return rgby2; } /*==========================================================================*/ #endif #endif ////#line 1 #ifndef SRGB_DEFINED #define SRGB_DEFINED #if HAS_SRGB vec3 to_srgb(vec3 lin) { return mix( 1.055 * pow(lin,vec3(0.41666)) - 0.055, 12.92 * lin, vec3(lessThanEqual(lin,vec3(0.0031308)))); } vec3 to_linear(vec3 srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),vec3(2.4)), srgb * (1.0 / 12.92), vec3(lessThanEqual(srgb,vec3(0.04045)))); } float to_srgb(float lin) { return mix( 1.055 * pow(lin,0.41666) - 0.055, 12.92 * lin, lin <= 0.0031308 ? 1.0 : 0.0); } float to_linear(float srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),2.4), srgb * (1.0 / 12.92), srgb <= 0.04045 ? 1.0 : 0.0); } #else vec3 to_linear(vec3 srgb) { return pow(srgb,vec3(2.2)); } vec3 to_srgb(vec3 lin) { return pow(lin,vec3(0.454545)); } float to_linear(float srgb) { return pow(srgb,2.2); } float to_srgb(float lin) { return pow(lin,0.454545); } #endif #endif /* SRGB_DEFINED */ //#line 27 VARYING(vec4, v_screen_pos) VARYING(vec4, v_color) #if TONE_MAP ////#line 1 #ifndef WANT_LIGHTMETER #define WANT_LIGHTMETER 0 #endif //////#line 1 #ifndef SRGB_DEFINED #define SRGB_DEFINED #if HAS_SRGB vec3 to_srgb(vec3 lin) { return mix( 1.055 * pow(lin,vec3(0.41666)) - 0.055, 12.92 * lin, vec3(lessThanEqual(lin,vec3(0.0031308)))); } vec3 to_linear(vec3 srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),vec3(2.4)), srgb * (1.0 / 12.92), vec3(lessThanEqual(srgb,vec3(0.04045)))); } float to_srgb(float lin) { return mix( 1.055 * pow(lin,0.41666) - 0.055, 12.92 * lin, lin <= 0.0031308 ? 1.0 : 0.0); } float to_linear(float srgb) { return mix( pow(srgb * (1.0/1.055f) + (0.055f/1.055),2.4), srgb * (1.0 / 12.92), srgb <= 0.04045 ? 1.0 : 0.0); } #else vec3 to_linear(vec3 srgb) { return pow(srgb,vec3(2.2)); } vec3 to_srgb(vec3 lin) { return pow(lin,vec3(0.454545)); } float to_linear(float srgb) { return pow(srgb,2.2); } float to_srgb(float lin) { return pow(lin,0.454545); } #endif #endif /* SRGB_DEFINED */ ////#line 8 #if !defined(VSHADER_TONE_ONLY) VARYING(float, v_tonemap_gain) #endif uniform sampler2D u_tex_tone; #define bias 0.75 #if WANT_LIGHTMETER VARYING(vec2, amp_meter) #endif #ifndef MODE_TONE_DYNAMIC #define MODE_TONE_DYNAMIC 2 #endif #ifndef TONE_MODE #define TONE_MODE 2 #endif #if TONE_MODE == MODE_TONE_DYNAMIC #define tonemap_gain u_tonemap_gain #else #define tonemap_gain vec4(0, 0, 0, 3.2) #endif float A = 0.15; float B = 0.50; float C = 0.10; float D = 0.20; float E = 0.02; float F = 0.30; vec3 W = vec3(3.2); vec3 Uncharted2Tonemap(vec3 x) { return ((x*(A*x+C*B)+D*E)/(x*(A*x+B)+D*F))-E/F; } float R_white = 3.0; vec3 Reinhard(vec3 L) { vec3 num = L * (L/(R_white*R_white) + 1.0); vec3 den = L + 1.0; return num/den; } // If we are going to tone map using Alex's epic equation, // go to sRGB-ish space first; his stuff is tuned to sRGB, // not linear. void tone_map_uncharted(inout vec3 rgb, in float gain) { // dumb clamping // rgb = clamp(rgb,0.0,1.0); // naive coolor Reinhardt - needs srgb // rgb = Reinhard(rgb); // Dawson - no srgb // vec3 x = max(rgb-0.004,0.0); // rgb = (x*(6.2*x+.5))/(x*(6.2*x+1.7)+0.06); // uncharted 2 float ExposureBias = gain; vec3 curr = Uncharted2Tonemap(ExposureBias*rgb); vec3 whiteScale = 1.0f/Uncharted2Tonemap(vec3(tonemap_gain.w)); rgb = curr*whiteScale; } void tone_map_gifford(inout vec3 rgb, in float gain) { // rgb = to_srgb(rgb); vec3 blend_rat = pow(clamp(rgb,0.0,1.0),vec3(bias)); rgb = mix(rgb,1.0 - pow(clamp(1.0 - rgb,0.0,1.0), vec3(gain)),blend_rat); } #if VSHADER #if !defined(VSHADER_TONE_ONLY) void calc_light_meter() { // Tone mapping light meter. // This samples the dynamic exposure texture (a 1x1 result of much copying) // and figures out how much gain to apply. Hoisted up into the vertex shader // to cut ops out of fragment processing - the tone map levels do not vary // by frame position!! // The light meter texture contains the _immediate_ brightness in "r" and // a blended value in "g" - the result is that g assymptotically r. // By sampling the max of both, we take the brighter of what we see now or // what we did see, for a fast adjustment when things get bright but slow // adjustment when things get dim. vec2 light_meter_raw = texture2DLod(u_tex_tone,vec2(0.5,0.5),0).rg; float overall_color = max(light_meter_raw.r,light_meter_raw.g); float amp = (overall_color); float amp_lim = clamp(amp,tonemap_gain.y,tonemap_gain.z); float total_gain = tonemap_gain.x / amp_lim; v_tonemap_gain = total_gain; #if WANT_LIGHTMETER amp_meter = vec2(amp,amp_lim); #endif } #endif void vshader_tone_map(inout vec3 rgb) { vec2 light_meter_raw = texture2DLod(u_tex_tone,vec2(0.5,0.5),0).rg; float overall_color = max(light_meter_raw.r,light_meter_raw.g); float amp = (overall_color); float amp_lim = clamp(amp,tonemap_gain.y,tonemap_gain.z); float total_gain = tonemap_gain.x / amp_lim; tone_map_uncharted(rgb, total_gain); } void vshader_tone_map_srgb(inout vec3 rgb) { rgb = to_linear(rgb); vshader_tone_map(rgb); rgb = to_srgb(rgb); } #endif #if FSHADER && !defined(VSHADER_TONE_ONLY) void tone_map(inout vec3 rgb) { tone_map_uncharted(rgb, v_tonemap_gain); // And now...Alex's epic equation. v_tonemap_gain is a gain // ratio (e.g. 1.0 is unity) and bias is a fixed constant // for where in the brightness curve we start to squash. // Generally lower biases squash lower brightness levels and // as a result make tings more, um, "crispy". // Light meter code - just for debugging - not particularly // well optimized. // How to read the light meter! // The yellow interval shows the range of clamping for the light // meter's input. This effectively limits how much it can increase // or decrease the gain (since the exposure is lowered by the // brightness of the input image). A wider range means dynamic // exposure can do more, and more space to the left means darks are // brought up more (because the system will go "whoa, the input image // is rully, rully dark) and crank it. // // The top bar is the light meter - left is darker, right is brighter. // The bar is green when it is within the unclamped range and is thus // affecting exposure; it is red when it is out of range and thus // small changes to brightness don't change exposure. // The bottom bar is the total gain applied to the frame, taking into // account both the fixed gain adjustment from skyc.cpp and the tone // mapping adjustment from the level meter. This bar is green when // the image is being made brighter and red when it is being made // darker. #if WANT_LIGHTMETER if(gl_FragCoord.y < 20) { float total_gain_log = log2(v_tonemap_gain); float amp_percent = (total_gain_log / 16.0) + 0.5; float pixel_pos = gl_FragCoord.x / 512.0; if(amp_percent < 0.5) { if(amp_percent > pixel_pos) rgb = vec3(1,0,0); } else { if(amp_percent > pixel_pos) rgb = vec3(0,1,0); } } else if (gl_FragCoord.y < 40) { float amp_log = log2(amp_meter.x); float amp_lim_log = log2(amp_meter.y); float meter_percent = (amp_log / 16.0) + 0.5; float pixel_pos = gl_FragCoord.x / 512.0; if(amp_lim_log != amp_log) { if(meter_percent > pixel_pos) rgb = vec3(1,0,0); } else { if(meter_percent > pixel_pos) rgb = vec3(0,1,0); } } else if(gl_FragCoord.y < 48) { vec2 bounds = log2(tonemap_gain.yz); vec2 range = vec2(bounds / 16.0) + 0.5; float pixel_pos = gl_FragCoord.x / 512.0; if(pixel_pos > range.x && pixel_pos < range.y) { rgb = vec3(1,1,0); } } #endif } void tone_map_srgb(inout vec3 rgb) { rgb = to_linear(rgb); tone_map(rgb); rgb = to_srgb(rgb); } void tone_map_srgb_to_linear(inout vec3 rgb) { rgb = to_linear(rgb); tone_map(rgb); } void tone_map_linear_to_srgb(inout vec3 rgb) { tone_map(rgb); rgb = to_srgb(rgb); } #endif #endif //#line 48 #if VSHADER ATTRIBUTE(vec4, a_color) ATTRIBUTE(vec4, a_vertex) void main (void) { STD_BROADCAST #if HAS_FULL_TRANSFORM gl_Position = u_mvp_matrix * (a_vertex - vec4(u_mv_offset,0.0)); #else gl_Position = a_vertex; #endif v_color = a_color; #if TONE_MAP calc_light_meter(); #endif } #endif #if FSHADER uniform sampler2D u_tex1; uniform sampler2D u_tex2; #define K 1.0 vec4 final_color_correct(vec4 in_color) { vec3 rgb = in_color.rgb; #if TONE_MAP tone_map(rgb); #endif return vec4(rgb,in_color.a); } void main (void) { #if MODE==MODE_BLUR gl_FragColor = u_kernel2.w * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - 7.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.z * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - 6.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.y * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - 5.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.x * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - 4.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel1.w * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - 3.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel1.z * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - 2.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel1.y * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) - (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel1.x * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) ) + u_kernel1.y * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel1.z * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + 2.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel1.w * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + 3.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.x * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + 4.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.y * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + 5.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.z * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + 6.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K) + u_kernel2.w * texture2D(u_tex1, (v_texcoord0.st * vec2(0.5, 1.0) + (uViewportTarget == 0 ? vec2(0.0, 0.0) : vec2(0.5, 0.0))) + 7.0 * (uFiScreenInv * vec2(0.5, 1.0)) * K); #endif #if MODE==MODE_COPY gl_FragColor = v_color * final_color_correct(texture2D(u_tex1, mod(gl_FragCoord.xy, uFiScreenSize) * (uFiScreenInv * vec2(0.5, 1.0)))); #endif #if MODE==MODE_GRAY // Grayscale shader. v_color's rgb is the weight ratios to convert to gray - usually // mostly green since that's how our eyes work. vec4 rgba = texture2D(u_tex1, mod(gl_FragCoord.xy, uFiScreenSize) * (uFiScreenInv * vec2(0.5, 1.0))); rgba.rgb = vec3(dot(v_color.rgb,rgba.rgb)); rgba = final_color_correct(rgba); rgba.a *= v_color.a; gl_FragColor = rgba; #endif #if MODE==MODE_HEAT vec2 screen_rat = mod(gl_FragCoord.xy, uFiScreenSize) * (uFiScreenInv * vec2(0.5, 1.0)); float t1 = fract(u_time); t1 = (t1 > 0.5) ? (1.0 - t1) : t1; t1 *= 2.0; float t2 = 1.0 - t1; vec2 noise_st1 = texture2D(u_tex2, screen_rat*0.5 +vec2(0,u_time)).rg * 2.0 - vec2(1.0); vec2 noise_st2 = texture2D(u_tex2, screen_rat*0.5 + vec2(0.5)+vec2(u_time,0)).rg * 2.0 - vec2(1.0); vec2 uv = screen_rat + noise_st1 * 0.005 * t1 + noise_st2 * 0.005 * t2; vec3 blur = texture2D(u_tex1, uv).rgb; gl_FragColor = vec4(blur,v_color.a); gl_FragColor = final_color_correct(gl_FragColor); #endif #if MODE==MODE_FXAA gl_FragColor = vec4(0.0); vec2 tex_ctr = mod(gl_FragCoord.xy, uFiScreenSize) * ((uFiScreenInv * vec2(0.5, 1.0)) * u_super_dim.xy); for(int x = 0; x < u_super_dim.x; ++x) for(int y = 0; y < u_super_dim.y; ++y) { vec2 sub_pos = tex_ctr + vec2(x,y) * (uFiScreenInv * vec2(0.5, 1.0)) - u_super_dim.zw; gl_FragColor += FxaaPixelShader( sub_pos, vec4(0.0), // console pos u_tex1, // tex u_tex1, // console tex-1 u_tex1, // console tex-2 (uFiScreenInv * vec2(0.5, 1.0)), // 1/screen size vec4(0.0), // console screen offsets vec4(0.0), // console screen offsets vec4(0.0), // console screen offsets 0.75, // sub px qual 0.166, // edge thesh 0.0833, // min thresh 0.0, // sharpness (console) 0.05, // console edge 0.0, // console edge min vec4(0.0)); // console directional constants } gl_FragColor *= (1.0 / (u_super_dim.x * u_super_dim.y)); gl_FragColor = final_color_correct(gl_FragColor); gl_FragColor.a = 1.0; #endif } #endif 0:00:48.445 I/FLT: Init dat_p0 type:loc_ram apt:L77 start at:Ramp 0:00:50.423 I/SCN: DSF load time: 761286 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+34-116.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 786322 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+32-116.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 1133225 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+32-118.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 3875862 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+34-118.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 4420497 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+32-118.dsf (99996 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 4886356 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+34-116.dsf (178696 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 5211907 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+32-116.dsf (204330 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 34886 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+34-115.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 2901513 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+34-118.dsf (246101 tris, 3 skipped for -3.6 m^2) 0:00:50.423 I/SCN: DSF load time: 21485 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+32-115.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 1297051 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+34-115.dsf (206201 tris, 3 skipped for -1.4 m^2) 0:00:50.423 I/SCN: DSF load time: 51032 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+32-117.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 1288155 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+32-115.dsf (209284 tris, 1 skipped for -7.7 m^2) 0:00:50.423 I/SCN: DSF load time: 992 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+33-115.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 80911 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+34-117.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 1630090 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+32-117.dsf (186031 tris, 3 skipped for -10.4 m^2) 0:00:50.423 I/SCN: DSF load time: 1172048 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+33-115.dsf (194302 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 1234378 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+34-117.dsf (176694 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 2441 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+33-116.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 232472 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+33-118.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 852462 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+33-116.dsf (170947 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 2783891 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+33-118.dsf (258475 tris, 6 skipped for -14.6 m^2) 0:00:50.423 I/SCN: DSF load time: 237680 for file Custom Scenery/KTRM - Jacqueline Cochran Regional Airport/Earth nav data/+30-120/+33-117.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 324765 for file Custom Scenery/Global Airports/Earth nav data/+30-120/+33-117.dsf (0 tris, 0 skipped for 0.0 m^2) 0:00:50.423 I/SCN: DSF load time: 1495804 for file Global Scenery/X-Plane 11 Global Scenery/Earth nav data/+30-120/+33-117.dsf (195729 tris, 0 skipped for 0.0 m^2) 0:01:15.550 I/OVR: Reading VR conf file: Aircraft/Laminar Research/Cessna 172SP/Cessna_172SP_G1000_vrconfig.txt 0:01:15.550 D/STM: Transitioning from state state_NormalLoading to state_Normal 0:01:15.550 I/OVR: User is leaving the holodeck. 0:03:45.555 E/SYS: +------------------------------------------------------------------------------- 0:03:45.555 E/SYS: | There was a problem processing a GLSL pixel shader. 0:03:45.555 E/SYS: | You may need to rerun the installer or update your graphics drivers. 0:03:45.555 E/SYS: | Resources/shaders/hdr.glsl 0:03:45.555 E/SYS: | Resources/shaders/hdr.glsl 0:03:45.555 E/SYS: | (OGL_shader.cpp:1059) 0:03:45.555 E/SYS: +------------------------------------------------------------------------------- 0:03:45.555 D/STM: Transitioning from state state_Normal to state_ShuttingDown G64: 221.081: Plugin Disabled. [SKUNKCRAFTS UPDATER INFO]: Disabling... [SKUNKCRAFTS UPDATER INFO]: [LUA]: "module" SkunkCrafts Updater stopped [SKUNKCRAFTS UPDATER INFO]: SASL plugin stopped... 0:03:45.555 I/JOY: Control response for axis joy_use_ptch: 0.500000 0:03:45.555 I/JOY: Control response for axis joy_use_roll: 0.500000 0:03:45.555 I/JOY: Control response for axis joy_use_hdng: 0.500000 0:03:45.555 I/JOY: Control response for axis joy_use_thro: 0.000000 0:03:45.555 I/JOY: Control response for axis joy_use_lbrk: 0.000000 0:03:45.555 I/JOY: Control response for axis joy_use_rbrk: 0.000000 0:03:45.555 I/JOY: Control response for axis joy_use_mixt: 0.000000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_ptch: 0.500000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_roll: 0.500000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_hdng: 0.500000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_thro: 0.000000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_lbrk: 0.000000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_rbrk: 0.000000 0:03:45.555 I/JOY: Stability augmentation for axis joy_use_mixt: 0.000000 0:03:45.555 I/JOY: UNREGISTER Joystick device: Control Manager ID #00 - VID:1678PID:49169 0:03:45.555 I/JOY: UNREGISTER Joystick device: Control Manager ID #00 - VID:1678PID:49170 0:03:45.555 I/JOY: UNREGISTER Joystick device: Control Manager ID #00 - VID:1678PID:49168 0:03:45.555 D/HID: HID Bridge Shutdown ----- X-Plane has shut down -----