From 42b7739321c97bde45648b82d5e8e358a4ab2354 Mon Sep 17 00:00:00 2001 From: Hugo Locurcio Date: Fri, 3 Feb 2023 12:33:17 +0100 Subject: Document unsupported features in Mobile and Compatibility rendering methods --- doc/classes/@GlobalScope.xml | 2 +- doc/classes/BaseMaterial3D.xml | 44 ++++++++++++++++--------------- doc/classes/CameraAttributesPhysical.xml | 1 + doc/classes/CameraAttributesPractical.xml | 2 ++ doc/classes/Control.xml | 2 +- doc/classes/Decal.xml | 1 + doc/classes/DisplayServer.xml | 7 ++--- doc/classes/Environment.xml | 14 +++++++--- doc/classes/Light3D.xml | 7 +++-- doc/classes/NinePatchRect.xml | 2 -- doc/classes/OmniLight3D.xml | 1 + doc/classes/ProjectSettings.xml | 21 ++++++++++++--- doc/classes/ReflectionProbe.xml | 1 + doc/classes/RenderingDevice.xml | 2 +- doc/classes/SpotLight3D.xml | 1 + doc/classes/TextureLayered.xml | 2 +- doc/classes/VoxelGI.xml | 1 + 17 files changed, 71 insertions(+), 40 deletions(-) (limited to 'doc/classes') diff --git a/doc/classes/@GlobalScope.xml b/doc/classes/@GlobalScope.xml index 55810156a1..a8860721d6 100644 --- a/doc/classes/@GlobalScope.xml +++ b/doc/classes/@GlobalScope.xml @@ -2818,7 +2818,7 @@ When duplicating a resource with [method Resource.duplicate], and this flag is set on a property of that resource, the property should never be duplicated, regardless of the [code]subresources[/code] bool parameter. - The property is only shown in the editor if modern renderers are supported (GLES3 is excluded). + The property is only shown in the editor if modern renderers are supported (the Compatibility rendering method is excluded). diff --git a/doc/classes/BaseMaterial3D.xml b/doc/classes/BaseMaterial3D.xml index 60225275f3..e72aa373f7 100644 --- a/doc/classes/BaseMaterial3D.xml +++ b/doc/classes/BaseMaterial3D.xml @@ -215,10 +215,11 @@ If [code]true[/code], the object is rendered at the same size regardless of distance. - If [code]true[/code], enables the vertex grow setting. See [member grow_amount]. + If [code]true[/code], enables the vertex grow setting. This can be used to create mesh-based outlines using a second material pass and its [member cull_mode] set to [constant CULL_FRONT]. See also [member grow_amount]. + [b]Note:[/b] Vertex growth cannot create new vertices, which means that visible gaps may occur in sharp corners. This can be alleviated by designing the mesh to use smooth normals exclusively using [url=https://wiki.polycount.com/wiki/Face_weighted_normals]face weighted normals[/url] in the 3D authoring software. In this case, grow will be able to join every outline together, just like in the original mesh. - Grows object vertices in the direction of their normals. + Grows object vertices in the direction of their normals. Only effective if [member grow] is [code]true[/code]. If [code]true[/code], uses parallax occlusion mapping to represent depth in the material instead of simple offset mapping (see [member heightmap_enabled]). This results in a more convincing depth effect, but is much more expensive on the GPU. Only enable this on materials where it makes a significant visual difference. @@ -277,7 +278,7 @@ If [code]true[/code], depth testing is disabled and the object will be drawn in render order. - If [code]true[/code], normal mapping is enabled. + If [code]true[/code], normal mapping is enabled. This has a slight performance cost, especially on mobile GPUs. The strength of the normal map's effect. @@ -344,7 +345,8 @@ Specifies the channel of the [member roughness_texture] in which the roughness information is stored. This is useful when you store the information for multiple effects in a single texture. For example if you stored metallic in the red channel, roughness in the blue, and ambient occlusion in the green you could reduce the number of textures you use. - Sets whether the shading takes place per-pixel or per-vertex. Per-vertex lighting is faster, making it the best choice for mobile applications, however it looks considerably worse than per-pixel. + Sets whether the shading takes place, per-pixel, per-vertex or unshaded. Per-vertex lighting is faster, making it the best choice for mobile applications, however it looks considerably worse than per-pixel. Unshaded rendering is the fastest, but disables all interactions with lights. + [b]Note:[/b] Setting the shading mode vertex shading currently has no effect, as vertex shading is not implemented yet. If [code]true[/code], enables the "shadow to opacity" render mode where lighting modifies the alpha so shadowed areas are opaque and non-shadowed areas are transparent. Useful for overlaying shadows onto a camera feed in AR. @@ -388,7 +390,7 @@ Repeat flags for the texture. See [enum TextureFilter] for options. - If [code]true[/code], transparency is enabled on the body. See also [member blend_mode]. + If [code]true[/code], transparency is enabled on the body. Some transparency modes will disable shadow casting. Any transparency mode other than Disabled has a greater performance impact compared to opaque rendering. See also [member blend_mode]. If [code]true[/code], enables parts of the shader required for [GPUParticles3D] trails to function. This also requires using a mesh with appropriate skinning, such as [RibbonTrailMesh] or [TubeTrailMesh]. Enabling this feature outside of materials used in [GPUParticles3D] meshes will break material rendering. @@ -431,7 +433,7 @@ If [code]true[/code], vertex colors are considered to be stored in sRGB color space and are converted to linear color space during rendering. If [code]false[/code], vertex colors are considered to be stored in linear color space and are rendered as-is. See also [member albedo_texture_force_srgb]. - [b]Note:[/b] Only effective when using the Vulkan Clustered or Vulkan Mobile backends. + [b]Note:[/b] Only effective when using the Forward+ and Mobile rendering methods, not Compatibility. If [code]true[/code], the vertex color is used as albedo color. @@ -523,31 +525,31 @@ Use [code]UV2[/code] with the detail texture. - The material will not use transparency. + The material will not use transparency. This is the fastest to render. - The material will use the texture's alpha values for transparency. + The material will use the texture's alpha values for transparency. This is the slowest to render, and disables shadow casting. - The material will cut off all values below a threshold, the rest will remain opaque. The opaque portions will be rendered in the depth prepass. + The material will cut off all values below a threshold, the rest will remain opaque. The opaque portions will be rendered in the depth prepass. This is faster to render than alpha blending, but slower than opaque rendering. This also supports casting shadows. - The material will cut off all values below a spatially-deterministic threshold, the rest will remain opaque. + The material will cut off all values below a spatially-deterministic threshold, the rest will remain opaque. This is faster to render than alpha blending, but slower than opaque rendering. This also supports casting shadows. Alpha hashing is suited for hair rendering. - The material will use the texture's alpha value for transparency, but will discard fragments with an alpha of less than 0.99 during the depth prepass and fragments with an alpha less than 0.1 during the shadow pass. + The material will use the texture's alpha value for transparency, but will discard fragments with an alpha of less than 0.99 during the depth prepass and fragments with an alpha less than 0.1 during the shadow pass. This also supports casting shadows. Represents the size of the [enum Transparency] enum. - The object will not receive shadows. + The object will not receive shadows. This is the fastest to render, but it disables all interactions with lights. - The object will be shaded per pixel. Useful for realistic shading effect. + The object will be shaded per pixel. Useful for realistic shading effects. - The object will be shaded per vertex. Useful when you want cheaper shaders and do not care about visual quality. + The object will be shaded per vertex. Useful when you want cheaper shaders and do not care about visual quality. Not implemented yet (this mode will act like [constant SHADING_MODE_PER_PIXEL]). Represents the size of the [enum ShadingMode] enum. @@ -623,13 +625,13 @@ Objects will not write their depth to the depth buffer, even during the depth prepass (if enabled). - Default cull mode. The back of the object is culled when not visible. Back face triangles will be culled when facing the camera. This results in only the front side of triangles being drawn. For closed-surface meshes this means that only the exterior of the mesh will be visible. + Default cull mode. The back of the object is culled when not visible. Back face triangles will be culled when facing the camera. This results in only the front side of triangles being drawn. For closed-surface meshes, this means that only the exterior of the mesh will be visible. - Front face triangles will be culled when facing the camera. This results in only the back side of triangles being drawn. For closed-surface meshes this means that the interior of the mesh will be drawn instead of the exterior. + Front face triangles will be culled when facing the camera. This results in only the back side of triangles being drawn. For closed-surface meshes, this means that the interior of the mesh will be drawn instead of the exterior. - No culling is performed. + No face culling is performed; both the front face and back face will be visible. Disables the depth test, so this object is drawn on top of all others drawn before it. This puts the object in the transparent draw pass where it is sorted based on distance to camera. Objects drawn after it in the draw order may cover it. This also disables writing to depth. @@ -639,7 +641,7 @@ Vertex colors are considered to be stored in sRGB color space and are converted to linear color space during rendering. See also [member vertex_color_is_srgb]. - [b]Note:[/b] Only effective when using the Vulkan Clustered or Vulkan Mobile backends. + [b]Note:[/b] Only effective when using the Forward+ and Mobile rendering methods. Uses point size to alter the size of primitive points. Also changes the albedo texture lookup to use [code]POINT_COORD[/code] instead of [code]UV[/code]. @@ -717,7 +719,7 @@ Toon blob which changes size based on roughness. - No specular blob. + No specular blob. This is slightly faster to render than other specular modes. Billboard mode is disabled. @@ -760,10 +762,10 @@ Smoothly fades the object out based on each pixel's distance from the camera using the alpha channel. - Smoothly fades the object out based on each pixel's distance from the camera using a dither approach. Dithering discards pixels based on a set pattern to smoothly fade without enabling transparency. On certain hardware this can be faster than [constant DISTANCE_FADE_PIXEL_ALPHA]. + Smoothly fades the object out based on each pixel's distance from the camera using a dithering approach. Dithering discards pixels based on a set pattern to smoothly fade without enabling transparency. On certain hardware, this can be faster than [constant DISTANCE_FADE_PIXEL_ALPHA]. - Smoothly fades the object out based on the object's distance from the camera using a dither approach. Dithering discards pixels based on a set pattern to smoothly fade without enabling transparency. On certain hardware this can be faster than [constant DISTANCE_FADE_PIXEL_ALPHA]. + Smoothly fades the object out based on the object's distance from the camera using a dithering approach. Dithering discards pixels based on a set pattern to smoothly fade without enabling transparency. On certain hardware, this can be faster than [constant DISTANCE_FADE_PIXEL_ALPHA] and [constant DISTANCE_FADE_PIXEL_DITHER]. diff --git a/doc/classes/CameraAttributesPhysical.xml b/doc/classes/CameraAttributesPhysical.xml index a61e735932..9512fd11a8 100644 --- a/doc/classes/CameraAttributesPhysical.xml +++ b/doc/classes/CameraAttributesPhysical.xml @@ -7,6 +7,7 @@ [CameraAttributesPhysical] is used to set rendering settings based on a physically-based camera's settings. It is responsible for exposure, auto-exposure, and depth of field. When used in a [WorldEnvironment] it provides default settings for exposure, auto-exposure, and depth of field that will be used by all cameras without their own [CameraAttributes], including the editor camera. When used in a [Camera3D] it will override any [CameraAttributes] set in the [WorldEnvironment] and will override the [Camera3D]s [member Camera3D.far], [member Camera3D.near], [member Camera3D.fov], and [member Camera3D.keep_aspect] properties. When used in [VoxelGI] or [LightmapGI], only the exposure settings will be used. The default settings are intended for use in an outdoor environment, tips for settings for use in an indoor environment can be found in each setting's documentation. + [b]Note:[/b] Depth of field blur is only supported in the Forward+ and Mobile rendering methods, not Compatibility. diff --git a/doc/classes/CameraAttributesPractical.xml b/doc/classes/CameraAttributesPractical.xml index 924b02fc79..cd914c9eb7 100644 --- a/doc/classes/CameraAttributesPractical.xml +++ b/doc/classes/CameraAttributesPractical.xml @@ -24,6 +24,7 @@ Enables depth of field blur for objects further than [member dof_blur_far_distance]. Strength of blur is controlled by [member dof_blur_amount] and modulated by [member dof_blur_far_transition]. + [b]Note:[/b] Depth of field blur is only supported in the Forward+ and Mobile rendering methods, not Compatibility. When positive, distance over which (starting from [member dof_blur_far_distance]) blur effect will scale from 0 to [member dof_blur_amount]. When negative, uses physically-based scaling so depth of field effect will scale from 0 at [member dof_blur_far_distance] and will increase in a physically accurate way as objects get further from the [Camera3D]. @@ -33,6 +34,7 @@ Enables depth of field blur for objects closer than [member dof_blur_near_distance]. Strength of blur is controlled by [member dof_blur_amount] and modulated by [member dof_blur_near_transition]. + [b]Note:[/b] Depth of field blur is only supported in the Forward+ and Mobile rendering methods, not Compatibility. When positive, distance over which blur effect will scale from 0 to [member dof_blur_amount], ending at [member dof_blur_near_distance]. When negative, uses physically-based scaling so depth of field effect will scale from 0 at [member dof_blur_near_distance] and will increase in a physically accurate way as objects get closer to the [Camera3D]. diff --git a/doc/classes/Control.xml b/doc/classes/Control.xml index 41c5b4fbe2..f45ddf2738 100644 --- a/doc/classes/Control.xml +++ b/doc/classes/Control.xml @@ -558,7 +558,7 @@ Steal the focus from another control and become the focused control (see [member focus_mode]). - [b]Note[/b]: Using this method together with [method Callable.call_deferred] makes it more reliable, especially when called inside [method Node._ready]. + [b]Note:[/b] Using this method together with [method Callable.call_deferred] makes it more reliable, especially when called inside [method Node._ready]. diff --git a/doc/classes/Decal.xml b/doc/classes/Decal.xml index b63f6e7252..8f2303cc9d 100644 --- a/doc/classes/Decal.xml +++ b/doc/classes/Decal.xml @@ -8,6 +8,7 @@ They are made of an [AABB] and a group of [Texture2D]s specifying [Color], normal, ORM (ambient occlusion, roughness, metallic), and emission. Decals are projected within their [AABB] so altering the orientation of the Decal affects the direction in which they are projected. By default, Decals are projected down (i.e. from positive Y to negative Y). The [Texture2D]s associated with the Decal are automatically stored in a texture atlas which is used for drawing the decals so all decals can be drawn at once. Godot uses clustered decals, meaning they are stored in cluster data and drawn when the mesh is drawn, they are not drawn as a post-processing effect after. [b]Note:[/b] Decals cannot affect an underlying material's transparency, regardless of its transparency mode (alpha blend, alpha scissor, alpha hash, opaque pre-pass). This means translucent or transparent areas of a material will remain translucent or transparent even if an opaque decal is applied on them. + [b]Note:[/b] Decals are only supported in the Forward+ and Mobile rendering methods, not Compatibility. When using the Mobile rendering method, only 8 decals can be displayed on each mesh resource. Attempting to display more than 8 decals on a single mesh resource will result in decals flickering in and out as the camera moves. [b]Note:[/b] When using the Mobile rendering method, decals will only correctly affect meshes whose visibility AABB intersects with the decal's AABB. If using a shader to deform the mesh in a way that makes it go outside its AABB, [member GeometryInstance3D.extra_cull_margin] must be increased on the mesh. Otherwise, the decal may not be visible on the mesh. diff --git a/doc/classes/DisplayServer.xml b/doc/classes/DisplayServer.xml index 8d2bbc4d29..fc5de1b771 100644 --- a/doc/classes/DisplayServer.xml +++ b/doc/classes/DisplayServer.xml @@ -1453,6 +1453,7 @@ Sets the V-Sync mode of the given window. See also [member ProjectSettings.display/window/vsync/vsync_mode]. See [enum DisplayServer.VSyncMode] for possible values and how they affect the behavior of your application. Depending on the platform and used renderer, the engine will fall back to [constant VSYNC_ENABLED] if the desired mode is not supported. + [b]Note:[/b] V-Sync modes other than [constant VSYNC_ENABLED] are only supported in the Forward+ and Mobile rendering methods, not Compatibility. @@ -1744,17 +1745,17 @@ [b]Note:[/b] This flag is implemented on macOS. - No vertical synchronization, which means the engine will display frames as fast as possible (tearing may be visible). Framerate is unlimited (nonwithstanding [member Engine.max_fps]). + No vertical synchronization, which means the engine will display frames as fast as possible (tearing may be visible). Framerate is unlimited (nonwithstanding [member Engine.max_fps]). Not supported when using the Compatibility rendering method. Default vertical synchronization mode, the image is displayed only on vertical blanking intervals (no tearing is visible). Framerate is limited by the monitor refresh rate (nonwithstanding [member Engine.max_fps]). - Behaves like [constant VSYNC_DISABLED] when the framerate drops below the screen's refresh rate to reduce stuttering (tearing may be visible). Otherwise, vertical synchronization is enabled to avoid tearing. Framerate is limited by the monitor refresh rate (nonwithstanding [member Engine.max_fps]). + Behaves like [constant VSYNC_DISABLED] when the framerate drops below the screen's refresh rate to reduce stuttering (tearing may be visible). Otherwise, vertical synchronization is enabled to avoid tearing. Framerate is limited by the monitor refresh rate (nonwithstanding [member Engine.max_fps]). Not supported when using the Compatibility rendering method. Displays the most recent image in the queue on vertical blanking intervals, while rendering to the other images (no tearing is visible). Framerate is unlimited (nonwithstanding [member Engine.max_fps]). - Although not guaranteed, the images can be rendered as fast as possible, which may reduce input lag (also called "Fast" V-Sync mode). [constant VSYNC_MAILBOX] works best when at least twice as many frames as the display refresh rate are rendered. + Although not guaranteed, the images can be rendered as fast as possible, which may reduce input lag (also called "Fast" V-Sync mode). [constant VSYNC_MAILBOX] works best when at least twice as many frames as the display refresh rate are rendered. Not supported when using the Compatibility rendering method. Display handle: diff --git a/doc/classes/Environment.xml b/doc/classes/Environment.xml index 33b6a786ae..7ee1f2dfd4 100644 --- a/doc/classes/Environment.xml +++ b/doc/classes/Environment.xml @@ -46,6 +46,7 @@ If [code]true[/code], enables the [code]adjustment_*[/code] properties provided by this resource. If [code]false[/code], modifications to the [code]adjustment_*[/code] properties will have no effect on the rendered scene. + [b]Note:[/b] Adjustments are only supported in the Forward+ and Mobile rendering methods, not Compatibility. The global color saturation value of the rendered scene (default value is 1). Effective only if [code]adjustment_enabled[/code] is [code]true[/code]. @@ -118,6 +119,7 @@ If [code]true[/code], the glow effect is enabled. + [b]Note:[/b] Glow is only supported in the Forward+ and Mobile rendering methods, not Compatibility. When using the Mobile rendering method, glow will look different due to the lower dynamic range available in the Mobile rendering method. The higher threshold of the HDR glow. Areas brighter than this threshold will be clamped for the purposes of the glow effect. @@ -126,10 +128,10 @@ The bleed scale of the HDR glow. - The lower threshold of the HDR glow. When using the OpenGL renderer (which doesn't support HDR), this needs to be below [code]1.0[/code] for glow to be visible. A value of [code]0.9[/code] works well in this case. + The lower threshold of the HDR glow. When using the Mobile rendering method (which only supports a lower dynamic range up to [code]2.0[/code]), this may need to be below [code]1.0[/code] for glow to be visible. A value of [code]0.9[/code] works well in this case. This value also needs to be decreased below [code]1.0[/code] when using glow in 2D, as 2D rendering is performed in SDR. - The overall brightness multiplier of the glow effect. When using the OpenGL renderer, this should be increased to [code]1.5[/code] to compensate for the lack of HDR rendering. + The overall brightness multiplier of the glow effect. When using the Mobile rendering method (which only supports a lower dynamic range up to [code]2.0[/code]), this should be increased to [code]1.5[/code] to compensate. The intensity of the 1st level of glow. This is the most "local" level (least blurry). @@ -166,7 +168,7 @@ If [code]true[/code], glow levels will be normalized so that summed together their intensities equal [code]1.0[/code]. - The strength of the glow effect. This applies as the glow is blurred across the screen and increases the distance and intensity of the blur. When using the OpenGL renderer, this should be increased to 1.3 to compensate for the lack of HDR rendering. + The strength of the glow effect. This applies as the glow is blurred across the screen and increases the distance and intensity of the blur. When using the Mobile rendering method, this should be increased to compensate for the lower dynamic range. The reflected (specular) light source. @@ -184,6 +186,7 @@ If [code]true[/code], enables signed distance field global illumination for meshes that have their [member GeometryInstance3D.gi_mode] set to [constant GeometryInstance3D.GI_MODE_STATIC]. SDFGI is a real-time global illumination technique that works well with procedurally generated and user-built levels, including in situations where geometry is created during gameplay. The signed distance field is automatically generated around the camera as it moves. Dynamic lights are supported, but dynamic occluders and emissive surfaces are not. + [b]Note:[/b] SDFGI is only supported in the Forward+ rendering method, not Mobile or Compatibility. [b]Performance:[/b] SDFGI is relatively demanding on the GPU and is not suited to low-end hardware such as integrated graphics (consider [LightmapGI] instead). To improve SDFGI performance, enable [member ProjectSettings.rendering/global_illumination/gi/use_half_resolution] in the Project Settings. [b]Note:[/b] Meshes should have sufficiently thick walls to avoid light leaks (avoid one-sided walls). For interior levels, enclose your level geometry in a sufficiently large box and bridge the loops to close the mesh. @@ -230,6 +233,7 @@ If [code]true[/code], the screen-space ambient occlusion effect is enabled. This darkens objects' corners and cavities to simulate ambient light not reaching the entire object as in real life. This works well for small, dynamic objects, but baked lighting or ambient occlusion textures will do a better job at displaying ambient occlusion on large static objects. Godot uses a form of SSAO called Adaptive Screen Space Ambient Occlusion which is itself a form of Horizon Based Ambient Occlusion. + [b]Note:[/b] SSAO is only supported in the Forward+ rendering method, not Mobile or Compatibility. The threshold for considering whether a given point on a surface is occluded or not represented as an angle from the horizon mapped into the [code]0.0-1.0[/code] range. A value of [code]1.0[/code] results in no occlusion. @@ -251,6 +255,7 @@ If [code]true[/code], the screen-space indirect lighting effect is enabled. Screen space indirect lighting is a form of indirect lighting that allows diffuse light to bounce between nearby objects. Screen-space indirect lighting works very similarly to screen-space ambient occlusion, in that it only affects a limited range. It is intended to be used along with a form of proper global illumination like SDFGI or [VoxelGI]. Screen-space indirect lighting is not affected by individual light's [member Light3D.light_indirect_energy]. + [b]Note:[/b] SSIL is only supported in the Forward+ rendering method, not Mobile or Compatibility. The brightness multiplier for the screen-space indirect lighting effect. A higher value will result in brighter light. @@ -269,6 +274,7 @@ If [code]true[/code], screen-space reflections are enabled. Screen-space reflections are more accurate than reflections from [VoxelGI]s or [ReflectionProbe]s, but are slower and can't reflect surfaces occluded by others. + [b]Note:[/b] SSR is only supported in the Forward+ rendering method, not Mobile or Compatibility. The fade-in distance for screen-space reflections. Affects the area from the reflected material to the screen-space reflection). Only positive values are valid (negative values will be clamped to [code]0.0[/code]). @@ -314,7 +320,7 @@ Enables the volumetric fog effect. Volumetric fog uses a screen-aligned froxel buffer to calculate accurate volumetric scattering in the short to medium range. Volumetric fog interacts with [FogVolume]s and lights to calculate localized and global fog. Volumetric fog uses a PBR single-scattering model based on extinction, scattering, and emission which it exposes to users as density, albedo, and emission. - [b]Note:[/b] Volumetric fog is only available in the forward plus renderer. It is not available in the mobile renderer or the compatibility renderer. + [b]Note:[/b] Volumetric fog is only supported in the Forward+ rendering method, not Mobile or Compatibility. Scales the strength of Global Illumination used in the volumetric fog's albedo color. A value of [code]0.0[/code] means that Global Illumination will not impact the volumetric fog. [member volumetric_fog_gi_inject] has a small performance cost when set above [code]0.0[/code]. diff --git a/doc/classes/Light3D.xml b/doc/classes/Light3D.xml index 95c39d535e..2a0b2da46c 100644 --- a/doc/classes/Light3D.xml +++ b/doc/classes/Light3D.xml @@ -54,8 +54,9 @@ If [code]true[/code], the light only appears in the editor and will not be visible at runtime. - The light's angular size in degrees. Increasing this will make shadows softer at greater distances. Only available for [DirectionalLight3D]s. For reference, the Sun from the Earth is approximately [code]0.5[/code]. + The light's angular size in degrees. Increasing this will make shadows softer at greater distances (also called percentage-closer soft shadows, or PCSS). Only available for [DirectionalLight3D]s. For reference, the Sun from the Earth is approximately [code]0.5[/code]. Increasing this value above [code]0.0[/code] for lights with shadows enabled will have a noticeable performance cost due to PCSS. [b]Note:[/b] [member light_angular_distance] is not affected by [member Node3D.scale] (the light's scale or its parent's scale). + [b]Note:[/b] PCSS for directional lights is only supported in the Forward+ rendering method, not Mobile or Compatibility. The light's bake mode. This will affect the global illumination techniques that have an effect on the light's rendering. See [enum BakeMode]. @@ -89,10 +90,12 @@ [Texture2D] projected by light. [member shadow_enabled] must be on for the projector to work. Light projectors make the light appear as if it is shining through a colored but transparent object, almost like light shining through stained-glass. [b]Note:[/b] Unlike [BaseMaterial3D] whose filter mode can be adjusted on a per-material basis, the filter mode for light projector textures is set globally with [member ProjectSettings.rendering/textures/light_projectors/filter]. + [b]Note:[/b] Light projector textures are only supported in the Forward+ and Mobile rendering methods, not Compatibility. - The size of the light in Godot units. Only available for [OmniLight3D]s and [SpotLight3D]s. Increasing this value will make the light fade out slower and shadows appear blurrier. This can be used to simulate area lights to an extent. + The size of the light in Godot units. Only available for [OmniLight3D]s and [SpotLight3D]s. Increasing this value will make the light fade out slower and shadows appear blurrier (also called percentage-closer soft shadows, or PCSS). This can be used to simulate area lights to an extent. Increasing this value above [code]0.0[/code] for lights with shadows enabled will have a noticeable performance cost due to PCSS. [b]Note:[/b] [member light_size] is not affected by [member Node3D.scale] (the light's scale or its parent's scale). + [b]Note:[/b] PCSS for positional lights is only supported in the Forward+ and Mobile rendering methods, not Compatibility. The intensity of the specular blob in objects affected by the light. At [code]0[/code], the light becomes a pure diffuse light. When not baking emission, this can be used to avoid unrealistic reflections when placing lights above an emissive surface. diff --git a/doc/classes/NinePatchRect.xml b/doc/classes/NinePatchRect.xml index 1592718c4b..4b406385e1 100644 --- a/doc/classes/NinePatchRect.xml +++ b/doc/classes/NinePatchRect.xml @@ -68,11 +68,9 @@ Repeats the center texture across the NinePatchRect. This won't cause any visible distortion. The texture must be seamless for this to work without displaying artifacts between edges. - [b]Note:[/b] Only supported when using the Vulkan renderer. When using the OpenGL renderer, this will behave like [constant AXIS_STRETCH_MODE_STRETCH]. Repeats the center texture across the NinePatchRect, but will also stretch the texture to make sure each tile is visible in full. This may cause the texture to be distorted, but less than [constant AXIS_STRETCH_MODE_STRETCH]. The texture must be seamless for this to work without displaying artifacts between edges. - [b]Note:[/b] Only supported when using the Vulkan renderer. When using the OpenGL renderer, this will behave like [constant AXIS_STRETCH_MODE_STRETCH]. diff --git a/doc/classes/OmniLight3D.xml b/doc/classes/OmniLight3D.xml index c0e10574c8..1df8c255dd 100644 --- a/doc/classes/OmniLight3D.xml +++ b/doc/classes/OmniLight3D.xml @@ -5,6 +5,7 @@ An Omnidirectional light is a type of [Light3D] that emits light in all directions. The light is attenuated by distance and this attenuation can be configured by changing its energy, radius, and attenuation parameters. + [b]Note:[/b] When using the Mobile rendering method, only 8 omni lights can be displayed on each mesh resource. Attempting to display more than 8 omni lights on a single mesh resource will result in omni lights flickering in and out as the camera moves. When using the Compatibility rendering method, only 8 omni lights can be displayed on each mesh resource by default, but this can be increased by adjusting [member ProjectSettings.rendering/limits/opengl/max_lights_per_object]. [b]Note:[/b] When using the Mobile or Compatibility rendering methods, omni lights will only correctly affect meshes whose visibility AABB intersects with the light's AABB. If using a shader to deform the mesh in a way that makes it go outside its AABB, [member GeometryInstance3D.extra_cull_margin] must be increased on the mesh. Otherwise, the light may not be visible on the mesh. diff --git a/doc/classes/ProjectSettings.xml b/doc/classes/ProjectSettings.xml index 0b589b4e57..986810edf2 100644 --- a/doc/classes/ProjectSettings.xml +++ b/doc/classes/ProjectSettings.xml @@ -704,7 +704,8 @@ Sets the V-Sync mode for the main game window. See [enum DisplayServer.VSyncMode] for possible values and how they affect the behavior of your application. - Depending on the platform and used renderer, the engine will fall back to [code]Enabled[/code] if the desired mode is not supported. + Depending on the platform and used renderer, the engine will fall back to [b]Enabled[/b] if the desired mode is not supported. + [b]Note:[/b] V-Sync modes other than [b]Enabled[/b] are only supported in the Forward+ and Mobile rendering methods, not Compatibility. [b]Note:[/b] This property is only read when the project starts. To change the V-Sync mode at runtime, call [method DisplayServer.window_set_vsync_mode] instead. @@ -1931,13 +1932,16 @@ Sets the number of MSAA samples to use for 2D/Canvas rendering (as a power of two). MSAA is used to reduce aliasing around the edges of polygons. A higher MSAA value results in smoother edges but can be significantly slower on some hardware. This has no effect on shader-induced aliasing or texture aliasing. + [b]Note:[/b] MSAA is only supported in the Forward+ and Mobile rendering methods, not Compatibility. Sets the number of MSAA samples to use for 3D rendering (as a power of two). MSAA is used to reduce aliasing around the edges of polygons. A higher MSAA value results in smoother edges but can be significantly slower on some hardware. See also bilinear scaling 3d [member rendering/scaling_3d/mode] for supersampling, which provides higher quality but is much more expensive. This has no effect on shader-induced aliasing or texture aliasing. + [b]Note:[/b] MSAA is only supported in the Forward+ and Mobile rendering methods, not Compatibility. Sets the screen-space antialiasing mode for the default screen [Viewport]. Screen-space antialiasing works by selectively blurring edges in a post-process shader. It differs from MSAA which takes multiple coverage samples while rendering objects. Screen-space AA methods are typically faster than MSAA and will smooth out specular aliasing, but tend to make scenes appear blurry. The blurriness is partially counteracted by automatically using a negative mipmap LOD bias (see [member rendering/textures/default_filters/texture_mipmap_bias]). Another way to combat specular aliasing is to enable [member rendering/anti_aliasing/screen_space_roughness_limiter/enabled]. + [b]Note:[/b] Screen-space antialiasing is only supported in the Forward+ and Mobile rendering methods, not Compatibility. If [code]true[/code], uses a fast post-processing filter to make banding significantly less visible in 3D. 2D rendering is [i]not[/i] affected by debanding unless the [member Environment.background_mode] is [constant Environment.BG_CANVAS]. @@ -1946,11 +1950,14 @@ Enables Temporal Anti-Aliasing for the default screen [Viewport]. TAA works by jittering the camera and accumulating the images of the last rendered frames, motion vector rendering is used to account for camera and object motion. Enabling TAA can make the image blurrier, which is partially counteracted by automatically using a negative mipmap LOD bias (see [member rendering/textures/default_filters/texture_mipmap_bias]). - [b]Note:[/b] The implementation is not complete yet, some visual instances such as particles and skinned meshes may show artifacts. + [b]Note:[/b] The implementation is not complete yet. Some visual instances such as particles and skinned meshes may show ghosting artifacts in motion. + [b]Note:[/b] TAA is only supported in the Forward+ rendering method, not Mobile or Compatibility. + If [code]true[/code], enables a spatial filter to limit roughness in areas with high-frequency detail. This can help reduce specular aliasing to an extent, though not as much as enabling [member rendering/anti_aliasing/quality/use_taa]. This filter has a small performance cost, so consider disabling it if it doesn't benefit your scene noticeably. + [b]Note:[/b] TAA is only supported in the Forward+ and Mobile rendering methods, not Compatibility. @@ -1968,7 +1975,7 @@ If [code]true[/code], performs a previous depth pass before rendering 3D materials. This increases performance significantly in scenes with high overdraw, when complex materials and lighting are used. However, in scenes with few occluded surfaces, the depth prepass may reduce performance. If your game is viewed from a fixed angle that makes it easy to avoid overdraw (such as top-down or side-scrolling perspective), consider disabling the depth prepass to improve performance. This setting can be changed at run-time to optimize performance depending on the scene currently being viewed. - [b]Note:[/b] Only supported when using the Vulkan Clustered backend or the OpenGL backend. When using Vulkan Mobile there is no depth prepass performed. + [b]Note:[/b] Depth prepass is only supported when using the Forward+ or Compatibility rendering method. When using the Mobile rendering method, there is no depth prepass performed. Thread model for rendering. Rendering on a thread can vastly improve performance, but synchronizing to the main thread can cause a bit more jitter. @@ -2174,12 +2181,15 @@ Max number of omnilights and spotlights renderable per object. At the default value of 8, this means that each surface can be affected by up to 8 omnilights and 8 spotlights. This is further limited by hardware support and [member rendering/limits/opengl/max_renderable_lights]. Setting this low will slightly reduce memory usage, may decrease shader compile times, and may result in faster rendering on low-end, mobile, or web devices. + [b]Note:[/b] This setting is only effective when using the Compatibility rendering method, not Forward+ and Mobile. Max number of elements renderable in a frame. If more elements than this are visible per frame, they will not be drawn. Keep in mind elements refer to mesh surfaces and not meshes themselves. Setting this low will slightly reduce memory usage and may decrease shader compile times, particularly on web. For most uses, the default value is suitable, but consider lowering as much as possible on web export. + [b]Note:[/b] This setting is only effective when using the Compatibility rendering method, not Forward+ and Mobile. Max number of positional lights renderable in a frame. If more lights than this number are used, they will be ignored. Setting this low will slightly reduce memory usage and may decrease shader compile times, particularly on web. For most uses, the default value is suitable, but consider lowering as much as possible on web export. + [b]Note:[/b] This setting is only effective when using the Compatibility rendering method, not Forward+ and Mobile. @@ -2271,7 +2281,8 @@ Determines how sharp the upscaled image will be when using the FSR upscaling mode. Sharpness halves with every whole number. Values go from 0.0 (sharpest) to 2.0. Values above 2.0 won't make a visible difference. - Sets the scaling 3D mode. Bilinear scaling renders at different resolution to either undersample or supersample the viewport. FidelityFX Super Resolution 1.0, abbreviated to FSR, is an upscaling technology that produces high quality images at fast framerates by using a spatially aware upscaling algorithm. FSR is slightly more expensive than bilinear, but it produces significantly higher image quality. FSR should be used where possible. + Sets the scaling 3D mode. Bilinear scaling renders at different resolution to either undersample or supersample the viewport. FidelityFX Super Resolution 1.0, abbreviated to FSR, is an upscaling technology that produces high quality images at fast framerates by using a spatially-aware upscaling algorithm. FSR is slightly more expensive than bilinear, but it produces significantly higher image quality. On particularly low-end GPUs, the added cost of FSR may not be worth it (compared to using bilinear scaling with a slightly higher resolution scale to match performance). + [b]Note:[/b] FSR is only effective when using the Forward+ rendering method, not Mobile or Compatibility. If using an incompatible rendering method, FSR will fall back to bilinear scaling. Scales the 3D render buffer based on the viewport size uses an image filter specified in [member rendering/scaling_3d/mode] to scale the output image to the full viewport size. Values lower than [code]1.0[/code] can be used to speed up 3D rendering at the cost of quality (undersampling). Values greater than [code]1.0[/code] are only valid for bilinear mode and can be used to improve 3D rendering quality at a high performance cost (supersampling). See also [member rendering/anti_aliasing/quality/msaa_3d] for multi-sample antialiasing, which is significantly cheaper but only smooths the edges of polygons. @@ -2295,9 +2306,11 @@ If [code]true[/code], forces vertex shading for all rendering. This can increase performance a lot, but also reduces quality immensely. Can be used to optimize performance on low-end mobile devices. + [b]Note:[/b] This setting currently has no effect, as vertex shading is not implemented yet. Lower-end override for [member rendering/shading/overrides/force_vertex_shading] on mobile devices, due to performance concerns or driver support. + [b]Note:[/b] This setting currently has no effect, as vertex shading is not implemented yet. The filtering quality to use for [Decal] nodes. When using one of the anisotropic filtering modes, the anisotropic filtering level is controlled by [member rendering/textures/default_filters/anisotropic_filtering_level]. diff --git a/doc/classes/ReflectionProbe.xml b/doc/classes/ReflectionProbe.xml index e912925cd2..cc48e0612b 100644 --- a/doc/classes/ReflectionProbe.xml +++ b/doc/classes/ReflectionProbe.xml @@ -7,6 +7,7 @@ Captures its surroundings as a cubemap, and stores versions of it with increasing levels of blur to simulate different material roughnesses. The [ReflectionProbe] is used to create high-quality reflections at a low performance cost (when [member update_mode] is [constant UPDATE_ONCE]). [ReflectionProbe]s can be blended together and with the rest of the scene smoothly. [ReflectionProbe]s can also be combined with [VoxelGI], SDFGI ([member Environment.sdfgi_enabled]) and screen-space reflections ([member Environment.ssr_enabled]) to get more accurate reflections in specific areas. [ReflectionProbe]s render all objects within their [member cull_mask], so updating them can be quite expensive. It is best to update them once with the important static objects and then leave them as-is. [b]Note:[/b] Unlike [VoxelGI] and SDFGI, [ReflectionProbe]s only source their environment from a [WorldEnvironment] node. If you specify an [Environment] resource within a [Camera3D] node, it will be ignored by the [ReflectionProbe]. This can lead to incorrect lighting within the [ReflectionProbe]. + [b]Note:[/b] Reflection probes are only supported in the Forward+ and Mobile rendering methods, not Compatibility. When using the Mobile rendering method, only 8 reflection probes can be displayed on each mesh resource. Attempting to display more than 8 reflection probes on a single mesh resource will result in reflection probes flickering in and out as the camera moves. [b]Note:[/b] When using the Mobile rendering method, reflection probes will only correctly affect meshes whose visibility AABB intersects with the reflection probe's AABB. If using a shader to deform the mesh in a way that makes it go outside its AABB, [member GeometryInstance3D.extra_cull_margin] must be increased on the mesh. Otherwise, the reflection probe may not be visible on the mesh. diff --git a/doc/classes/RenderingDevice.xml b/doc/classes/RenderingDevice.xml index a09940d30e..64be8e6807 100644 --- a/doc/classes/RenderingDevice.xml +++ b/doc/classes/RenderingDevice.xml @@ -8,7 +8,7 @@ On startup, Godot creates a global [RenderingDevice] which can be retrieved using [method RenderingServer.get_rendering_device]. This global RenderingDevice performs drawing to the screen. Internally, [RenderingDevice] is used in Godot to provide support for several modern low-level graphics APIs while reducing the amount of code duplication required. [b]Local RenderingDevices:[/b] Using [method RenderingServer.create_local_rendering_device], you can create "secondary" rendering devices to perform drawing and GPU compute operations on separate threads. - [b]Note:[/b] [RenderingDevice] is not available when running in headless mode or when using the OpenGL renderer. + [b]Note:[/b] [RenderingDevice] is not available when running in headless mode or when using the Compatibility rendering method. diff --git a/doc/classes/SpotLight3D.xml b/doc/classes/SpotLight3D.xml index 9dff742748..6bb4ae548c 100644 --- a/doc/classes/SpotLight3D.xml +++ b/doc/classes/SpotLight3D.xml @@ -5,6 +5,7 @@ A Spotlight is a type of [Light3D] node that emits lights in a specific direction, in the shape of a cone. The light is attenuated through the distance. This attenuation can be configured by changing the energy, radius and attenuation parameters of [Light3D]. + [b]Note:[/b] When using the Mobile rendering method, only 8 spot lights can be displayed on each mesh resource. Attempting to display more than 8 spot lights on a single mesh resource will result in spot lights flickering in and out as the camera moves. When using the Compatibility rendering method, only 8 spot lights can be displayed on each mesh resource by default, but this can be increased by adjusting [member ProjectSettings.rendering/limits/opengl/max_lights_per_object]. [b]Note:[/b] When using the Mobile or Compatibility rendering methods, spot lights will only correctly affect meshes whose visibility AABB intersects with the light's AABB. If using a shader to deform the mesh in a way that makes it go outside its AABB, [member GeometryInstance3D.extra_cull_margin] must be increased on the mesh. Otherwise, the light may not be visible on the mesh. diff --git a/doc/classes/TextureLayered.xml b/doc/classes/TextureLayered.xml index 8f6dff7acf..7a0940edd4 100644 --- a/doc/classes/TextureLayered.xml +++ b/doc/classes/TextureLayered.xml @@ -8,7 +8,7 @@ Data is set on a per-layer basis. For [Texture2DArray]s, the layer specifies the array layer. All images need to have the same width, height and number of mipmap levels. A [TextureLayered] can be loaded with [method ResourceLoader.load]. - Internally, Godot maps these files to their respective counterparts in the target rendering driver (Vulkan, GLES3). + Internally, Godot maps these files to their respective counterparts in the target rendering driver (Vulkan, OpenGL3). diff --git a/doc/classes/VoxelGI.xml b/doc/classes/VoxelGI.xml index 4347c5845f..cc78426736 100644 --- a/doc/classes/VoxelGI.xml +++ b/doc/classes/VoxelGI.xml @@ -5,6 +5,7 @@ [VoxelGI]s are used to provide high-quality real-time indirect light and reflections to scenes. They precompute the effect of objects that emit light and the effect of static geometry to simulate the behavior of complex light in real-time. [VoxelGI]s need to be baked before having a visible effect. However, once baked, dynamic objects will receive light from them. Furthermore, lights can be fully dynamic or baked. + [b]Note:[/b] [VoxelGI] is only supported in the Forward+ rendering method, not Mobile or Compatibility. [b]Procedural generation:[/b] [VoxelGI] can be baked in an exported project, which makes it suitable for procedurally generated or user-built levels as long as all the geometry is generated in advance. For games where geometry is generated at any time during gameplay, SDFGI is more suitable (see [member Environment.sdfgi_enabled]). [b]Performance:[/b] [VoxelGI] is relatively demanding on the GPU and is not suited to low-end hardware such as integrated graphics (consider [LightmapGI] instead). To improve performance, adjust [member ProjectSettings.rendering/global_illumination/voxel_gi/quality] and enable [member ProjectSettings.rendering/global_illumination/gi/use_half_resolution] in the Project Settings. To provide a fallback for low-end hardware, consider adding an option to disable [VoxelGI] in your project's options menus. A [VoxelGI] node can be disabled by hiding it. [b]Note:[/b] Meshes should have sufficiently thick walls to avoid light leaks (avoid one-sided walls). For interior levels, enclose your level geometry in a sufficiently large box and bridge the loops to close the mesh. To further prevent light leaks, you can also strategically place temporary [MeshInstance3D] nodes with their [member GeometryInstance3D.gi_mode] set to [constant GeometryInstance3D.GI_MODE_STATIC]. These temporary nodes can then be hidden after baking the [VoxelGI] node. -- cgit v1.2.3