Implement Octahedral Compression for normal/tangent vectors
*Oct32 for uncompressed vectors
*Oct16 for compressed vectors
Reduces vertex size for each attribute by
*Uncompressed: 12 bytes, vec4<float32> -> vec2<unorm16>
*Compressed: 2 bytes, vec4<unorm8> -> vec2<unorm8>
Binormal sign is encoded in the y coordinate of the encoded tangent
Added conversion functions to go from octahedral mapping to cartesian
for normal and tangent vectors
sprite_3d and soft_body meshes write to their vertex buffer memory
directly and need to convert their normals and tangents to the new oct
format before writing
Created a new mesh flag to specify whether a mesh is using octahedral
compression or not
Updated documentation to discuss new flag/defaults
Created shader flags to specify whether octahedral or cartesian vectors
are being used
Updated importers to use octahedral representation as the default format
for importing meshes
Updated ShaderGLES2 to support 64 bit version codes as we hit the limit
of the 32-bit integer that was previously used as a bitset to store
enabled/disabled flags
If the user changed the portal Z scale in the editor the portal margin display could become incorrectly sized.
This is because the portal margin is measured in world space units, and has to be back calculated into model space using the inverse global transform of the portal node. The model space size of the margin is thus tied to the current scale of the node.
This PR forces updating the gizmo each time the transform is changed. This isn't super efficient, but as this is an editor only feature it should be okay, and it is unlikely to be a performance problem.
Due to a quirk in CSG Shapes, updating is usually deferred to the next frame. This is problematic as we need to read back the geometry on the first frame when converting levels.
This PR adds a function to CSGShape to force immediate updating (if dirty), and calls it during room conversion.
The default autoplace algorithm places instances in the highest priority Room. It became apparent that there are some situations in which users will want to override this and force placement in a Room from a particular RoomGroup, especially an "outside" RoomGroup.
This setting allows the user to specify a preference for Room priority. When set to 0, the setting is ignored and the highest priority Room is chosen.
Although explicit portals did a check to detect internal portals, this check was missing from autolinked portals. This meant they were incorrectly clipping the room bounds of the enclosing outer room.
This PR adds a check for internal rooms during the autolinking and sets the internal flag where needed.
The default portal margin is stored in the RoomManager. Previously this was propagated to Portals when the value was changed, and Portals each stored this default value.
This caused a bug during loading, if the RoomManager was loaded before Portals, the value was never propagated.
This PR makes the default margin a static value stored in the RoomManager, and the Portals now read directly from the static value when required, and do not store locally. This gets around the problem.
Higher values will make indirect lighting brighter.
A value of 1.0 represents physically accurate behavior, but higher values
can be used to make indirect lighting propagate more visibly when using
a low number of bounces.
This can be used to speed up bake times by lowering the number of bounces
then increasing the bounce indirect energy. Unlike BakedLightmapData's
energy property, this property does not affect direct lighting
emitted by light nodes or emissive materials.
Due to an oversight in the autoplace recursive search for static objects, static objects were getting added twice to the portal renderer, which meant they were being rendered twice, lowering performance.
This PR corrects this horrendous error.
Allows users to have the RoomManager as the roomlist.
Fixes a couple of bugs dealing with situations where users attempt to link Portals to Rooms outside the roomlist.
Adds a PortalEditorPlugin allowing you to flip individual portals.
In order to make level building easier, the system can now support STATIC and DYNAMIC objects in the roomlist that are not placed in rooms. The system will automatically place them in the appropriate room.
* Safe and unsafe motion are calculated by dichotomy with a limited
number of steps. It's good for performance, but on long motions that
either collide near the beginning or near the end, the result can be
very imprecise.
* Now a factor 0.25 or 0.75 is used to converge faster when this case
happens, which allows longer motions to get more accurate collision
detection.
* Makes snap collision more precise, and helps with cases where diagonal collision on the border of a platform can lead to the character being stuck.
Additional improvements to move_and_slide:
* Handle slide canceling in move_and_collide with 0 velocity instead of
not applying it.
* Better handling of snap with custom logic to cancel sliding.
* Remove small jittering when using stop on slope, by canceling the
motion completely when the resulting motion is less than margin instead
of always projecting to the up direction (in both body motion and snap).
Co-authored-by: fabriceci <fabricecipolla@gmail.com>
Make sure the direction of the motion is preserved, unless the depth is
higher than the margin, which means the body needs depenetration in any
direction.
Also changed move_and_slide to avoid sliding on the first motion, in
order to avoid issues with unstable position on ground when jumping.
Co-authored-by: fabriceci <fabricecipolla@gmail.com>
- Specular can now be set above 1.0.
- Some properties now allow more precise values.
- The Bake Mode property hint was changed to be more descriptive.
- The maximum shadow bias values were adjusted to follow the property
hints in `master`.
It turned out the new autolinking feature was linking portals AFTER the static meshes had been added to rooms in the PortalRenderer. This meant that large meshes weren't being sprawled across these portals. The fix involves doing the autolinking BEFORE adding the static meshes.
Fixes a bug in the warning for portals being in the wrong direction, they should have only been checkout for outgoing portals. This was resulting in erroneous warnings.
Also the room conversion logs are refined to be more compact and informative.
A warning icon is also added in the gizmo for portals where autolink fails.
Clean: remove duplicate and interior vertices (uses Bullet algorithm)
Simplify: modify the geometry for further simplification (uses VHACD
algorithm)
In the editor, single convex hull now uses the clean option.
Added a new editor entry to create a simplified convex hull, can be
useful for creating convex hull from highly tessellated triangle meshes.
Specific change for 3.x:
Add support for Vector<Vector3> and PoolVector<Vector3> in the convex hull generator.
This backports the improved RayCast debug drawing functionality
from the `master` branch.
`ArrayMesh.clear_surfaces()` was also backported from the `master`
branch and exposed because the new debug drawing code requires it.
Although the visibility enabler worked to turn on and off AnimationPlayer as it enters and exits the view frustum, this was of little use as bones animation and especially software skinning still take place driven by the AnimationTree node.
This PR adds the ability to turn on and off AnimationTree, and AnimationTreePlayer nodes as they enter or exit the view frustum, which achieves the intention of switching off expensive animation processing.
This commits adds a new emitter type for particles material
and 3D CPU particles. The new emitter is called "ring"
and it can emit either in a ring or cylinder fashion.
This adds the following properties for the emitter:
1. ring_emitter_axis: the axis along which the ring/cylinder
will be constructed
2. ring_emitter_radius: outer radius of the ring/cylinder
3. ring_emitter_inner_radius: inner radius of the cylinder.
when set to zero, particles will emit in the full volume.
4. ring_emitter_height: height of the ring/cylinder emitter.
Co-authored-by: Rémi Verschelde <rverschelde@gmail.com>
We've been using standard C library functions `memcpy`/`memset` for these since
2016 with 67f65f6639.
There was still the possibility for third-party platform ports to override the
definitions with a custom header, but this doesn't seem useful anymore.
Backport of #48239.
Fix for a regression from software skinning support:
instance_attach_skeleton wasn't called in set_mesh before, and it's
causing issues when the mesh instance is loaded from a thread.
1. Call from a thread queues instance_attach_skeleton with RID() in the
visual server.
2. Call from the main thread when entering tree calls
instance_attach_skeleton immediately with a valid skeleton
3. Queued instance_attach_skeleton resets the attached skeleton
This change prevents that to happen by making sure
instance_attach_skeleton is not called on set_mesh as it was doing
before, but there might be a more general problem to solve in how
visual server commands are executed when resources are loaded from
a different thread.
* Better handling of the scene's environment energy in the lightmapper
bakes.
* Fixed a bug where ProceduralSky::get_panorama() returned a reference
instead of a copy.
* Removed includes to Embree's internal header files.
MeshInstance added as child nodes for CollisionObject debug shapes can
be invalidated while deleting the collision object (child nodes are
deleted first), which caused accesses to invalid memory in
shape_owner_remove_shape that lead to random crashes.
Also optimized accesses to shapes to avoid copy-on-write on each
iteration.
There was a specific case where the node path wasn't checked for
validity before trying to access the attachment node.
It could cause lots of error log noise in both editor and game.
If the editor was started with --debug-collisions, 3d shapes were
displayed twice, both with the gizmo and debug shapes. Some shapes could
also persist after being removed due to the usage of queue_free() to
destroy the debug shapes.
- Fix objects with no material being considered as fully transparent by the lightmapper.
- Added "environment_min_light" property: gives artistic control over the shadow color.
- Fixed "Custom Color" environment mode, it was ignored before.
- Added "interior" property to BakedLightmapData: controls whether dynamic capture objects receive environment light or not.
- Automatically update dynamic capture objects when the capture data changes (also works for "energy" which used to require object movement to trigger the update).
- Added "use_in_baked_light" property to GridMap: controls whether the GridMap will be included in BakedLightmap bakes.
- Set "flush zero" and "denormal zero" mode for SSE2 instructions in the Embree raycaster. According to Embree docs it should give a performance improvement.
When one of the bodies exited the tree, the corresponding node path was
reset instead of just resetting the joint from the physics server. That
was causing the node path to be reset on scene switch when one of the
bodies is under the joint in the scene tree.
- Based on C++11's `atomic`
- Reworked `SafeRefCount` (based on the rewrite by @hpvb)
- Replaced free atomic functions by the new `SafeNumeric<T>`
- Replaced wrong cases of `volatile` by the new `SafeFlag`
- Platform-specific implementations no longer needed
Co-authored-by: Hein-Pieter van Braam-Stewart <hp@tmm.cx>
- Based on C++11's `mutex`
- No more need to allocate-deallocate or check for null
- No pointer anymore, just a member variable
- Platform-specific implementations no longer needed
- Simpler for `NO_THREADS`
- `BinaryMutex` added for special cases as the non-recursive version
- `MutexLock` now takes a reference. At this point the cases of null `Mutex`es are rare. If you ever need that, just don't use `MutexLock`.
- `ScopedMutexLock` is dropped and replaced by `MutexLock`, because they were pretty much the same.
Added BakedLightmap.use_hdr and BakedLightmap.use_color properties
that can reduce the flie size of lightmap texture at the expense of quality.
Changed the denoiser to work in a single buffer, reducing RAM
usage. Also added the `-mstackrealign` flag in the denoiser compilation
for MinGW builds. This flag helped fix a bug in Embree, so I want to see
if it will help fix GH #45296.
- Fix crash when a ray hits a texel with a UV2 coordinate exactly
equal to 1.0.
- Take BakedLightmap extents into account.
- Clear capture data between bakes.
- Fix minor issues with seam correction.
When visual instances are hidden, there is no need to update their global transforms and send these to the visual server. This only needs to be done when the objects are reshown.
After adding `_change_notify("extents");` to `set_extents()` method in `BakedLightmap` class
the issue was fixed
Fix disconnection between Gizmo and inspector values on BakedLightmap
Remove unnecessary `_change_notify("bake_extents")` call from `set_extents()`
- Fix Embree runtime when using MinGW (patch by @RandomShaper).
- Fix baking of lightmaps on GridMaps.
- Fix some GLSL errors.
- Fix overflow in the number of shader variants (GLES2).
Fix Raycast node render debug not showing in editor camera preview.
Use dynamic mesh update to change the ray on-the-fly without too much
extra cost when collision debug is enabled.
Fixes#43571
Completely re-write the lightmap generation code:
- Follow the general lightmapper code structure from 4.0.
- Use proper path tracing to compute the global illumination.
- Use atlassing to merge all lightmaps into a single texture (done by @RandomShaper)
- Use OpenImageDenoiser to improve the generated lightmaps.
- Take into account alpha transparency in material textures.
- Allow baking environment lighting.
- Add bicubic lightmap filtering.
There is some minor compatibility breakage in some properties and methods
in BakedLightmap, but lightmaps generated in previous engine versions
should work fine out of the box.
The scene importer has been changed to generate `.unwrap_cache` files
next to the imported scene files. These files *SHOULD* be added to any
version control system as they guarantee there won't be differences when
re-importing the scene from other OSes or engine versions.
This work started as a Google Summer of Code project; Was later funded by IMVU for a good amount of progress;
Was then finished and polished by me on my free time.
Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
This change does two things:
1. Properly update the internal shape data using _update_in_shape_owner
when updating a shape (in 2D it was resetting one way collision)
2. Avoid unnecessary updates when calling set_shape with the same shape,
which happens each time a shape property is modified
(e.g shape.extents.x = ...)
Fixes#45090
(cherry picked from commit 4b43cd17c5)
Happy new year to the wonderful Godot community!
2020 has been a tough year for most of us personally, but a good year for
Godot development nonetheless with a huge amount of work done towards Godot
4.0 and great improvements backported to the long-lived 3.2 branch.
We've had close to 400 contributors to engine code this year, authoring near
7,000 commit! (And that's only for the `master` branch and for the engine code,
there's a lot more when counting docs, demos and other first-party repos.)
Here's to a great year 2021 for all Godot users 🎆
(cherry picked from commit b5334d14f7)
See #36285 which mistakenly added documentation for the whole C++ API, while
some of it is meant to be and stay private as it's not exposed to scripts.
The access modifiers and method prefix were not used properly.
Cleanup code, and rename wrong `group_name` parameters to `method`, as it's a
method name which is being broadcast.
This is a very old class from pre-open source days, chances are that it was
just forgotten and not meant to be kept as is and undocumented.
(cherry picked from commit 79ba70f7ee)
See #43689.
Also 'fixed' some spelling for behavior in publicly visible strings.
(Sorry en_GB, en_CA, en_AU, and more... Silicon Valley won the tech spelling
war.)
(cherry picked from commit a655de89e3)
This fixes a regression from PR #40313 (support for software skinning in MeshInstance).
Before, the base mesh was always updated on load even if not skinning
was used, which caused mesh instance materials to be reset on the
rendering side.
Now the base mesh is set only when it has been modified, or when
switching software skinning on or off. In this case the mesh instance
materials are always updated properly afterwards.
My mistake I missed this before the PR got merged:
is_visible_in_tree() should be avoided being called per frame unless absolutely necessary, because it is a recursive function that traverses the scene tree. It should be used when deciding on rare occasions whether to switch on or off skeleton processing, but it is better to use the cheaper is_visible() check on the per frame update.
Don't apply lighting to objects when they have a lightmap texture and
the light is set to BAKE_ALL. This prevents applying the same direct
light twice on the same object and makes setting up scenes with mixed
lighting much easier.
Option in MeshInstance to enable software skinning, in order to test
against the current USE_SKELETON_SOFTWARE path which causes problems
with bad performance.
Co-authored-by: lawnjelly <lawnjelly@gmail.com>
InterpolatedCamera has already been removed from the `master` branch.
This adds a deprecation notice to inform people about the upcoming removal
in Godot 4.0.
Its functionality could be replicated in a GDScript add-on with relative
ease.
Changed CPU velocity calculation for EMISSION_SHAPE_DIRECTED_POINTS
to follow the same logic as in the GPU version:
mat2 rotm;
rotm[0] = texelFetch(emission_texture_normal, emission_tex_ofs, 0).xy;
rotm[1] = rotm[0].yx * vec2(1.0, -1.0);
VELOCITY.xy = rotm * VELOCITY.xy;
Now both CPUParticles2D & CPUParticles3D (z disabled) show the same results
as their GPU counterparts and take the initial velocity settings into account.
(cherry picked from commit 1c231cacb3)
Add optional parameter to specify whether applying rotation to the
PathFollow3D's Transform is necessary, preventing erroneous updates.
(cherry picked from commit be3a1769fe)
Configured for a max line length of 120 characters.
psf/black is very opinionated and purposely doesn't leave much room for
configuration. The output is mostly OK so that should be fine for us,
but some things worth noting:
- Manually wrapped strings will be reflowed, so by using a line length
of 120 for the sake of preserving readability for our long command
calls, it also means that some manually wrapped strings are back on
the same line and should be manually merged again.
- Code generators using string concatenation extensively look awful,
since black puts each operand on a single line. We need to refactor
these generators to use more pythonic string formatting, for which
many options are available (`%`, `format` or f-strings).
- CI checks and a pre-commit hook will be added to ensure that future
buildsystem changes are well-formatted.
(cherry picked from commit cd4e46ee65)
(cherry picked from commit 38085f2f6982c491935a434bb45e358dbebe1714)
(cherry picked from commit b9c280b73ff6a13ea490d2da0f2728bcef3038dc)
(cherry picked from commit 895ed2aed7)
In the case where a ConcavePolygonShape is used as a shape for a RigidBody
in another mode than static, a configuration warning will appear in the
editor.
(cherry picked from commit ad227d9a85)
Fixes#26637.
Fixes#19900.
The viewport_size returned by get_viewport_size was previously incorrect, being half the correct value. The function is renamed to get_viewport_half_extents, and now returns a Vector2.
Code which called this function has also been modified accordingly.
This PR also fixes shadow culling when using ortho cameras, because the correct input for CameraMatrix::set_orthogonal should be the full HEIGHT from get_viewport_half_extents, and not half the width.
It also fixes state.ubo_data.viewport_size in rasterizer_scene_gles3.cpp to be the width and the height of the viewport in pixels as stated in the documentation, rather than the current value which is half the viewport extents in worldspace, presumed to be a bug.
