Used native hardware MSAA x4
Full multisampling support including custom blend and compositions.
Must be verified on web and 4K resolution for performance issues
Full review of clipping workflow.
Before we are used separate render target for each clip path and compute shader to find clip path intersections with AND logic.
Now we are using depth buffer and transactions from depth to stencil and back to get AND logic intersections,
Compute shaders, layouts and pipelines was removed
Since we've separated ClipPath and Masking,
Masking now has a distinct and independent purpose.
API Modification:
- enum class CompositeMethod -> enum class MaskMethod
- Result Paint::composite(std::unique_ptr<Paint> target, CompositeMethod method) -> Result Paint::mask(std::unique_ptr<Paint> target, MaskMethod method)
- CompositeMethod Paint::mask(const Paint** target) const -> MaskMethod Paint::mask(const Paint** target) const
issue: https://github.com/thorvg/thorvg/issues/1372
- refactored the Fill matrix to hold internal data statically.
- refactored for clean & neat the raster engine / svg loader logic.
API Modification:
- Matrix Fill::transform() const -> Matrix& Fill::transform() const
issue: https://github.com/thorvg/thorvg/issues/1372
Scene effects are typically applied to modify
the final appearance of a rendered scene,
such as adding a blur effect.
Each effect would have a different number of parameters
to control its visual properties. The Scene::push() interface
uses variadic arguments to accommodate various cases.
Users should refer to the SceneEffect API documentation
and pass the parameters exactly as required for the specific
effect type. For instance, GaussianBlur expects 3 parameters
which are:
- sigma(float)[greater than 0]
- direction(int)[both: 0 / horizontal: 1 / vertical: 2]
- border(int)[extend: 0 / wrap: 1]
- quality(int)[0 ~ 100]
and, scene->push(SceneEffect::GaussianBlur, 5.0f, 0, 0, 100);
New Experimental APIs:
- SceneEffect::ClearAll
- SceneEffect::GaussianBlur
- Result Scene::push(SceneEffect effect, ...);
Example:
- examples/SceneEffect
issue: https://github.com/thorvg/thorvg/issues/374
In cases where the distance between points is 0, further
processing of the points results in division by zero.
To avoid this check, we ensure that duplicate points are
not added during trimming.
Streaming model for massive vertex and index creations: minimize memory allocations, range checks and other conditions
Reduce number of segments length calculations (sqrt) and bbox (min and max).
Update distances and bboxes on a whole buffer and only if necessary. For shapes without strokes compute distances not necessary at all. bbox can be updated only on the final stage of geometry workflow, but not on the each stage.
Using stack memory instead of heap. its more cache friendly and did not fragment memory, faster memory allocations (weak place of realization)
Using cache for points distances and whole path length. Updates only if necessary
Validation of geometry consistency executes only on the final stage of path life cicle. It more friendly for data streaming: no any conditions and branches.
Using binary search for strokes trimming
Pre-cached circles geometry for caps and joints
Refactored strokes elements generation functions. Code is more readable and modifiable in general. Can be easily fixed if some geometry issues will be finded
Fixed detection of close vertices using comparison with epsilon and related artifacts in stencil buffer as extra pixels.
Fixed incorrect tessellation of curves using scaling.
Vertex ordering during line join needed adjustment.
Vertices were forming two partialy overlapping triangles
instead of a rectangle. The issue was rarely visible since
the resulting rectangle often overlapped significantly
with other rectangles.
Emscripten 3.1.66 includes support for WebGPU's premultiplied canvas.
Surface configurations have been updated with premultiplied alpha mode to support this feature.
see: c82a307c61
Corrected the alpha interpolation order during blending.
This also corrected the hard mix blending result in the guitar sample.
issue: https://github.com/thorvg/thorvg/issues/2704
move instance, adapter and device creation from renderer to application
its necessary for web integration, because browser have its own mechanics to create hardware handles
this changes makes webgpu canvas more universal to use in case of system and web applications
issue: https://github.com/thorvg/thorvg/issues/2410
The anti-aliased outline color was incorrectly blended
at the multiply option.
The fix can be observed in the example:
'examples/lottie/resourcesguitar.json'
in order to do this, RenderMehthod::blend() method introduced
`bool direct` for figuring out the intermediate composition.
- used hardware blending stage for scene blending
- used AABB for scene blending
- reduced number of offfscreen buffers coping
- reduced number of render pass switching
- used render pipelines abilities to convert offscreen pixel format to screen pixel format
- removed unused shaders
Spec out this incomplete experimental feature,
this is a still promising one, we will reintroduce
this officially after 1.0 release
size: -2kb
issue: https://github.com/thorvg/thorvg/issues/1372
Deep shader refactoring for the following purposes:
* used pre-calculated gradient texture instead of per-pixel gradient map computation
* used HW wrap samples for fill spread setting
* unified gradient shader types
* used single shader module for composition instead of signle module per composition type
* used single shader module for blending for each of fill type (solid, gradient, image) instaed of signle module per blend type
* much easier add new composition and blend equations
* get rided std::string uasge
* shaders code is more readable
* bind groups creation in real time removed - performance boost
* blend and composition shaders decomposed - performance boost
* shader modules and pipeline layouts generalized - less memory usage
* shared single stencil buffer used - less memory usage
* bind groups usage simplified
* general context API simplified and generalized
* all rendering logic moved into new composition class
* ready for hardware MSAA (in next steps)
* ready for direct mask applience (in next steps)
