Abstract: Graphics processing includes setting up a plurality of objects in a scene in virtual space, each object being defined by a set of vertices. A unique object identifier is assigned to each object and written to an ID buffer. Draw calls are issued to draw the objects associated with the object identifiers. Parameter values of the vertices are manipulated to output vertex parameter values. Primitives are set up from the vertices, each primitive being defined by one or more of the vertices. Each primitive belongs to one or more of the objects. Each primitive is rasterized at a plurality of pixels. Processing the pixels includes spatial or temporal anti-aliasing that utilizes the one or more object identifiers of the plurality of object identifiers. The pixels are processed for each rasterized primitive to generate an output frame.

Abstract: Graphics processing systems and methods are disclosed which may minimize invocations to a pixel shader in order to improve efficiency in a rendering pipeline. In implementations of the present disclosure, a plurality of samples within a pixel may be covered by a primitive. The plurality of samples may include one or more color samples and a plurality of depth samples. The nature of the samples which were covered by the primitive may be taken into account before invoking a pixel shader to perform shading computations on the pixel. In implementations of the present disclosure, if at least one sample is covered by a primitive, but none of the samples are color samples, an invocation to a pixel shader may be avoided.

Abstract: A graphics processing unit (GPU) is configured to receive metadata specifying an active sample configuration for a particular region of a display device among a plurality of regions of the display device and receive pixel data for one or more pixels in the particular region. The pixel data specifies the same number of color samples for each pixel. For each pixel in the particular region, the GPU invokes a pixel shader only for color samples specified to be active samples by the configuration.

Abstract: Graphics processing includes setting up a plurality of objects in a scene in virtual space, each object being defined by a set of vertices. A unique object identifier is assigned to each object and written to an ID buffer. Draw calls are issued to draw the objects associated with the object identifiers. Parameter values of the vertices are manipulated to output vertex parameter values. Primitives are set up from the vertices, each primitive being defined by one or more of the vertices. Each primitive belongs to one or more of the objects. Each primitive is rasterized at a plurality of pixels. Processing the pixels includes spatial or temporal anti-aliasing that utilizes the one or more object identifiers of the plurality of object identifiers. The pixels are processed for each rasterized primitive to generate an output frame.

Abstract: Morphological anti-aliasing (MLAA) of a re-projection of a two-dimensional image can be implemented in a way that produces a better result while using fewer processor resources. One or more discontinuities between each neighboring pixel of the two-dimensional image are determined. One or more pre-defined patterns formed by the one or more discontinuities are identified. A blend amount is calculated for each pixel neighboring the identified pre-defined patterns. A re-projection is applied to the two-dimensional image and to the blend amount for each pixel thereby generating re-projected blend amounts. The neighboring pixels of the re-projection are then blended according to the re-projected blend amounts.

Abstract: Morphological anti-aliasing (MLAA) of a re-projection of a two-dimensional image can be implemented in a way that produces a better result while using fewer processor resources. One or more discontinuities between each neighboring pixel of the two-dimensional image are determined. One or more pre-defined patterns formed by the one or more discontinuities are identified. A blend amount is calculated for each pixel neighboring the identified pre-defined patterns. A re-projection is applied to the two-dimensional image and to the blend amount for each pixel thereby generating re-projected blend amounts. The neighboring pixels of the re-projection are then blended according to the re-projected blend amounts.