Abstract: An interactive tile-based ML terrain generation method is disclosed. At a first phase of a painting of a digital environment using a brush tool, a modification to a terrain surface of the digital environment is approximated. The approximating includes decomposing a stroke of the brush tool into one or more stamps. Each of the one or more stamps changes a height of a portion of terrain surface as the brush tool passes over the portion of the terrain surface. At a second phase of the painting of the digital environment, details are added to the portion of the terrain surface passed over by each of the one or more stamps. The adding of the details includes dividing work associated with the adding of the details into one or more tiles and processing the one or more tiles.

Abstract: Methods, systems, and computer-readable media for rendering light probes in a virtual environment are disclosed. Noisy lighting data is accessed in a data structure associated with a light probe in a set of light probes in an environment. The noisy lighting data is provided as an input to a neural network. The neural network is trained to output an estimate of non-noisy lighting data based on the input. The noisy lighting data is replaced in the data structure with the estimated non-noisy lighting data.

Abstract: A graphics-processing unit is used to perform mesh simplification. A vertex shader receives a dataset for an input mesh that portrays a three-dimensional graphics object. The vertex shader generates from the dataset vertices for primitives that make up the input mesh. The input mesh is divided into a grid of cells. A geometry shader receives the vertices from the vertex shader and generates from the received vertices a simplified mesh that portrays the three-dimensional graphics object in less detail than the input mesh. Before the input mesh is divided into grid cells, a warping function can be applied to the input mesh based on a weighting function to warp the input mesh, thereby increasing sampling at a region of interest. A projective warping can be performed on the grid to produce grid cells of different volumes in accordance with a camera position.

Abstract: A method carried out by graphics processing circuitry includes generating animated coarse mesh vertex information based on instanced coarse mesh data; and tessellating instanced coarse mesh data based on the animated coarse mesh vertex information to produce instances of a three dimensional object for display. A graphics processing circuitry includes programmable shader logic operative to execute programmable instructions that when executed cause the programmable shader logic to generate animated coarse mesh vertex information based on instanced coarse mesh data; and tessellate instanced coarse mesh data based on the animated coarse mesh vertex information to produce instances of a three dimensional object for display. Another method carried out by graphics processing circuitry includes determining density of a plurality of three dimensional objects in a current view on a display; setting a tessellation level based on said density; and tessellating said plurality of three dimensional objects.

Abstract: A graphics-processing unit is used to perform mesh simplification. A vertex shader receives a dataset for an input mesh that portrays a three-dimensional graphics object. The vertex shader generates from the dataset vertices for primitives that make up the input mesh. The input mesh is divided into a grid of cells. A geometry shader receives the vertices from the vertex shader and generates from the received vertices a simplified mesh that portrays the three-dimensional graphics object in less detail than the input mesh. Before the input mesh is divided into grid cells, a warping function can be applied to the input mesh based on a weighting function to warp the input mesh, thereby increasing sampling at a region of interest. A projective warping can be performed on the grid to produce grid cells of different volumes in accordance with a camera position.