Abstract: In a technique for rendering non-linear BRDFs that are stable in both the temporal and spatial domains, without serious interruption to the content creation pipeline used in most games, non-linear content is linearized by rendering in texture space at a fixed resolution. A MIP-map chain is calculated from this texture. The complete MIP-map chain is used for rendering on a display device. Low resolution reflectance parameters are used to approximate the highest resolution reflectance parameters as the object becomes smaller on the display device. The low resolution reflectance parameters are calculated using non linear fitting techniques.

Abstract: A multi-threaded graphics processor is configured to use to extrapolate low resolution mipmaps stored in physical memory to produce extrapolated texture values while high resolution nonresident mipmaps are retrieved from a high latency storage resource and converted into resident mipmaps. The extrapolated texture values provide an improved image that appears sharper compared with using the low resolution mipmap level texture data in place of the temporarily unavailable high resolution mipmap level texture data. An extrapolation threshold LOD is used to determine when extrapolated magnification or minification texture filtering is used. The extrapolation threshold LOD may be used to smoothly transition from using extrapolated filtering to using interpolated filtering when a nonresident mipmap is converted to a resident mipmap.

Abstract: A multi-threaded graphics processor is configured to use to extrapolate low resolution mipmaps stored in physical memory to produce extrapolated texture values while high resolution mipmaps are retrieved from a high latency storage resource. The extrapolated texture values provide an improved image that appears sharper compared with using the low resolution mipmap level texture data in place of the temporarily unavailable high resolution mipmap level texture data.

Abstract: In a method for determining weighting factors for the color calculation of a color value of texels for a footprint covering a plurality of texels in a texel grid, in a graphic system, form information of the footprint is determined at first. Afterwards, the edges of the footprint are determined and the edges determined in this way are approximated by a staircase function. The texels of the texel grid contacted by the staircase function are determined and a weighting factor is determined for each texel containing a portion of the staircase function, depending on the subarea of the respective texel covered by the footprint.

Abstract: A method of organizing memory for storage of texture data, in accordance with one embodiment of the invention, includes accessing a size of a mipmap level of a texture map. A block dimension may be determined based on the size of the mipmap level. A memory space (e.g., computer-readable medium) may be logically divided into a plurality of whole number of blocks of variable dimension. The dimension of the blocks is measured in units of gobs and each gob is of a fixed dimension of bytes. A mipmap level of a texture map may be stored in the memory space. A texel coordinate of said mipmap level may be converted into a byte address of the memory space by determining a gob address of a gob in which the texel coordinate resides and determining a byte address within the particular gob.

Abstract: A display device includes a wavelet transform unit for transforming an original image into wavelet coefficients using a Harr wavelet transform formula, a level obtaining unit for obtaining a mipmap level of a mipmap image. The display device also includes an inverse wavelet transform unit for receiving and subjecting at least a portion of the wavelet coefficients obtained by the transformation by the wavelet transform unit, to an inverse transform using an inverse Harr wavelet transform formula until an order having a value equal to the mipmap level is obtained, and outputting an image represented by at least a portion of a low-frequency component of wavelet coefficients having the order having the value equal to the mipmap level. Additionally, the display device includes a polygon drawing unit for drawing the image output by the inverse wavelet transform unit, as a mipmap image, on the polygon image.

Abstract: The invention provides techniques for wrapping a two-dimensional texture conformally onto a surface of a three dimensional virtual object within an arbitrarily-shaped, user-defined region. The techniques provide minimum distortion and allow interactive manipulation of the mapped texture. The techniques feature an energy minimization scheme in which distances between points on the surface of the three-dimensional virtual object serve as set lengths for springs connecting points of a planar mesh. The planar mesh is adjusted to minimize spring energy, and then used to define a patch upon which a two-dimensional texture is superimposed. Points on the surface of the virtual object are then mapped to corresponding points of the texture. The invention also features a haptic/graphical user interface element that allows a user to interactively and intuitively adjust texture mapped within the arbitrary, user-defined region.

Abstract: A system, method and computer program product are provided for computer graphics processing. In use, a value is modified based on an algorithm. An operation is subsequently performed on pixel data taking into account the modified value.

Abstract: Circuits, methods, and apparatus that provide texture caches and related circuits that store and retrieve texels in a fast and efficient manner. One such texture circuit provides an increased number of bilerps for each pixel in a group of pixels, particularly when trilinear or aniso filtering is needed. For trilinear filtering, texels in a first and second level of detail are retrieved for a number of pixels during a clock cycle. When aniso filtering is performed, multiple bilerps can be retrieved for each of a number of pixels during one clock cycle.

Abstract: A graphics system supports arrays of cube map textures. In one implementation, a cube map texture is utilized as an index into a set of cube map textures. The set of cube map textures may further be arranged into an atlas of two-dimensional textures.

Abstract: Systems and methods for border color handling in a graphics processing unit are disclosed. In one embodiment, the system includes a border color register that stores at least one border color pointer. A border color pointer indicates an address in an external memory at which border color information is located. Border color information is populated within external memory and retrieved by the texture cache controller if the texture filter unit requires a border color for texture mapping operations.

Abstract: There is provided a system and method for filter kernel interpolation for seamless mipmap filtering. There is provided a method of implementing a filter kernel interpolation for seamless filtering of transitions within a plurality of mipmaps derived from a base-image texture filtered using a prefilter, the method comprising choosing a filter kernel, determining a filter width for the filter kernel, selecting a first mipmap and a second mipmap from the plurality of mipmaps, applying interpolation on the filter kernel based on the prefilter, the first mipmap, and the second mipmap to generate an interpolated filter kernel, and applying the interpolated filter kernel to the first mipmap to generate a seamless filtered texture. Two alternative methods of interpolation are introduced, including filter kernel value interpolation and filter kernel position interpolation with x-lerping. By avoiding access to the second mipmap, greater efficiency and image quality can be achieved versus conventional interpolation.

Abstract: Existing map solutions using tiled maps with geographic data in vector format typically has such a volume of data that not all geographic data can be stored in memory. A primary bottleneck for map rendering speed in such solutions is the time needed to load all data necessary for replacing one rendered map with the next rendered map. Utilizing the structure and methods of the invention herein when loading geographic vector data, the average time to load all necessary geographic data needed for one rendered map to the next will be greatly reduced—the loading time reduction will be especially large when pan and zoom operations are the main transitions from one rendered map to the next.

Abstract: The current invention involves new systems and methods for increasing texture filtering performance by reorganizing a texture sampling order used to read and filter texels when anisotropic filtering is used. Texel read performance is improved for anisotropic filtering by reorganizing texel reads when a texel cache is used. The texel reads are paired based on a major axis alignment in pixel space. The paired texel reads for a pixel footprint may also be ordered to improve texel coherency, thereby improving a texture cache hit rate.

Abstract: A method for determining a pose of an object from a range image is presented. A distribution of poses for an object is determined from a model of the object. A set of reference poses is selected according to the distribution, such that more reference poses are selected for more probable poses than less probable poses. A reference image for each reference pose is generated, and the reference pose used to generate the reference image that best matches a range image including the object is assigned as the pose of the object.

Abstract: An image special effect device includes: a graphic processor and a video processing block; the graphic processor converting coordinates in virtual three-dimensional space into two-dimensional coordinates on a display screen in accordance with a supply of information on the shape of a model in the virtual three-dimensional space, computing texture coordinates of an image that is pasted on the display screen and a reduction rate of the model on the display screen, and outputting information on the texture coordinates and reduction rate from an image data output unit; the video processing block writing input video data YUV into a memory after filtered by using a pre-filter coefficient corresponding to information on the reduction rate supplied from the graphic processor, and reading out the video data from the memory by using the supplied texture coordinates as information of read-address.

Abstract: A first command is retrieved from a script containing one or more commands written for a first color space. The first command is associated with zero or more input buffers and zero or more output buffers. The first command has zero or more parameters. A behavior of the first command in the first color space and in a second color space is determined. The behavior comprises one of unique behavior, transparent behavior, and different behavior. The first command has the unique behavior when the first command only operates in the first color space. The first command has the transparent behavior when the first command generates similar results in the first color space and in the second color space. The first command has the different behavior when the first command generates different results in the first color space and in the second color space. Using the behavior of the first command, an operation associated with the first command is processed.

Abstract: The current invention involves new systems and methods for increasing texture filtering performance based on pixel coverage. When half of the pixels in a 2×2 pixel quad are not covered, texel coordinates for the uncovered pixels are not output. Therefore, the texels for the uncovered pixels are not read or processed, allowing the texel filtering processing throughput to be used to produce filtered results for covered pixels. This optimization is particularly useful when anisotropic filtering is used since the number of texels needed to produce a filtered result for a pixel increases as the anisotropic ratio increases. Elimination of unnecessary texel processing for uncovered pixels may improve texture filtering performance.

Abstract: In a technique for rendering non-linear BRDFs that are stable in both the temporal and spatial domains, without serious interruption to the content creation pipeline used in most games, non-linear content is linearized by rendering in texture space at a fixed resolution. A MIP-map chain is calculated from this texture. The complete MIP-map chain is used for rendering on a display device. Low resolution reflectance parameters are used to approximate the highest resolution reflectance parameters as the object becomes smaller on the display device. The low resolution reflectance parameters are calculated using non linear fitting techniques.

Abstract: Embodiments include a texture mapping processor incorporating a dynamic level of detail map for use in a graphics processing system. Level of detail values are defined, with 0 being the finest and corresponding to the largest mipmap level. Each bound texture in a graphics object is assigned an identifier. This identifier is used as an index into a minimum-LOD value tracking table that is updated whenever a texel is fetched. A texture processing module controls when the tracking table is initialized and read back, and which identifiers are tracked. The minimum-LOD values in the tracking table are accompanied by a coarse region access mask to associate a minimum LOD value with a specific region of the image or object. A clamping table contains LOD clamp values for each region and a region code that specifies the coarseness of the LOD associated with each region of the texture.

Abstract: A memory controller having graphic processing function that includes a graphic processing unit operating in response to a selection signal from a master, and a memory interface for storing outputs of the graphic processing unit in an external memory at and receiving graphic data from the external memory to provide the graphic data to the graphic processing unit.

Abstract: A still image producing apparatus in which a user selects one piece of information, and production of still images and music associated with the mood in time of the production is provided. A still image producing apparatus 100 comprises: an image feature extraction unit 102 which analyzes image data, so as to extract feature; a music feature extraction unit 106 which extracts a music feature quantity from music data; and a mapping process unit 109 which uses each feature, so as to respectively map image data and music data on each sensitivity map using two axes. The user selects, from the sensitivity map, via a menu input unit 111, a quadrant for music or images desired in a slide show. Thus, the slide show associated with the mood is automatically produced in a slide show generation unit 110.

Abstract: Circuits, methods, and apparatus that provide texture caches and related circuits that store and retrieve texels in a fast and efficient manner. One such texture circuit provides an increased number of bilerps for each pixel in a group of pixels, particularly when trilinear or aniso filtering is needed. For trilinear filtering, texels in a first and second level of detail are retrieved for a number of pixels during a clock cycle. When aniso filtering is performed, multiple bilerps can be retrieved for each of a number of pixels during one clock cycle.

Abstract: Methods and systems for texture mapping in a computer-implemented graphics pipeline are described. A sample group is identified as including a divergent pixel. A determination is made whether an operand of an instruction executing on the divergent pixel satisfies a condition. A scheme for determining a level of detail for the texture mapping is selected depending on whether or not the condition is satisfied.

Abstract: A pixel is textured by storing a first texel reference value, a second texel reference value, and texel mapping values where each texel mapping value represents a k-tuple of (ternary) references to the first texel reference value, the second texel reference value and a third texel reference value to thereby represent a block of texels. A pixel value for the pixel is generated from the stored texel values and the pixel is displayed responsive to the generated pixel value. In some embodiments, respective pluralities of texel reference values and texel mapping values that map thereto are stored for respective ones of a plurality of overlapping blocks of texels. In further embodiments, a first mipmap value for a pixel is bilinearly interpolated from the retrieved texel values for the set of nearest neighbor texels. A second mipmap value for the pixel is generated by averaging the retrieved texel values for the set of nearest neighbor texels.

Abstract: An accuracy enhancing system may include an image processor cooperating with a database for generating a reference geospatial image corresponding to the collected geospatial image, a change detector cooperating with the image processor for detecting a change between the collected geospatial image and the reference geospatial image, and an accuracy enhancer. The accuracy enhancer may cooperate with the change detector for generating at least one enhanced accuracy value corresponding to the at least one geospatial collection value based upon the change detected between the collected geospatial image and the reference geospatial image. The airborne platform may traverse an actual flight path based upon a planned flight path, and the image processor may generate the reference geospatial image based upon a desired match with the collected geospatial image.

Abstract: A multi-resolution texture mapping system suitable for large scale terrain rendering using commodity graphics processing units (GPU). The GPU vertex and fragment shaders are used to implement the clip-mapping functionality. The terrain texture is represented by a combination of a mip-map and a multi-level clip-map having independent origins and off-set values. The independent clip-map levels may be independently updated. The offset values allow the origins to be associated with a reference point in a scene to be rendered. The desired clip-map level to be used to render a particular fragment may be determined using the base 2 logarithm of the maximum screen-space derivative of the source texture required by the terrain geometry to be drawn. If the desired clip-map level is non-integer and lies between two clip-map levels, appropriate texel data is created by interpolating between the bounding clip-map levels. This interpolation allows a multi-resolution texture mapping to be displayed.

Abstract: A texture unit of a graphics processing unit provides the ability to switch among different filter modes depending upon shader program instructions that are received by the texture unit. One filter mode has the capability to extract filter weights that have been specified in a received shader program instruction rather than calculating the weights within the texture unit itself.

Abstract: Systems and methods for positioning bilinear texture samples to produce an anisotropically filtered texture mapped pixel may improve texture mapping performance and image quality. The bilinear texture samples are positioned along a major axis of anisotropy to approximate an elliptical footprint, ensuring that the bilinear texture samples span the entire axis of anisotropy without extending beyond the major axis of anisotropy. An additional bilinear texture sample or a pair of additional bilinear texture samples is positioned in the center of the axis of anisotropy dependent on the anisotropic ratio. The additional bilinear texture samples are weighed less than the other bilinear texture samples and all of the bilinear textures samples lie within the anisotropic footprint.

Abstract: A method and system for performing texture mapping across adjacent texture maps. Specifically, one embodiment of the present invention sets forth a method, which includes the steps of determining whether a texel crosses a boundary of a first texture map, examining a first texture state identifier associated with the first texture map, and requesting for a second texture state identifier associated with a second texture map that is adjacent to the first texture map to enable traversal to the second texture map to access the texel if the first texture state identifier includes a mode indicative of wrapping to an adjacent texture map and texture adjacency information that points to a second texture map.

Abstract: One embodiment of the present invention relates to a system for reconstructing an image. During operation, the system receives an original image, wherein the original image includes a defective area to be reconstructed. The system also receives a reference to a texture image, which contains a texture to be used during the reconstruction. The system then divides the values of pixels in the defective area and surrounding boundary pixels, by the values of corresponding pixels in the texture image to produce a first intermediate image. Next, the system solves a partial differential equation (PDE) for non-boundary pixels in the first intermediate image subject to values of the boundary pixels in the first intermediate image to produce a second intermediate image. The system then multiplies the values of pixels in the second intermediate image by the values of corresponding pixels in the texture image to produce a replacement image.

Abstract: An image change detecting system may include an image processor cooperating with the database for generating a reference geospatial image corresponding to the collected geospatial image, an image enhancer for enhancing at least one of the reference geospatial image and the collected geospatial image based upon at least one environmental condition, and a change detector cooperating with the image processor and the image enhancer. The change detector may detect a change between the collected geospatial image and the reference geospatial image with at least one thereof enhanced by the image enhancer based upon the at least one environmental condition. The environmental condition may include a weather condition, a time of day, or a time of year. The environmental condition may be typically associated with the collected geospatial image.

Abstract: A device and method for controlling generation of a final pixel utilizes a conditional statement, referred to as an IF_NEIGHBOR statement, which when compiled, causes a programmable pixel shader to perform mip map texture lookups even if a pixel of interest does not meet the condition of the conditional statement. As such, any neighboring pixels needed for mip map selection have their associated shader code guaranteed to execute even though the pixel of interest may fail the conditional portion of the conditional statement. The device and method executes texture address calculations for pixels within a region and for pixels outside of a region but only those necessary to determine the mip map level corresponding to a pixel within the region. Execution of shader code for a current pixel is executed if any of the surrounding neighboring pixels meet the desired condition even if the current pixel does not meet the condition.

Abstract: A method and apparatus for synthesizing and expressing multiple effects and textures in a 3-dimensional graph image. The method includes: defining nodes in VRML (Virtual Reality Modeling Language), the nodes being used to express multiple effects; inputting VRML files corresponding to the defined nodes, the VRML files having information on the 3-dimensional image and the multiple effects to be expressed in the 3-dimensional image; generating the 3-dimensional image by using the input VRML files; and expressing the multiple effects in the generated 3-dimensional image by using the information on the multiple effects in the input VRML files. Accordingly, the VRML nodes and fields are defined to synthesize multiple effects and multiple textures in the 3-dimensional image, so that it is possible to effectively apply the multiple effects and multiple texture effects in the 3-dimensional image in cooperation with a conventional 3-dimensional image expression method and apparatus.

Abstract: A method for implementing LOD (level of detail) filtering in a cube mapping application. The method includes accessing a first sample and a second sample for a cube map. A cube map path is computed between the first sample and the second sample. A distance is computed between the first sample and the second sample, wherein the distance is measured using the cube map path. LOD filtering is then implemented by using the distance between the first sample and the second sample.

Abstract: A method and apparatus for mapping a texture to a surface of a computer generated object dependent upon the geometric shape of the object. The method includes the steps of approximating a true pixel color by performing a number of texturing operations and averaging the results of the texturing operations. The apparatus including a texturing unit comprising a control unit for determining a set of N footprint textel locations, a RAM for determining N sets of textel values with each set of textel values being associated with one footprint textel location, an interpolator for interpolating the N sets of textel values for each textel location, and an averaging unit for determining an average value from the interpolated values for each textel location.

Abstract: A system maintains data from different resolution levels of an image in textures of a graphics processing unit (GPU). Image data is organized into multiple resolution layers of an image. Data from the lower resolution level(s) is used to process an image while higher resolution data is loaded. In one embodiment, a first resolution representation of the image having a lowest resolution level using data resident in the GPU memory is drawn prior to drawing a portion of a second resolution representation having a higher resolution level using data resident in the GPU memory.

Abstract: A mipmap generator generates pairs of mipmaps that are each of a lower resolution that its respective source image. A single-pass, gradient-based motion vector generator generates an image motion vector map having values that represent the motion trajectories for pixels in the first and second source images. An image interpolator generates an interpolated image based on the source images and the image motion vector map. A motion detector generates a motion factor map based on a pair of mipmaps from those generated by the mipmap generator that represents a detected degree of motion between the first and second source images. The blending module generates a blended, upconverted new image using the motion factor map, the interpolated image and one of the first and second motion maps.

Abstract: A first command is retrieved from a script containing one or more commands written for a first color space. The first command is associated with zero or more input buffers and zero or more output buffers. The first command has zero or more parameters. A behavior of the first command in the first color space and in a second color space is determined. The behavior comprises one of unique behavior, transparent behavior, and different behavior. The first command has the unique behavior when the first command only operates in the first color space. The first command has the transparent behavior when the first command generates similar results in the first color space and in the second color space. The first command has the different behavior when the first command generates different results in the first color space and in the second color space. Using the behavior of the first command, an operation associated with the first command is processed.

Abstract: As for a multi-dimensional texture constituted by a plurality of textures different in photographing environment such as a viewing position or lighting conditions, texture images constituting the multi-dimensional texture are rearranged block by block or pixel by pixel, and converted into an anisotropic texture expressing anisotropic variations of the multi-dimensional texture in an area corresponding to the block or the pixel.

Abstract: An apparatus for mapping primitives of a 3D graphics model from a texture space to a screen space. The apparatus includes a texture memory for storing texture maps. A resampler resamples, for each primitive, data from a texture map that corresponds to the primitive to corresponding pixel data defining a portion of a display image that corresponds to the primitive. The texture space resampler and/or the screen space resampler is operative to select a resampling algorithm for performing the resampling from a respective set of at least two distinct resampling algorithms. The selection is done in dependence on a size of the primitive.

Abstract: Ripmapping and footprint assembly are used to anisotropically filter texture maps. A subset of the set of ripmaps associated with a base texture is created and stored. The subset includes ripmaps selected to maximize anisotropic texture sampling performance and to minimize the texture memory requirements. For pixel footprints not aligned with the anisotropy of ripmaps or requiring a ripmap outside of the subset, footprint assembly is used to perform anisotropic filtering by taking multiple isotropic probes from a mipmap. For texture samples aligned within a tolerance range of the anisotropy of a ripmap, footprint assembly constructs an anisotropic texture sample from one or more samples of a ripmap. Ripmap statistics are collected during texture mapping to dynamically determine an optimal subset of ripmaps, and additional ripmaps can be added to the subset on demand if warranted. A graphics driver can analyze ripmap statistics to determine the subset of ripmaps.

Abstract: Provided is a method and apparatus for rendering 3D graphics data. By calculating the size of a primitive, which is a basic constituent unit of objects indicated by the graphics data, on a screen, selecting one of a plurality of resolutions supported by a video stream according to the calculated size, generating a video frame image having the selected resolution from the video stream, and rendering the graphics data using the generated video frame image, the amount of memory space used and power consumed are reduced. In addition, since when rendering is performed using a video frame image decoded at a low resolution, a processing speed increases, and since rendering can be performed using video frame images decoded at various resolutions, image quality increases.

Abstract: A multi-threaded graphics processor is configured to use to extrapolate low resolution mipmaps stored in physical memory to produce extrapolated texture values while high resolution nonresident mipmaps are retrieved from a high latency storage resource and converted into resident mipmaps. The extrapolated texture values provide an improved image that appears sharper compared with using the low resolution mipmap level texture data in place of the temporarily unavailable high resolution mipmap level texture data. An extrapolation threshold LOD is used to determine when extrapolated magnification or minification texture filtering is used. The extrapolation threshold LOD may be used to smoothly transition from using extrapolated filtering to using interpolated filtering when a nonresident mipmap is converted to a resident mipmap.

Abstract: An image processing apparatus and method which can achieve a reduction in size of the crossbar circuit and achieve a higher speed of processing, which perform DDA processing (ST11), then read out texture data from a memory (ST12), perform sub-word reallocation processing (ST13), then perform texture-filtering (ST14), then globally distribute data by the crossbar circuit 13 to a first operation processing element of each processing module (ST15), then perform processing at the pixel level, specifically use the texture data after filtering and the various types of data after rasterization to perform operations by pixel units and draw the pixel data passing the various types of tests in the processing at the pixel level to a frame buffer on a memory module (ST16).

Abstract: A method and apparatus to generate one or more graphics textures of a video image, wherein graphics textures have differing resolutions, and to render the video image as a graphics texture using the one or more graphics textures of differing resolutions. The one or more graphics textures have different resolutions, for example, a base resolution and one or more scaled resolutions. The graphics texture may be rendered and displayed on a surface of a multi-dimensional graphics object that may be manipulated on a display.

Abstract: A method for improving hit rates of a cache memory for storing texture data during graphics rendering is disclosed. In response to a request for a mipmap block from a first mipmap to render a texture, a determination is made whether or not the mipmap block from the first mipmap is already stored in a cache memory. If the mipmap block from the first mipmap is already in the cache memory, the mipmap block from the first mipmap already in the cache memory is utilized to render the texture. Otherwise, if the mipmap block from the first mipmap is not in a cache memory, another determination is made whether or not a bias value associated with the mipmap block from the first mipmap is set. If a bias value associated with the mipmap block from the first mipmap is not set, then the mipmap block is obtained from the first mipmap to render the texture.

Abstract: A method of displaying a graphical texture comprises determining whether an update exists for a particular content element from among one or more content elements, wherein each content element is associated with one or more textures; if an update exists for the particular content element, performing the steps of determining a particular texture from among the one or more textures with which the information is associated; obtaining the particular texture, wherein the particular texture is generated based on the particular content element; and displaying the particular texture.

Abstract: Systems and methods are presented for real-time realistic river modeling. In one implementation, when provided with the physical features of a modeled terrain, the systems and methods described herein automatically determine a watercourse route in relation to the terrain and calculate river borders. The systems and methods then calculate a surface structure for the river between the borders and map texture coordinates to the surface structure. The systems and methods apply textures at the texture coordinates to provide an animated, view-dependent representation of wave geometries for the modeled water surface including reflection, refraction, and Fresnel effects.

Abstract: An image rendering apparatus is provided to reduce the amount of data communicated between a shader and a texture unit. The texture unit performs a texture mapping operation for mapping a texture to an object in a three-dimensional space. An input unit receives a single texture address parameter from the shader unit. A coordinate generation unit generates a plurality of texel coordinates based on the texture address parameter. A retrieval unit retrieves a plurality of texel values corresponding respectively to the plurality of generated texel coordinates in a texture cache or in an external texture buffer. An output unit outputs the retrieved texel values to the shader unit in a batch. Since a plurality of texel values are read from a single texture address parameter input, the bandwidth for communications from the shader to the texture unit may be saved.