WO2011065886A1 - Decoding system and method operable on encoded texture element blocks - Google Patents
Decoding system and method operable on encoded texture element blocks Download PDFInfo
- Publication number
- WO2011065886A1 WO2011065886A1 PCT/SE2010/050217 SE2010050217W WO2011065886A1 WO 2011065886 A1 WO2011065886 A1 WO 2011065886A1 SE 2010050217 W SE2010050217 W SE 2010050217W WO 2011065886 A1 WO2011065886 A1 WO 2011065886A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- adder
- output
- operator
- texel
- decoder
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/04—Texture mapping
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/127—Prioritisation of hardware or computational resources
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/14—Coding unit complexity, e.g. amount of activity or edge presence estimation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/182—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a pixel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/436—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation using parallelised computational arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/44—Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/80—Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/16—Indexing scheme for image data processing or generation, in general involving adaptation to the client's capabilities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/08—Bandwidth reduction
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/156—Availability of hardware or computational resources, e.g. encoding based on power-saving criteria
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/167—Position within a video image, e.g. region of interest [ROI]
Definitions
- the present invention generally relates to decoding, and in particular to a decoding system and method operable on encoded texture element blocks.
- an embodiment relates to a decoding system for generating a decoded pixel based on at least one encoded texel block.
- the decoding system comprises N parallel decoders that each are configured to decode an encoded texel block to generate one or more texel values.
- the circuitries of the N decoders are different as compared to each other so that each decoder of the N decoders has a unique circuitry that is different from the circuitry of the other N - 1 decoders. This enables the decoder circuitries to have different levels of complexities since the decoders have different capabilities in terms of the texel positions within a texel block they can handle.
- the decoding system also comprises a value selector connected to the N decoders and configured to selectively output N texel values. These N texel values are further output from one or more of the N decoders.
- An embodiment defines a method of generating a pixel by decoding at least one encoded texel block.
- the method comprises selectively outputting N texel values from at least one of N different decoders, each having a circuitry that is unique and different from the circuitry of the other N - i decoders.
- Each decoder further outputs at least one texel value based on an input encoded texel block.
- the selection of which texel values from the N decoders to use for the calculation of the pixel value of the pixel to decode is performed based on the position of the N texels associated with the N texel values relative the boundary of a texel block.
- the selected N texel values are then used for calculating the pixel value.
- FIG. 2 is a schematic overview of a decoding arrangement comprising a decoding system according to an embodiment
- Fig. 3 is an illustration of a decoding system according to an embodiment
- Fig. 4 schematically illustrates an encoded texel block according to an embodiment
- Fig. 5 illustrates the concept of employing a planar approximation of texel property values in a texel block
- Fig. 6 is a hardware implementation example of the fourth decoder in Fig. 3;
- Fig. 7 is a hardware implementation example of the third decoder in Fig. 3;
- Fig. 8 is a hardware implementation example of the second decoder in Fig. 3;
- Fig. 9 is a hardware implementation example of the first decoder in Fig. 3;
- Fig. 10 is a hardware implementation example of a seven-to-eight bit extender;
- Fig. 11 is a hardware implementation of a clamper;
- Fig. 12 is another hardware implementation of a clamper
- Fig. 13 is a hardware implementation example of the value selector of Fig. 3; and Fig. 14 is a flow diagram illustrating a decoding method according to an embodiment.
- the embodiments generally relate to decoding or decompression of encoded or compressed texture element (texel) blocks for the purpose of generating decoded pixel values.
- the decoding is well adapted for usage with three-dimensional (3D) graphics, such as games, 3D maps and scenes, 3D messages, e.g. animated messages, screen savers, man-machine interfaces (MMIs), etc., but is not limited thereto.
- 3D graphics such as games, 3D maps and scenes, 3D messages, e.g. animated messages, screen savers, man-machine interfaces (MMIs), etc.
- MMIs man-machine interfaces
- the embodiments could also be employed for decoding other types of images or graphics, e.g. one-dimensional (1 D), two-dimensional (2D) or 3D images.
- a texture encoding method or system is typically employed.
- Such an encoding system should result in more efficient usage of expensive on-chip memory and lower memory bandwidth during rendering and, thus, in lower power consumption and/or faster rendering.
- This reduction in bandwidth and processing power requirements is particularly important for thin clients, such as mobile units and telephones, with a small amount of memory, little memory bandwidth and limited power, i.e. powered by batteries.
- textures are divided into so-called texel blocks, each comprising a number of texels.
- a texel block is typically rectangular or quadratic, such as a size of MRXMC texels, where MR, MC are positive integer numbers with the proviso that both M R and Mc are not simultaneously one.
- M R Mc and preferred such block embodiments could be 4 ⁇ 4 texels.
- a texel is associated with a texel value and a pixel is associated with a pixel value.
- the texel and pixel values can advantageously be color values.
- a color value can be represented in various color formats, such as red, green, blue (RGB) color or as luminance and chrominance.
- RGB red, green, blue
- a texel value could represent other properties of a texel, such as normal or Z value.
- a texel value is non-limitedly exemplified as a color value.
- a typical rasterizer does not access the texture a single texel at a time. Instead a small neighborhood of texels is employed, such as 2x2 texels.
- bilinear filtering For instance, during bilinear filtering, four such neighboring texels are read to produce a bilinearly filtered value, i.e. a pixel value.
- a bilinear patch For instance, a 2x2 texel group employed in bilinear filtering is denoted a bilinear patch.
- This concept can be extended further, for instance, to trilinear filtering. Trilinear filtering uses two bilinear patches from two different mipmaps.
- FIG. 1A illustrates the case where a bilinear patch 20 of four texels 30-36 are found in a single texel block 10.
- Figs. 1 B and 1 C illustrate another situation where the bilinear patch 20 spans two pixel blocks 10, 12, 14 either present side-by-side (Fig. 1 B) or on top of each other (Fig. 1 C). The extreme case is depicted in Fig. 1 D with one texel 30-36 of the bilinear patch 20 in a respective texel block 10-16.
- the traditional way to solve this problem is to use four identical decoders, each of which decodes a predetermined texel in the bilinear patch.
- the encoded texel block representing the texel block 10 is input in the four decoders.
- the first decoder output texel (1 , 2) 30, texel (2, 2) is obtained from the second decoder with the third and fourth decoders generating texel (1 , 3) and (2, 3) 36 of the bilinear patch, respectively.
- a coordinate system in which the origin is in the top left texel of the texel block 10 is assumed.
- Fig. 1A the encoded texel block representing the texel block 10 is input in the four decoders.
- the first decoder output texel (1 , 2) 30, texel (2, 2) is obtained from the second decoder with the third and fourth decoders generating texel (1 , 3) and (2, 3) 36 of
- the decoding system comprises N different decoders.
- N is an integer equal to or larger than two, preferably equal to four.
- Each decoder 100, 200, 300, 400 has a circuitry that is different from the circuitry of the other N - 1 decoders.
- the decoding system 1 does not comprise identical decoders as in the prior art but in clear contrast different decoders 100, 200, 300, 400 with different unique circuitries.
- the difference in circuitries of the N decoders 100, 200, 300, 400 imply that the combination of the circuit elements of one of the decoders is different from the combination of the circuit elements in the other decoders.
- the decoders 100, 200, 300, 400 therefore have different implementation cost and complexity.
- Each decoder 100, 200, 300, 400 is further configured to decode an encoded texel block to generate at least one texel value.
- the decoding system 1 also comprises a value selector 500 connected to the N decoders 100, 200, 300, 400 and configured to selectively output at least N texel values, preferably N texel values. These at least N texel values are output from at least one of the N decoders 100, 200, 300, 400.
- the first decoder 100 of the N decoders 100, 200, 300, 400 is configured to decode an input encoded texel block to generate N texel values.
- the value selector 500 is then configured to selectively output these N texel values from the first decoder 100 and further to the pixel calculator 500 if the N texels are positioned inside the boundaries of a texel block.
- the four texels 30-36 of a bilinear patch 20 are all positioned inside a single texel block 10
- the four texel values are all generated by the first decoder 100.
- the first decoder 100 has a first circuitry configured to decode an input encoded texel block to generate four texel values.
- the second decoder 200 has a second circuitry and outputs two texel values based on an input encoded texel block.
- the third 300 and fourth 400 decoders comprise a third circuitry and a fourth circuitry, respectively.
- the third 300 and fourth 400 decoders each output a single texel value given an input encoded texel block.
- Fig. 1 D all preferably four decoders 100, 200, 300, 400 output texel values that are selected by the value selector 500.
- the value selector 500 is then configured to selectively output a first texel value from the first decoder 100, a second texel value from the second decoder 200, a third texel value from the third decoder 300 and finally a fourth texel value from the fourth decoder 400.
- each of the four texels 30-36 is positioned within a respective texel block 10-16.
- Fig. 2 schematically illustrates an implementation structure with a decoding system 1 according to an embodiment.
- the structure comprises a memory 5, typically a RAM , having memory locations comprising encoded texel blocks 40.
- a graphical processing unit (GPU) 4 is connected to the memory 5 through a memory bus.
- the GPU 4 includes texel block identifier 2 for identifying a texel block, the encoded block 40 of which should be fetched from the memory 5.
- the texel block identifier 2 basically calculates the address of the encoded texel block 40 in the memory 5 and sends a memory access request over the bus to the memory 5.
- the encoded texel block 40 present in the memory location assigned or dedicated to the texel block is read from the memory 5 and transferred over the bus to the decoding system 1 or to a cache 3 of the GPU 4.
- the decoding system 1 comprises or is connected to at least one cache 3 having cache locations or areas for storing encoded texel blocks or decoded texel values.
- the GPU 4 could comprise multiple caches 3, such as a texture cache and color buffer cache.
- the GPU 4 is typically connected to a display screen 6, on which a decoded image is displayed.
- the GPU 4 and its included units 1 , 2, 3 may be provided in one or more chips, for instance of a graphics card, video card or a motherboard .
- the implementation structure of Fig. 2 can be implemented in any image or texture processing unit or device including, for instance, a computer, a game console, a portable device, such as mobile telephone or media processing device.
- PLANAR is designed to handle texel blocks with smoothly varying chrominances. PLANAR approximates the color components (red, green, blue) in the texel block with a planar approximation. In order to specify a plane, it suffices to specify the colors in three locations in the texel block.
- Fig. 5 illustrates this concept with three red components Ro, RH and Rv at certain positions relative a texel block 10.
- red component can be calculated anywhere in the texel block 10 using the following equation:
- the blue and green components are calculated the same way as the red component.
- An encoded texel block 40 as illustrated in Fig. 4 therefore comprises three color codewords 41 -49 per color component, i.e. Xo, XH and X v , where X is R for the red component, G for the green component and B for the blue component.
- X is R for the red component
- G for the green component
- B for the blue component.
- six bits are preferably spent on each of the red codewords
- the first decoder 100 of Fig. 3 can decode any 2x2 bilinear patch that has all four texels inside the texel block, see Fig. 1A.
- the second decoder 200 is less advanced and produces texel values for texels in the first row or the first column of a texel block, see texel 32, 36 in 15 Fig. 1 B and texels 34, 36 in Fig. 1 C.
- the second decoder 200 therefore preferably outputs two texel values at once, either a 2x1 patch or a 1 x2 patch.
- the fourth decoder is preferably configured to only decode texels at origin. This means that the circuitry of the fourth decoder can be made extremely simple.
- Fig. 6 illustrates the circuitry of the fourth decoder 400 for one of the color components.
- the only structure is an optional but preferred extender 480 that is configured to extend the input first color codeword Xo to a defined bit length, preferably eight bits.
- such an extender 480 can be implemented without any hardware structures and instead only with wires from where the respective color components are stored to the output of the fourth decoder 400.
- the third decoder 300 should, according to the discussion above, preferably handle texels at the
- This expression is more easily implemented in hardware as compared to the above presented one. Shift operators can be implemented by connecting wires differently. The only remaining operations needed to calculate x(3,o) are three additions.
- Fig . 7 illustrates an embodiment of the circuitry of the third decoder 300.
- the third decoder 300 comprises a left shift operator 310 for shifting the second color codeword XH one bit position to the left to get a left-shifted bit sequence.
- the left shift operator 310 is connected to a first adder 320 having carry in set to one.
- This first adder 320 adds the left-shifted bit sequence to the second color codeword XH.
- the resulting output is added to the first color codeword Xo in a second adder 322 having carry in set to one and being connected to the first adder 320.
- the second adder 322 is connected to a right shift operator 330 configured to shift the bit sequence corresponding to the output of the second adder 322 two bit positions to the right to form the final output of third decoder 300, i.e. the texel value
- the first color codeword Xo and the second color codeword XH are preferably first extended or expanded from six or seven bits into eight bits.
- a first extender 380 is preferably arranged connected to the second adder 322 and configured to extend the first color codeword Xo into a pre-defined bit size, preferably eight bits.
- a second extender 382 is preferably arranged connected to the left shift operator 310 and the first adder 320 and is configured to extend or expand the second color codeword XH into the pre-defined bit size, i.e. preferably eight bits.
- FIG. 10 is a schematic illustration of an implementation of the first and second extenders 380, 382 suitable for the green color component. As is seen in the figure, the most significant bit of the input 7-bit sequence is copied to the least significant bit position to get an output 8-bit sequence. For the red and blue components the two most significant bits are copied to the two least significant bit positions to go from a 6-bit sequence to an 8-bit sequence.
- the hardware circuitry of the third decoder 300 can be implemented with a total of 250.8 gates, if each adder costs 4.4 gates per bit:
- Fig. 8 illustrates an embodiment of a hardware implementation of the second decoder 200.
- the second decoder 200 comprises a first multiplexor 250 configured to output one of the second color codeword XH and the third color codeword Xv.
- the selection of color codeword is dependent on whether half of the four texels are positioned in the first row or first column inside the boundary of the texel block, the encoded texel block of which is input to the second decoder 200. If the two pixels are positioned in the first row, the second color codeword XH is output from the first multiplexor 250 otherwise the third color codeword Xv is output.
- a negation operator 240 is configured to negate the first color codeword Xo.
- a first adder 220 is connected to the first multiplexor 250 and the negation operator 240 and adds the output from these units 240, 250 to form the difference XH-XO or X v -Xo.
- the negation operator 240 is replaced by a NOT operator with the first adder 220 being an adder with carry in set to one. Usage of a NOT operator together with an adder with carry in set to one is generally preferred over a negation operator with a normal adder in terms of hardware complexity.
- a second multiplexor 252 is configured to output one of a row coordinate and a column coordinate of a texel.
- This texel is the first texel of the two texels present in the texel block, the encoded texel block of which is input to the second decoder 200.
- Sx denotes the column coordinate
- SY denotes the row coordinate of this texel. If the two texels are positioned in the first row, the column coordinate Sx is output from the second multiplexor 252 otherwise the row coordinate SY.
- a third multiplexor 254 is connected to the left shift operator 210 and the first adder 220. The third multiplexor 254 selects between i.e.
- An OR operator 360 is connected to the second multiplexor 352 and is configured to perform an OR operation on the most significant bit and the least significant bit of the row or column coordinate from the second multiplexor 352.
- the OR operator 360 and the third multiplexor 254 are further connected to an AND operator 270.
- This AND operator 270 performs an AND operation of the multiplexor output bitwise with the single bit from the OR operator 260. The result here is that the multiplexor output is preserved if the OR output is 1 bin and the output from the AND operator 270 is zero if the OR output is Obin. Hence, the output from the AND operator 270 is 2d, d or 0.
- a second adder 222 is connected to the AND operator 270 and is configured to add the output of the AND operator 270 and a bit sequence having 10bin as the least two significant bits preceded by the first color codeword Xo.
- the output from the second adder 222 is therefore 2d+f, d+f or f.
- a third adder 224 is connected to the first adder 220 and the second adder 222 and is configured to add the outputs from these adders 220, 222. This third adder 224, hence, gives the result 3d+f, 2d+f or d+f.
- a first right shift operator 230 is connected to the third adder 224 and is configured to shift the bit sequence from the third adder 224 two bit positions to the right to get the texel value of the lower or the right texel of the two texels in the first column or the first row in the bilinear patch.
- the right shift operators 230, 232 are preferably connected to a respective clamper or clamp operator 290, 292.
- These clampers 290, 292 are configured to clamp the output from the right shift operators 230, 232 between a minimum texel value and a maximum texel value.
- the clampers 290, 292 make sure that the output bit sequence from the second decoder 200 is within an allowed value interval.
- An example of such an allowed value interval is from 0 to 255 if the output sequences from the second decoder 200 should be of 8 bits.
- Fig. 1 1 illustrates an example of a hardware implementation of a clamper 190-196, 290, 292, such as employed in the second decoder of Fig. 8 or the first decoder of Fig. 9.
- This clamper embodiment is designed to clamp a 1 1-bit sequence potentially representing a value in the interval of [-1024, 1023] to a 8-bit sequence representing a value in the interval of [0, 255].
- OR operators 701 -708 each receiving one of the eight least significant bits of the input bit sequence.
- the other respective inputs to the OR operators 701 -708 are connected to the output of another OR operator 700.
- This OR operator 700 receives the two most significant bits of the input bit sequence.
- the outputs of the eight OR operators 701 -708 are each connected to a respective AND operator 71 1 -718.
- the other respective input of the AND operators 71 1 -718 receives the sign bit of the
- FIG. 12 An alternative implementation embodiment for the clamper 190-196, 290, 292 is illustrated in Fig. 12.
- the eight OR operators are replaced by eight N OR operators 721 -724.
- the eight AND operators of Fig. 1 1 are also replaced by eight NOR operators 731 - 20 738.
- a NOT operator 740 is connected between the eight NOR operators 731 -738 and the input for the sign bit of the input bit sequence.
- the number of gates for the implementation example of the second decoder 200 in Fig. 8, excluding the clampers, is:
- Fig. 9 is an example of a hardware implementation of the first decoder 100.
- the first decoder 15 100 comprises a negation operator 140 configured to negate the first color codeword Xo.
- the negation operator 140 is replaced by a NOT operator.
- the first adder 20 120 and the second adder 122 then both have carry in set to one.
- a first left shift operator 1 10 is connected to the first adder 120 and shifts the adder output one bit position to the left get a left-shifted bit sequence corresponding to 2g.
- a second left shift operator 1 12 is connected to the second adder 122 and shifts the adder output one bit position to the left to get the bit
- a first multiplexor 150 is connected to the first left shift operator 1 10 and the first adder 120 and is configured to select between g and 2g based on the most significant bit of a column coordinate Sx.
- a second multiplexor 152 is correspondingly connected to the second left shift operator 1 12 and the second adder 122 and selects between h and 2h based on the most significant bit of a row coordinate SY.
- the column and row coordinates are the coordinates of the top left texel in the 2x2
- a first OR operator 160 performs an OR operation on the most significant bit and the least significant bit of the column coordinate Sx.
- a second OR operator 162 performs an OR operation on the most significant bit and the least significant bit of the row coordinate SY.
- the first OR operator 160 and the first multiplexor 150 are connected to a first AND operator 170 that performs a bitwise AND operation on the multiplexor output with the single bit from the first OR operator 160.
- the output from the first AND operator 170 is therefore 2g, g or 0.
- a second AND operator 172 is connected to the second multiplexor 152 and the second OR operator 162.
- the second AND operator 172 performs a bitwise AND operation on the multiplexor output with the bit from the second OR operator 162 to get an output of 2h, h or 0.
- a fourth adder 126 is connected to the first adder 120 and the third adder 124 and is configured to add the outputs from these adders 120, 124 to get 3g+f, 2g+f or g+f.
- a fifth adder 128 is connected to the second adder 122 and the second AND operator 172 and adds the outputs from these two units 122, 172 to get 3h, 2h or h.
- a sixth adder 121 is connected to the third adder 124 and the second AND operator 172 and is configured to add the outputs from these units 124, 172.
- a first right shift operator 130 is connected to the sixth adder 121 and is configured to shift the bit sequence of the adder output two bit positions to the right to get texel value of texel 30 in Fig. 1 A.
- a seventh adder 123 is connected to the fourth adder 126 and the second AND operator 172 and is configured to add the outputs form these units 126, 172.
- a second right shift operator 132 is connected to the seventh adder 123 and is configured to shift the bit sequence of the adder output two bit positions to the right to get texel value of texel 32 in Fig. 1A.
- An eighth adder 125 is connected to the third adder 124 and the fifth adder 128 and is configured to add the outputs form these adders 124, 128.
- a third right shift operator 134 is connected to the eighth adder 125 and is configured to shift the bit sequence of the adder output two bit positions to the right to get texel value of texel 34 in Fig. 1A.
- a ninth adder 127 is connected to the fourth adder 126 and the fifth adder 128 and is configured to add the outputs form these adders 126, 128.
- a fourth right shift operator 136 is connected to the fourth adder 127 and is configured to shift the bit sequence of the adder output two bit positions to the right to get texel value of texel 36 in Fig. 1A.
- the input color codewords Xo, XH, XV are first extended or expanded from six or seven bits into eight bits.
- a first bit extender 180 is preferably arranged connected to the negation operator 140 and further configured to extend or expand the first color codeword Xo into the pre-defined bit sequence, preferably eight bits, before 10bin is added as least significant bits to get a 10-bit sequence.
- a second extender 182 is correspondingly arranged connected to the first adder 120 and configured to extend the second color codeword XH.
- a third extender 184 is connected to the second adder 122 and is configured to extend the third color codeword Xv.
- each right shift operator 130-136 is preferably connected to a respective clamper or clamp operator 190-196.
- These clampers 190-196 are configured to clamp the output from the right shift operators 130-136 between a minimum texel value and a maximum texel value.
- the clampers 190-196 make sure that the output bit sequence from the first decoder 100 is within an allowed value interval.
- An example of such an allowed value interval is from 0 to 255 if the output sequences from the first decoder 100 should be of 8 bits.
- the number of gates for the implementation example of Fig. 8 excluding the optional clampers 190-196 is:
- a third multiplexor 530 is connected to the third and fourth outputs of the first decoder 100, the first output of the second decoder 200 and the single output of the third decoder 300.
- the third multiplexor 530 outputs the texel value for the third texel 34 in the bilinear patch.
- the fourth multiplexor 540 is connected to the fourth output of the first decoder 100, the second output of the second decoder 200 and the single output of the fourth decoder 400.
- the fourth multiplexor 540 selects and outputs the texel value of the fourth texel 36 in the bilinear patch.
- 1 B implies that the first multiplexor 510 selects the bit sequence of the second output of the first decoder 100.
- the second multipelxor 520 selects the bit sequence of the first output of the second decoder 200.
- the third decoder 300 selects the bit sequence of the fourth output from the first decoder 100 and the fourth decoder 400 selects the bit sequence from the second output of the second decoder 200.
- the concept can then be applied in a similar way to the 15 two other patch positions as illustrated in Figs. 1 C and 1 D.
- the pixel calculator of the decoding system can be implemented in hardware according to techniques well-known in the art. In more detail, assume that the four texel values, in the case of bilinear filtering, from the value selector are denoted A, B, C and D.
- the above described equation is performed in parallel for the red, green and blue color components.
- a hardware implementation can then be defined based on the equation. For instance, assume that the position of the sample point can be expressed with three binary decimals, e.g. the dx position can assume the following values [0.000, 0.125, 0.250, 0.375, 0.500, 0.625, 0.750, 0.875]. In such a case, the position can be defined in eights so that the values above can be represented by [0, 1 , 2, 3, 4, 5, 6, 7].
- the output from the multiplication operator should then be multiplied by the texel value B, preferably a number between 0 and 255.
- This multiplication can be handled by a multiplication operator taking two 8-bit inputs and outputs a 16-bit bit sequence. However, only the fourteen least significant bits from the multiplication operator are needed since the largest possible number is 56x255 ⁇ 2 14 .
- a corresponding implementation is then performed for the other terms in the equation and the outputs from the respective final multiplication operators are added and shifted six steps to the right in order to get the final 8-bit pixel value.
- the above described implementation examples should merely be seen as preferred but non-limiting embodiments of how the units of the decoding system can be implemented in hardware. Various variants thereof are possible and within the scope of the invention. The important feature is though that the circuitry of the including decoders are different with increasing complexity and that each decoder is configured to handle and output the texel value(s) of one or more texels at predefined positions in the patch.
- Fig. 14 is a flow diagram illustrating a method of generating a decoded pixel by decoding at least one encoded texel block.
- the method starts in step S1 there the position of at least N texel values, i.e. a patch, relative a texel block border is determined.
- a respective encoded texel block is input to N decoders to get decoded texel values.
- a next step S2 selectively outputs at least N texel values from at least one of the N decoders.
- Step S2 preferably involves selectively outputting all N texel values from a first decoder of the N decoders if the N texels are positioned inside the boundary of a texel block. If however half of the N texels are positioned inside one texel block and the remaining half of the texels are instead found inside a neighboring texel block, half of the N texel values are preferably selected from the output of the first decoder with the remaining half from a second decoder.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Graphics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Image Generation (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
Claims
Priority Applications (10)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020127015254A KR101709822B1 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable on encoded texture element blocks |
EP10833652.0A EP2504814B1 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable on encoded texture element blocks |
CN201080053340.0A CN102687175B (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable on encoded texture element blocks |
JP2012539846A JP2013512482A (en) | 2009-11-24 | 2010-02-24 | Decoding system and method that can be used for encoded texture element blocks |
PL10833652T PL2504814T3 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable on encoded texture element blocks |
ES10833652.0T ES2635540T3 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method usable in blocks of coded texture elements |
RU2012126120/08A RU2510939C2 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable to encode texture element blocks |
DK10833652.0T DK2504814T3 (en) | 2009-11-24 | 2010-02-24 | DECODING SYSTEM AND PROCEDURE TO USE ON CODED TEXTURE ELEMENT BLOCKS |
BR112012012273-8A BR112012012273B1 (en) | 2009-11-24 | 2010-02-24 | decoding system and method for generating a decoded pixel |
US13/510,992 US8655063B2 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable on encoded texture element blocks |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US26392209P | 2009-11-24 | 2009-11-24 | |
US61/263,922 | 2009-11-24 | ||
US26631609P | 2009-12-03 | 2009-12-03 | |
US61/266,316 | 2009-12-03 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2011065886A1 true WO2011065886A1 (en) | 2011-06-03 |
WO2011065886A8 WO2011065886A8 (en) | 2012-07-19 |
Family
ID=44066774
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/SE2010/050217 WO2011065886A1 (en) | 2009-11-24 | 2010-02-24 | Decoding system and method operable on encoded texture element blocks |
Country Status (11)
Country | Link |
---|---|
US (1) | US8655063B2 (en) |
EP (1) | EP2504814B1 (en) |
JP (1) | JP2013512482A (en) |
KR (1) | KR101709822B1 (en) |
CN (1) | CN102687175B (en) |
BR (1) | BR112012012273B1 (en) |
DK (1) | DK2504814T3 (en) |
ES (1) | ES2635540T3 (en) |
PL (1) | PL2504814T3 (en) |
RU (1) | RU2510939C2 (en) |
WO (1) | WO2011065886A1 (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013009224A1 (en) * | 2011-07-08 | 2013-01-17 | Telefonaktiebolaget L M Ericsson (Publ) | Multi-mode processing of texture blocks |
EP3396627A1 (en) * | 2017-04-28 | 2018-10-31 | Imagination Technologies Limited | Decoder unit for texture decompression |
EP3396628A1 (en) * | 2017-04-28 | 2018-10-31 | Imagination Technologies Limited | Multi-output decoder for texture decompression |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10148972B2 (en) * | 2016-01-08 | 2018-12-04 | Futurewei Technologies, Inc. | JPEG image to compressed GPU texture transcoder |
US10055810B2 (en) * | 2016-03-04 | 2018-08-21 | Samsung Electronics Co., Ltd. | Cache architecture for efficiently accessing texture data using buffers |
US11244492B2 (en) * | 2018-11-02 | 2022-02-08 | Facebook Technologies, Llc. | Parallel texture sampling |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1227444A1 (en) * | 1999-04-15 | 2002-07-31 | SP3D Chip Design GmbH | Method of rasterizing a graphics basic component |
WO2005059837A1 (en) * | 2003-12-19 | 2005-06-30 | Telefonaktiebolaget Lm Ericsson (Publ) | Alpha image processing |
WO2005059839A1 (en) * | 2003-12-19 | 2005-06-30 | Telefonaktiebolaget Lm Ericsson (Publ) | Multi-mode image processing |
WO2006006915A1 (en) * | 2004-07-08 | 2006-01-19 | Telefonaktiebolaget Lm Ericsson (Publ) | Multi-mode image processing |
WO2007084062A2 (en) * | 2006-01-23 | 2007-07-26 | Telefonaktiebolaget Lm Ericsson (Publ). | Image processing |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6111582A (en) * | 1996-12-20 | 2000-08-29 | Jenkins; Barry L. | System and method of image generation and encoding using primitive reprojection |
US5956431A (en) | 1997-10-02 | 1999-09-21 | S3 Incorporated | System and method for fixed-rate block-based image compression with inferred pixel values |
JPH11341496A (en) * | 1998-05-28 | 1999-12-10 | Matsushita Electric Ind Co Ltd | Image processing method, image processing unit and data storage medium |
US6968006B1 (en) * | 2001-06-05 | 2005-11-22 | At&T Corp. | Method of content adaptive video decoding |
SE526226C2 (en) * | 2003-12-19 | 2005-08-02 | Ericsson Telefon Ab L M | Image Processing |
WO2006126949A1 (en) | 2005-05-27 | 2006-11-30 | Telefonaktiebolaget Lm Ericsson (Publ) | Weight based image processing |
WO2007116551A1 (en) * | 2006-03-30 | 2007-10-18 | Kabushiki Kaisha Toshiba | Image coding apparatus and image coding method, and image decoding apparatus and image decoding method |
EP2023637A4 (en) * | 2006-05-24 | 2010-06-16 | Panasonic Corp | Image decoding device |
US7890750B2 (en) * | 2006-07-06 | 2011-02-15 | Accenture Global Services Limited | Encryption and decryption on a graphics processing unit |
US9967590B2 (en) * | 2008-04-10 | 2018-05-08 | Qualcomm Incorporated | Rate-distortion defined interpolation for video coding based on fixed filter or adaptive filter |
-
2010
- 2010-02-24 KR KR1020127015254A patent/KR101709822B1/en active IP Right Grant
- 2010-02-24 PL PL10833652T patent/PL2504814T3/en unknown
- 2010-02-24 EP EP10833652.0A patent/EP2504814B1/en active Active
- 2010-02-24 US US13/510,992 patent/US8655063B2/en active Active
- 2010-02-24 RU RU2012126120/08A patent/RU2510939C2/en active
- 2010-02-24 WO PCT/SE2010/050217 patent/WO2011065886A1/en active Application Filing
- 2010-02-24 BR BR112012012273-8A patent/BR112012012273B1/en active IP Right Grant
- 2010-02-24 JP JP2012539846A patent/JP2013512482A/en active Pending
- 2010-02-24 ES ES10833652.0T patent/ES2635540T3/en active Active
- 2010-02-24 DK DK10833652.0T patent/DK2504814T3/en active
- 2010-02-24 CN CN201080053340.0A patent/CN102687175B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1227444A1 (en) * | 1999-04-15 | 2002-07-31 | SP3D Chip Design GmbH | Method of rasterizing a graphics basic component |
WO2005059837A1 (en) * | 2003-12-19 | 2005-06-30 | Telefonaktiebolaget Lm Ericsson (Publ) | Alpha image processing |
WO2005059839A1 (en) * | 2003-12-19 | 2005-06-30 | Telefonaktiebolaget Lm Ericsson (Publ) | Multi-mode image processing |
WO2006006915A1 (en) * | 2004-07-08 | 2006-01-19 | Telefonaktiebolaget Lm Ericsson (Publ) | Multi-mode image processing |
WO2007084062A2 (en) * | 2006-01-23 | 2007-07-26 | Telefonaktiebolaget Lm Ericsson (Publ). | Image processing |
Non-Patent Citations (3)
Title |
---|
DATABASE INSPEC Database accession no. 7257600 * |
PARK S.-J. ET AL: "A reconfigurable multilevel parallel texture cache memory with 75-GB/s parallel cache replacement bandwidth", IEEE JOURNAL OF SOLID-STATE CIRCUITS, vol. 37, no. 5, May 2002 (2002-05-01), pages 612 - 623, XP011065764 * |
See also references of EP2504814A4 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013009224A1 (en) * | 2011-07-08 | 2013-01-17 | Telefonaktiebolaget L M Ericsson (Publ) | Multi-mode processing of texture blocks |
US8494291B2 (en) | 2011-07-08 | 2013-07-23 | Telefonaktiebolaget Lm Ericsson (Publ) | Multi-mode processing of texture blocks |
US8837842B2 (en) | 2011-07-08 | 2014-09-16 | Telefonaktiebolaget Lm Ericsson (Publ) | Multi-mode processing of texture blocks |
EP3396627A1 (en) * | 2017-04-28 | 2018-10-31 | Imagination Technologies Limited | Decoder unit for texture decompression |
EP3396628A1 (en) * | 2017-04-28 | 2018-10-31 | Imagination Technologies Limited | Multi-output decoder for texture decompression |
CN108810544A (en) * | 2017-04-28 | 2018-11-13 | 想象技术有限公司 | Multi output decoder for texture decompression |
US10636195B2 (en) | 2017-04-28 | 2020-04-28 | Imagination Technologies Limited | Decoder unit for texture decompression |
GB2562042B (en) * | 2017-04-28 | 2020-07-15 | Imagination Tech Ltd | Decoder unit for texture decompression |
GB2584764A (en) * | 2017-04-28 | 2020-12-16 | Imagination Tech Ltd | Decoder unit for texture decompression |
US10937198B2 (en) | 2017-04-28 | 2021-03-02 | Imagination Technologies Limited | Multi-output decoder for texture decompression |
US11043020B2 (en) | 2017-04-28 | 2021-06-22 | Imagination Technologies Limited | Decoder unit for texture decompression |
GB2584764B (en) * | 2017-04-28 | 2021-12-29 | Imagination Tech Ltd | Decoder unit for texture decompression |
EP3940646A1 (en) * | 2017-04-28 | 2022-01-19 | Imagination Technologies Limited | Decoder unit for texture decompression |
US11544892B2 (en) | 2017-04-28 | 2023-01-03 | Imagination Technologies Limited | Decoder unit for texture decompression |
US11836830B2 (en) | 2017-04-28 | 2023-12-05 | Imagination Technologies Limited | Multi-output decoder for texture decompression |
CN108810544B (en) * | 2017-04-28 | 2024-03-01 | 想象技术有限公司 | Multi-output decoder for texture decompression |
Also Published As
Publication number | Publication date |
---|---|
EP2504814B1 (en) | 2017-05-03 |
US20120281915A1 (en) | 2012-11-08 |
KR101709822B1 (en) | 2017-02-23 |
PL2504814T3 (en) | 2017-10-31 |
RU2510939C2 (en) | 2014-04-10 |
US8655063B2 (en) | 2014-02-18 |
BR112012012273A2 (en) | 2018-04-17 |
ES2635540T3 (en) | 2017-10-04 |
EP2504814A1 (en) | 2012-10-03 |
CN102687175B (en) | 2015-02-04 |
DK2504814T3 (en) | 2017-08-21 |
CN102687175A (en) | 2012-09-19 |
JP2013512482A (en) | 2013-04-11 |
BR112012012273B1 (en) | 2020-11-03 |
WO2011065886A8 (en) | 2012-07-19 |
EP2504814A4 (en) | 2016-04-06 |
RU2012126120A (en) | 2013-12-27 |
KR20120097515A (en) | 2012-09-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP2005393B1 (en) | High quality image processing | |
US9159114B2 (en) | Texture decompression for graphics processors | |
US8204324B2 (en) | Method and apparatus for compressing and decompressing data | |
US8625910B2 (en) | Compression of image data | |
US20050264573A1 (en) | Graphics texture processing methods, apparatus and computer program products using texture compression, block overlapping and/or texture filtering | |
EP2504814B1 (en) | Decoding system and method operable on encoded texture element blocks | |
JP2006014341A (en) | Method and apparatus for storing image data using mcu buffer | |
CN110782387B (en) | Image processing method and device, image processor and electronic equipment | |
EP2797049B1 (en) | Color buffer compression | |
US9111328B2 (en) | Texture compression and decompression | |
WO2008123823A1 (en) | Vector-based image processing | |
EP2119245B1 (en) | Programmable pattern-based unpacking and packing of data channel information | |
US20180097527A1 (en) | 32-bit hdr pixel format with optimum precision | |
US9129543B2 (en) | Texture compression and decompression | |
Krause | ftc—floating precision texture compression | |
EP2652707A1 (en) | Tile encoding and decoding | |
EP3085096A1 (en) | Techniques for image bitstream processing | |
WO2024137230A1 (en) | Workload packing in graphics texture pipeline | |
Nawandhar | 3D Graphics Texture Compression and Its Recent Trends |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201080053340.0 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10833652 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 4221/DELNP/2012 Country of ref document: IN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012539846 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13510992 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
REEP | Request for entry into the european phase |
Ref document number: 2010833652 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010833652 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 20127015254 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012126120 Country of ref document: RU |
|
REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112012012273 Country of ref document: BR |
|
ENP | Entry into the national phase |
Ref document number: 112012012273 Country of ref document: BR Kind code of ref document: A2 Effective date: 20120522 |