CN101273360B - Error diffusion for display frame buffer saving electricity - Google Patents
Error diffusion for display frame buffer saving electricity Download PDFInfo
- Publication number
- CN101273360B CN101273360B CN2005800517021A CN200580051702A CN101273360B CN 101273360 B CN101273360 B CN 101273360B CN 2005800517021 A CN2005800517021 A CN 2005800517021A CN 200580051702 A CN200580051702 A CN 200580051702A CN 101273360 B CN101273360 B CN 101273360B
- Authority
- CN
- China
- Prior art keywords
- pixel
- frame buffer
- section
- error diffusion
- color
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Control Of Indicators Other Than Cathode Ray Tubes (AREA)
- Liquid Crystal Display Device Control (AREA)
Abstract
Methods and apparatuses for error diffusion for display frame buffer power saving are described herein. According to one embodiment, pixels of a color plane of image data are stored in a first segment and a second segment of a frame buffer during a normal power state. During a low power state, an error diffusion operation is performed on the pixels to reduce a color depth of the pixels. Thereafter, at least a portion of the pixels with reduced color depth is stored in the first segment of the frame buffer during the low power state without accessing the second segment of the frame buffer. Other methods and apparatuses are also described.
Description
Technical field
Embodiments of the invention relate to data handling system; Relate more specifically to be used for the error diffusion of the display frame buffer of data handling system.
Background technology
Now, P/PC, PDA (personal digital assistant), mobile phone, digital camera and camcorder all are equipped with colored TFT (thin film transistor (TFT)), LCD (LCD), STN (super-twist nematic) LCD or OLED (Organic Light Emitting Diode) that power efficiency is high.Display subsystem has become main big power consumer, even system disposition has powerful CPU and the SDRAM of tens of MB.In LCD and OLED, frame buffer is indispensable ingredient, and it has consumed most of energy.
One of low power frame buffer techniques is the dynamic-color-depth control technology, has wherein proposed a kind of new pixel organization and has realized LSD (minimum effective equipment) in the closure frame impact damper.In this case, reduced the power consumption of frame Buffer Memory and interrelated logic.Yet this technology has caused the pseudo-shadow of the obvious false contouring in the image/video,
Also proposed other technology and reduced power in the display subsystem, comprised that the variable duty ratio that reduces power by the refresh rate that reduces frame buffer refreshes and change by reducing backlight illumination simultaneously and regulating the brightness and contrast that brightness of image reduces power.Yet these technology have caused tangible image/video quality degeneration.
In addition, proposed to reduce the frame buffer that mobile phone is used based on the technology of shadow tone.In this technology, 24 original RGB color data is converted into 18 or 15 s' color data, and uses the ordered dithering technology to remove to quantize the pseudo-shadow of false contouring that causes.Yet it has produced the pseudo-shadow of the model of being introduced by the fixed threshold matrix (pattern).
In addition, the power consumption that a kind of frame buffer compression technique reduces display subsystem has been proposed.In this technology, frame buffer is divided into not page compression and page compression.During normal power mode, raw data is sent to not in the page compression at first.Under low-power mode, scrambler with compressing original data in page compression, and LCD panel refresh operation only visit have the still less page compression of position.Therefore, the quantity of frame buffer bit access and corresponding power consumption have all obtained reduction.Yet this technology is invalid on image.
In addition, error diffusion technique has been widely used in the printing field.Error diffusion technique is the ad hoc approach of shadow tone, and it uses a small amount of color to represent more color.In this technology, to each image pixel execution algorithm function.For the red, green and blue component of each pixel, the error between the value after calculating original value and the quantification.Then, appear at the error diffusion of center pixel to surrounding pixel with being defined as.Yet this technology comprises the complicated calculations of processing power that need be big relatively.
Description of drawings
By with reference to following instructions and the accompanying drawing that is used to illustrate various embodiments of the present invention, can understand embodiments of the invention best.In the accompanying drawings:
Fig. 1 shows as conventional method with according to the result's of the method for some embodiment various images;
Fig. 2 is the figure that illustrates according to the data layout of the frame buffer of an embodiment;
Fig. 3-the 5th illustrates the block scheme of the various embodiment of display subsystem;
Fig. 6 A is the process flow diagram that illustrates according to the process example of its subsystem of demonstration of an embodiment;
Fig. 6 B is the process flow diagram that illustrates according to the process example of the error diffusion operation of an embodiment;
Fig. 7 is the pseudo-code according to the error diffusion operation of an embodiment;
Fig. 8 is the process flow diagram that illustrates according to the process example of the error diffusion operation of another embodiment;
Fig. 9 is the pseudo-code according to the error diffusion operation of another embodiment; And
Figure 10 is the block scheme that the example of the data handling system that can be used as the embodiment use is shown.
Embodiment
This paper has described the method and apparatus of the error diffusion that is used for display frame buffer saving electricity.According to some aspect of the present invention, adopt the LSD of effective low-power display technique with closure frame buffer memory when keeping picture quality.According to an embodiment, utilized the error diffusion algorithm simplified (for example, be implemented as " error diffusion encoder "), it will have panchromatic dark image and be encoded into the image of the color depth with reduction under the situation that does not produce the pseudo-shadow of false contouring, as utilize shown in the image 103 that the image 102 of Fig. 1 generated." error diffusion encoder " can be at display driver, realizes in application software and/or the hardware module such as display controller, GMCH (figure and memory controlling hub) and/or equivalent chip.
Compare with the dynamic-color-depth control technology of routine, the techniques described herein can overcome the pseudo-shadow of the caused false contouring of the LSD of direct closure frame buffer memory, as utilize shown in the image 101 that the image 102 of Fig. 1 generated.Compare with the run-length encoding frame buffer compression technique of routine, the techniques described herein are more effective for image and video content and text and graphical content.Though routine techniques is only effective to text and simple graphical content, this is because run-length encoding algorithm compressed image and video content very effectively, and this makes that the economize on electricity of image and video is very limited.Compare with the dither technique of routine, the complicacy of the techniques described herein is lower, and it can be with color depth, for example, be reduced to 8 or 9 from 24, and routine techniques only is reduced to color depth 15 or 18.
In the following description, set forth a large amount of details and come more in depth to introduce embodiments of the invention.Yet, it will be apparent for a person skilled in the art that embodiments of the invention can implement under the situation of these details not having.In other example, for fear of embodiments of the invention are beyonded one's depth, illustrate rather than at length show known structure and equipment with the form of block scheme.
" embodiment " who quotes in the instructions or " embodiment " are meant that described in conjunction with the embodiments special characteristic, structure or characteristics comprise at least one embodiment of the present invention.The phrase " in one embodiment " that occurs in many places in the instructions might not all refer to same embodiment." image " quoted in the instructions is meant a frame of image, video or the 2D/3D graphical content that is used to show of storing with image data format.
According to an embodiment, proposed a kind of error diffusion mechanism and handled the view data that writes frame Buffer Memory, so that when saving power consumption, remove these pseudo-shadows.
In one embodiment, can in the hardware of display subsystem and/or software, realize error diffusion encoder, to remove image artifacts when the LSD of closure frame buffer memory in order to reduce power.In addition, according to an embodiment, utilize contiguous error-filter of 2-and fault bit short-cut method to reduce the computation complexity of classical error diffusion algorithm, thereby reduce the energy consumption and the cost of error diffusion encoder.In addition, can utilize some software of " error diffusion encoder " and/or the display subsystem that hardware embodiment provides low-power working mode.Although error diffusion encoder can realize in display controller, similar design can be applied in the display subsystem with different color depths and data memory format, to realize similar function.
The use error diffusion technique can represent to have number of colors by a small amount of color under the situation that does not have the pseudo-shadow of false contouring image. in one embodiment of the invention, use error diffusion technique cause has the pixel of the color depth of reduction and represents to have panchromatic dark original image. in a particular embodiment, with improved algorithm (for example as scrambler, " error diffusion encoder "), be used for needs close or the power of LSD that reduces frame Buffer Memory when reducing power consumption, original image is transformed into the image with low color depth, to be fit to the MSD. of frame buffer
According to an embodiment, can recombinate to the position order in the frame buffer.With reference to Fig. 2, conventional physical arrangement 201 comprises 24 colored pixels.Suppose in low-power working mode this color of pixel degree of depth to be compressed to 8 (being respectively 3,3 and 2 for R, G and B), the physical arrangement of this pixel position can be reassembled as 202, and wherein MSD can realize in for example different memory banks with LSD.Therefore, can realize closing LSD.Be different from conventional dynamic-color-depth control technology, the embodiment of this method in the hardware module of display driver, display controller and/or equivalence (for example, software, hardware or the two combination) the middle error diffusion module of simplifying (" error diffusion encoder ") that increases, to remove the image artifacts that causes by the color depth that reduces.
" error diffusion encoder " can be implemented in display driver and/or for example in processor or controller (for example, firmware, DSP or CPU etc.) in the application software of operation.When closing LSD or reduce the power of LSD, the video data of revising is packed among the MSD.
Fig. 3 is the block scheme that display subsystem according to an embodiment of the invention is shown.In one embodiment, display system 300 is including, but not limited to the display controller 303 that is coupled to the error diffusion encoder 301 of frame buffer 302 and is coupled to the display device 304 such as LCD or OLED display panel.In this embodiment, error diffusion encoder 301 may be implemented as software encoder, and it may be implemented as application program, device driver and/or firmware.Frame buffer 302 can be the storer of the specific assigned of primary memory (not shown).Perhaps, frame buffer 302 can be the private memory in chipset (for example, Memory Controller, not shown) or the display controller 303.In addition, frame buffer 302 can be a Stand Alone Memory chipset special-purpose in the display subsystem.Display controller 303 can be implemented in chipset or the independent displaying controller (for example, PCI (peripheral parts interconnected), PCI-Express or AGP (Accelerated Graphics Port) compatible controller).In one embodiment, software diffusion encoder 301 processes and displays data and the video data of handling is stored in the frame buffer 302.After this, display controller 303 fetches data from frame buffer and video data is sent to display device 304 to show.
With reference to Fig. 3, according to an embodiment, display subsystem 300 provides two kinds of display modes: 1) normal mode of operation or color power mode and 2) low-power working mode.In normal mode of operation, raw display data is similar to the position order 202 of Fig. 2 by the position order of directly packing into and is not added in the frame buffer of any change.The display controller 303 that software (for example, comprising error diffusion encoder 301) is controlled can be selected path (1) to visit whole frame buffers to refresh to carry out display panel.When being in low-power working mode, software " error diffusion encoder " 301 is transformed into original image the pixel with reduction color depth, it is fit to size and the tissue of MSD under the situation that does not produce the pseudo-shadow of false contouring, software " error diffusion encoder " 301 writes result data among the MSD of frame buffer 302 then.When display controller 303 refreshed display device 304, it just took out data by path (2) from the MSD of frame buffer 302, and with these data display device 304 of packing into, and the data of least significant bit (LSB) are used as zero or predetermined value.
Under low-power working mode, only power to the MSD and the interrelated logic of frame buffer, and all other parts (for example, the LSD of frame buffer) can cut off the power supply. according to an embodiment, (for example coloured image with 24 or 16 color depths can be converted to 9,3 in every kind of color) or 8 (for example, R and G are 3, B is 2), and do not have tangible pseudo-shadow. therefore, this embodiment can close about 43%～66% frame buffer to save the power consumption of display subsystem about 43%～66%. and this embodiment typically is used for picture material and changes those application relatively slowly, for example, these application comprise e-book, the picture/text editing machine, web browser etc.
Fig. 4 illustrates the block scheme of display subsystem according to another embodiment of the present invention.In this embodiment, " error diffusion encoder " can be implemented in the hardware.For example, error diffusion encoder may be implemented as the part of display controller and/or chipset.In one embodiment, display system 400 including, but not limited to: be implemented in the error diffusion encoder 401 in the display controller 403, wherein display controller 403 is coupled to one or more frame buffers 402 and the display device 404 such as LCD or OLED display panel.Frame buffer 402 can be the storer of the specific assigned of primary memory (not shown).Perhaps, frame buffer 402 can be the private memory in chipset (for example, Memory Controller, not shown) or the display controller 403.In addition, frame buffer 402 can be a Stand Alone Memory chipset special-purpose in the display subsystem.Display controller 403 can be implemented in chipset or the independent displaying controller (for example, PCI, PCI-Express or AGP compatible controller).In one embodiment, error diffusion encoder 401 processes and displays data and the video data of handling is stored in the frame buffer 402.After this, display controller 403 fetches data from frame buffer 402, and video data is sent to display device 404, to show.
Fig. 4 shows the example of display controller implementation, and wherein hardware module is also carried out error diffusion algorithm except the reorganization of carrying out frame buffer bit.Under normal mode of operation, with reference to Fig. 4, data stream is advanced through (1) along the road, raw image data is write among the MSD and LSD of frame buffer 402 by this path (1), and raw image data is all addressable, refreshes so that carry out display panel.Under low-power working mode, data stream along the path (2) advance." error diffusion encoder " 401 is transformed into the pixel with reduction color depth with original image, and it is fit to MSD under the situation that does not produce the pseudo-shadow of false contouring size and tissue then, write result data among the MSD.Under low-power mode, can remove or reduce the power of LSD.
When display controller 403 refreshed display device 404, display controller 403 just took out data by path (2) from MSD, and the data of least significant bit (LSB) are used as zero or predetermined value.Apparently, in this case, same section of frame Buffer Memory (for example LSD) and associated bus lines can be cut off the power supply.In this embodiment, hardware simulation results shows that the power consumption of " error diffusion encoder " is probably low to 0.05% of frame Buffer Memory, and this can be left in the basket when the economize on electricity of computing system.Therefore, in this embodiment, approximately can reduce by 43%～66% power consumption of frame buffer and interrelated logic.This embodiment typically is used for various application, for example line-based browser, image viewer, movie player etc.
According to another embodiment, can use a plurality of frame buffers or frame buffer zone.Fig. 5 illustrates the block scheme of display subsystem according to another embodiment of the present invention.In this embodiment, a plurality of frame buffers or frame buffer zone have been realized.In one embodiment, display system 500 including, but not limited to: be implemented in the error diffusion encoder 501 in the display controller 503, wherein display controller 503 is coupled to a plurality of frame buffer 502A-502B and the display device 504 such as LCD or OLED display panel.
Frame buffer 502A and 502B can be the storeies of the specific assigned of primary memory (not shown).Perhaps, frame buffer 502A and 502B can be the private memories in chipset (for example, Memory Controller, not shown) or the display controller 503.In addition, frame buffer 502A and 502B can be Stand Alone Memory chipsets special-purpose in the display subsystem.In a particular embodiment, frame buffer 502A and 502B can power on separately and cut off the power supply.Display controller 503 can be implemented in chipset or the independent displaying controller (for example, PCI, PCI-Express or AGP compatible controller).In one embodiment, error diffusion encoder 501 processes and displays data, and the video data of handling is stored among frame buffer 502A and the 502B.After this, display controller 503 fetches data from frame buffer 502A and 502B, and video data is sent to display device 504, to show.
With reference to Fig. 5, in one embodiment, use " unpressed frame buffer " (UFB) 502A is stored original image pixels during normal power mode (being also referred to as first power rating), and use other " frame buffer of compression " (CFB) 502B under low-power working mode, store result images .UFB 502A and have MSD and LSD with low bit depth, the two can be organized according to traditional bit order shown in Figure 2 201 or the position that proposes order 202.CFB502B only has MSD, and the size of the MSD in its size and the configuration shown in Figure 4 is similar.Under normal mode of operation, according to an embodiment, view data is written among the UFB 502A, reads view data by path (1) then, refreshes to carry out display panel.
When being in low-power working mode (being also referred to as the low power state or second power rating), " error diffusion encoder " 501 is transformed into original image the pixel with reduction color depth, it is fit to size and the tissue of CFB under the situation that does not produce the pseudo-shadow of false contouring, then, result data is write among the CFB 502B.Simultaneously, raw image data also is written among the UFB 502A, for reading back in the future.When display controller 503 refreshed display device 504, it just passed through path (2) sense data from CFB502B, and the data of least significant bit (LSB) are used as zero or predetermined value.
In the frame buffer read and write, this embodiment provides the transparent interface with host computer system.In this example, write at frame buffer, low-power working mode may consume more energy than normal mode of operation, and this is because " error diffusion encoder " 501 and CFB 502B consume extra energy.But because the reduced size of CFB 502B and interrelated logic, so display device refreshes the consumption less power.As a result, this embodiment typically is used for picture material change those application very slowly.Such exemplary application comprises e-book, picture/text editing machine, web browser etc.
Although above-mentioned hardware " error diffusion encoder " is embedded in the display controller, at some up-to-date handheld architectures, for example OMAP
TM2 Hes
Among the PXA27X, the part in the primary memory can be assigned with or be mapped as display frame buffer, and data write by Memory Controller and/or equivalent modules control.In this case, except such as some attachment features such as mode of operation switch in the display controller, " error diffusion encoder " also can be embedded in Memory Controller and/or the equivalent modules (for example, chipset).
In addition, according to an embodiment, can in error diffusion encoder, realize being used for the algorithm of planes of color (color plane).For the coloured image with 3 planes of color, algorithm is similar for each face.In one embodiment, according to the consecutive order that begins from top left pixel and from left to right be performed until the bottom of image each image pixel is carried out error diffusion algorithm.Perhaps, can adopt a plurality of scramblers to handle a plurality of planes of color.
Fig. 6 is the process flow diagram that illustrates according to the example of the error diffusion process of an embodiment.Process 600 can be carried out by processing logic, and this processing logic can comprise hardware (circuit, special logic etc.), software (for example operating in the software on general-purpose computing system or the custom-built machine) or the two combination.For example, process 600 can be carried out by the display subsystem shown in Fig. 3-5.
With reference to Fig. 6 A,, during low power state, for example during the power rating of low or reduction, pixel is carried out error diffusion operation to reduce the color depth of pixel at square frame 602 places.At square frame 604 places, at least a portion in the pixel that color depth is reduced is stored in first section of frame buffer, and not to second section conduct interviews (for example, write or read) of frame buffer.As a result, during low power state, can reduce or cut off second section the power that leads to frame buffer, to save power.Can also carry out other operation.Attention: equipment can be operated under normal power state and low power state independently, and can for example any time switch to another kind of power rating from a kind of power rating by hardware, software or both being combined in.Like this, a kind of power rating needn't depend on another kind of power rating, and vice versa.
Fig. 6 B is the process flow diagram that illustrates according to the example of the error diffusion process of an embodiment. process 650 can be carried out by processing logic, this processing logic can comprise hardware (circuit, special logic etc.), software (for example operating in the software on general-purpose computing system or the custom-built machine) or both combinations. for example, process 650 can be carried out by the display subsystem shown in Fig. 3-5, specifically, the part operation that comprises in the square frame 602 as Fig. 6 A.
With reference to Fig. 6 B,,, calculate output valve according to source pixel value for each pixel at square frame 652 places.In a particular embodiment, output pixel value can comprise 2 to 3 highest significant positions of source pixel value.At square frame 654 places, calculate the error between source pixel value and the output pixel value.At square frame 656 places, can for example comprise with this error diffusion to neighborhood pixels (for example, the pixel more than 4), upgrade the value of each neighborhood pixels by the error that increases weighting.The model of propagated error also is called as error-filter.Can also carry out other operation.
Typically, classical error diffusion is the algorithm that is similar to 2 dimension convolution, and its computation complexity height and storage buffer visit are concentrated.For example, in conventional algorithm, each pixel need be carried out 8 memory access at least, with error diffusion to 4 neighborhood pixels.Also need a storage buffer, so that the error information of interim storage diffusion with two image line sizes.These have brought " error diffusion encoder " expensive and high power consumption own, and this power technique for the portable set such as handheld device is unacceptable.
Therefore, according to an embodiment, the error diffusion algorithm that has designed a kind of simplification solves this problem, and it comprises " the contiguous error-filter of 2-" and " fault bit simplification ".In one embodiment, in " the contiguous error-filter of 2-", utilize the contiguous error-filter of 2-to replace the classical error wave filter, as follows:
| |
1/2 |
| 1/2 |
As implied above, when processed pixels X, according to an embodiment, the output color of X and the error between the primitive color only are diffused into right pixel and bottom pixel, and for isotropy diffusion and the convenience that realizes, weighting coefficient is set to 1/2.Because neighborhood pixels is reduced to 2, thus only need an impact damper of an image line size to store the error that is diffused into bottom pixel, and the error that is diffused into right pixel can be temporarily stored in the register.
Attention: although according to from top to bottom, from left to right carrying out said process, if image access order different (for example, from the lower-left to upper right), according to some embodiment of the present invention, neighborhood pixels can be right pixel and top pixels so.
Fig. 7 shows color depth from 8 pseudo-codes that are reduced to 3 instantiation procedure.For example, pseudo-code can be carried out by above-mentioned error diffusion encoder.With reference to Fig. 7, in 5-6 is capable, sues for peace to the original pixel value of X and from the left pixel of X and the weighted error of top pixels diffusion, and be stored among the register temp_sum temporarily.In the 7th row, calculate output pixel value by low 5 (being set to zero) of leaving out temp_sum, and output pixel value is write in the frame buffer (for example, FrameBuffer[x]).In eighth row, original pixels of calculate revising and the error between the output pixel, it is exactly low 5 of temp_sum.In the 9th and 10 row, with error divided by 2 and be stored in respectively in the impact damper (for example, buffer[x]) and register (register) of bottom pixel and right pixel.Apparently, if we only consider the extra computation of comparing with image copy, so the complexity of each pixel the chances are 2 times " addition ", 2 times " with ", 2 times " displacement " and one " condition evaluation " operation.Extra memory access also be reduced to 2 times (for example, the 5th and 10 the row in read and write buffer[x]).
In addition, according to an embodiment, except the contiguous error-filter of 2-, aspect the hardware realization, can also further simplify this error diffusion algorithm to reduce buffer memory access, most of energy of its consumption " error diffusion encoder ". for example, suppose that 8 color depths are reduced to 3, the figure place of error is 5, and be 4. can leave out or clip low 3 after multiply by 1/2 weighting coefficient, and only keep the error information that high 2 conducts are diffused into right pixel and bottom pixel. the result, reduced the memory access by the position counting.
And, because very little (the display of storage requirement for 640 * 480, each planes of color only needs 640 * 2 positions or 160 bytes), therefore storage buffer can save local the realization by SRAM (static RAM), and the energy that SRAM consumes lacks than DRAM (dynamic ram).
Fig. 8 shows the process flow diagram according to the process example of the error diffusion algorithm with the contiguous error-filter of 2-and fault bit simplification of an embodiment.Attention: the variable temp_sum here only is the symbol that is used to illustrate algorithm, rather than physical register or storer, temp_sum[7:5] the 5th, 6 and 7 data line of expression, and temp_sum[4:3] data line of the 3rd and 4 of expression.
Realize comparing with software, hardware solution has lower complexity and the energy efficiency of Geng Gao, this be because: (1) makes register and impact damper support the visit of 2 bit data on hardware is realized easily, this difficulty relatively in software solution; (2) on hardware was realized, it did not need temporary register and " displacement " operating unit.Instead be that these two functions can realize by suitably connecting data line.As a result, for each pixel, hardware " error diffusion encoder " only needs 2 times 2 bit memory visits and 2 times 2 bit register visits, and this consumes energy still less.
Top explanation is based on the image with a planes of color.For coloured image (for example 24 true color figures, 16 color graphicss etc.), in the configuration of different display systems, the RGB sub-pixel data can by by sub-pixel ground or by planes of color in the display system of packing into.For the former situation, can use 3 copies of " error diffusion encoder " to handle R, G and B sub-pixel independently.For the latter's situation, can use a copy of scrambler sequentially to handle R, G and B planes of color.Can there be other configuration.
In a particular embodiment, " error diffusion encoder " is implemented among the XC3S400-FG320 of Xilinx Spartan Series FPGA.According to an embodiment, Fig. 9 shows the hardware pseudo-code at a planes of color, wherein, 8 the pixel data of importing is encoded to 3 data.
For true color image with 3 planes of color.24 images are packed in the frame buffer by pixel ground, and wherein each pixel comprises B, G and the R color data with 8 respectively.In this case, utilize three copies and the corresponding internal buffer of " error diffusion encoder ".Use the block RAM (in unbound document, defining) among the FPGA to realize internal memory buffer, be stored in the sweep trace with error with diffusion.Block RAM is the dedicated block of real two-port RAM.This implementation is compared with the logical resource of FPGA and has been saved power.The frame rate that realizes was approximately for 30 frame/seconds, and picture size is 640 * 480 pixels.For encoder core logic, only use block RAM 3840,379 (slice) and 154 triggers, they have the door number with individual door equivalent 32723 (comprising block RAM).
Use real image data to carry out the hardware " error diffusion encoder " that FPGA emulation proposes with checking.Original image is 24 true color, and after " error diffusion encoder ", color depth is lowered to 9, and wherein RGB is 3: 3: 3.Simulation result shows that the image of hardware " error diffusion encoder " output is identical with image in the above-mentioned software experimentation.
Because " error diffusion encoder " be integrated in the lcd controller, so the core logic of additional encoder is used for power budget, thereby do not comprise the I/O part.The core logic of error diffusion encoder only consumes the power of about 2.4mW.By using above-mentioned one or more technology, can save about 5/8 frame buffer storage room, and can reduce their power or cut off their power.But consider that actual SDRAM frame buffer is 4,8 or 16 bit wides,, can save about 1/2 power for the frame buffer of 4 bit wides.For the frame buffer of 16M position, can save the power consumption of about 8M position.For example, if the SDRAMK4S161622E of 16M position that uses Samsung as frame buffer, its power consumption is 462mW under operator scheme so, and can save the power of 231mW.
For the performance of equivalence, FPGA will approximately consume the power of 20 times of ASIC.If error diffusion encoder directly is integrated in the lcd controller by ASIC, power consumption will be reduced to 0.12mW so.On an average, for the SDRAM K4S161622E of 16M position, one will consume 2.88 * 10-5mW.The power of hardware error diffusion can be converted into about 4155 power consumption among the SDRAM of 16M position.For above-mentioned error diffusion encoder, be cost with the power of 4k position, can save the frame buffer power of 8M position, this cost be approximately saving frame buffer power 0.05%.
The hardware " error diffusion encoder " that utilizes us to propose will be reduced to 9 from 24 as the numerical digit number, and we not only can save the power of frame buffer, but also can save the power of frame buffer data bus and LCD panel bus 5/8.Energy proportion based on the conventional system parts of handing embedded system, we can estimate the general power that the hardware error diffusion encoder saved ratio (since additional encoder power only be approximately saving frame buffer power 0.05%, so it is left in the basket in estimation).The LCD panel bus is the type with 4 additional synchronous/control signals.These data are shown in the following table.From following table, we can find out: the hardware " error diffusion encoder " that utilizes us to propose, can approximately save 10.57% of whole hand-held embedded system power.
We only calculate the power proportions that handheld device with independent frame buffer chip is saved.For the architecture of the part of primary memory with the display frame buffer of being mapped as, utilize hardware " error diffusion encoder solution ", still kept the saving power on display memory data bus and LCD panel bus.
Be about 24 color image encodings being become the implementation and the analysis of 9 coloured images above.Hardware design is similar with the coding of handling from 24 bit images to 8 bit images, from 16 bit images to 8 bit images, from 16 bit images to 9 bit images etc.Attention:, above-mentioned technology can be applied to planes of color for illustration purpose.But be not limited only to this.Should recognize that above-mentioned technology also can be applied to the gray level image plane, wherein the gray level image plane can be regarded as the planes of color of Special Category.
Figure 10 can use have at least one the above-mentioned functions parts block scheme of example computer system of embodiment of display subsystem of (for example above-mentioned error diffusion encoder).In one embodiment, computer system 1000 comprises communication agency, is used for interconnection and/or the bus 1011 and the integrated circuit components of exchange message, for example with bus or interconnect 1011 the coupling so that the Main Processor Unit of process information 1012.Main Processor Unit 1012 can comprise as the cooperative one or more processors in unit or handle core logic.
Alphanumeric Entry Device (keyboard) 1022 comprises alphanumeric key and other key, and this Alphanumeric Entry Device 1022 can also be coupled to bus 1011, to be used for that information and command selection are sent to Main Processor Unit 1012.Other user input device is a cursor control device 1023, for example mouse, tracking ball, touch pad, stylus or cursor direction key, it is coupled to bus 1011, being used for that directional information and command selection are sent to Main Processor Unit 1012, and be used to control the motion of cursor on display device 1021.Chipset can carry out interface with input-output device.
Another equipment that can be coupled to bus 1011 is hard copy devices 1024, and it can be used for going up print command, data or out of Memory at medium (for example medium of paper, film or similar type).In addition, SoundRec and playback apparatus, for example loudspeaker and/or microphone (not shown) can randomly be coupled to bus 1011, to be used for carrying out audio interface with computer system 1000.Another equipment that can be coupled to bus 1011 is wire/wireless communication equipment 1025.In addition, according to an embodiment, system 1000 comprises image acquisition equipment 1030, for example digital camera, video camera and/or scanner etc.Image acquisition equipment 1030 can obtain image stream, and system 1000 can use one or more above-mentioned technology to handle the image that is obtained.
According to an embodiment, the error diffusion encoder with one or more above-mentioned algorithms can realize in system 1000.For example, error diffusion encoder can be embodied as application software, it can be stored in the nonvolatile memory 1006, and is carried out handling the video data in one or more frame buffers by primary memory 1004, and in display device 1021 video data.Perhaps, error diffusion encoder can realize in firmware 1003 or device driver (for example, display driver).In addition, error diffusion encoder can realize in the hardware such as display controller that display controller can be realized in chipset 1036 and/or processor 1012.One or more frame buffers can be the storeies of the specific assigned of storer 1004.Perhaps, one or more frame buffers can be embodied as independent private memory in hardware and/or the firmware (for example combination of chipset 1036 and/or firmware 1003 or said structure).Can also comprise other parts.
Like this, this paper has described the method and apparatus of the error diffusion that is used for display frame buffer saving electricity.Some parts during the front specifically describes is to introduce with the algorithm and the symbolic representation of the computing that the data bit in the computer memory is carried out.These arthmetic statements and expression are that the data processing field technician is used for the content mode that the most effective others skilled in the art of conveying to adopted with their work.Herein and under normal conditions, algorithm is considered to cause the self-congruent sequence of operation of expected results.These operations are the operations that need carry out physical treatment to physical quantity.Usually (although not necessarily), the form that this tittle adopts and can be stored, transmits, makes up, compares and carry out the electric signal or the magnetic signal of other processing.Verified: as, these signals to be called bit, value, element, symbol, character, term, numerical value etc. normally easily mainly due to the reason of public use.
Yet, should keep firmly in mind, all these terms and similar terms all are associated with suitable physical quantity, unless and only be the convenience mark that is applied to this tittle. other special declaration, otherwise it is evident that by top argumentation, be to be appreciated that in whole instructions, the argumentation of employing such as " processing " or " calculating " or " computing " or " determining " or terms such as " demonstrations " refers to the action and the process of computer system or similar electronic computing device, and the data processing that its interior physics (electronics) of RS that will be expressed as computer system is measured also converts storer or register or other this information stores that is represented as computer system equally to, other data of physical quantity in transmission or the display device.
Embodiments of the invention can relate to the device that is used to carry out operation herein.Can be for required purpose this device of special configuration, perhaps this device can comprise multi-purpose computer, this multi-purpose computer is optionally activated or is reconfigured by the computer program of storing in this computing machine.This computer program can be stored in the computer-readable recording medium, the disc of (but being not limited to) any type for example comprises floppy disk, CD, CD-ROM and magneto-optic disk, ROM (read-only memory) (ROM), random-access memory (ram), erasable programmable ROM (EPROM), electrically erasable ROM (EEPROM), magnetic or light-card or is suitable for the store electrons instruction and each can be coupled to the medium of any type on the computer system bus.
Algorithm that this paper proposed and display are not associated with any specific computing machine or other device inherently.Various general-purpose systems can confirm that perhaps it is easily that the more specialized device of structure is carried out needed method operation with using based on the program of this paper instruction.By following description, will present the desired structure that is used for various these systems.In addition, embodiments of the invention are not described at any specific programming language.Should be appreciated that the instruction that to use various programming languages to realize embodiments of the invention as described herein.
Machine readable media can comprise any mechanism that is used for storage of the readable form of machine (for example computing machine) or transmission information.For example, machine readable media comprises ROM (read-only memory) (" ROM "); Random access memory (" RAM "); Magnetic disk storage medium; Optical storage media; Flash memory device; The transmitting signal of electricity, light, sound or other form (for example, carrier wave, infrared signal, digital signal etc.) etc.
In aforesaid instructions, embodiments of the invention have been described with reference to its concrete exemplary embodiment.Clearly, under the situation of the broader spirit of the embodiments of the invention that can in not breaking away from, propose and scope it is carried out various modifications by following claim.Correspondingly, that instructions and accompanying drawing should be thought exemplary meaning but not restrictive, sense.
Claims (17)
1. method that is used for error diffusion comprises:
Under normal power state, the pixel of the planes of color of view data directly is stored in first section and second section of frame buffer;
Under low power state, described pixel is carried out error diffusion operation to reduce the color depth of described pixel, described normal power state and low power state are separate and can switch mutually; And
During described low power state, at least a portion in the described pixel that color depth is reduced is stored in described first section of described frame buffer, and described second section of described frame buffer is not conducted interviews.
2. the method for claim 1 also comprises: during described low power state, reduce the described second section power that leads to described frame buffer.
3. the method for claim 1 also comprises:
During described normal power state, from described first and second sections of described frame buffer, get described pixel, to show; And
During described low power state, from described first section of described frame buffer, get the described pixel that color depth reduces and show, and described second section of described frame buffer do not conducted interviews.
4. method as claimed in claim 3, wherein, described first section is the highest effective equipment (MSD) of described frame buffer, and described second section is minimum effective equipment (LSD) of described frame buffer.
5. method as claimed in claim 4, wherein, during described low power state, the pixel of using color depth to reduce shows as the data relevant with described MSD, and use predetermined value to show as the data relevant with described LSD, and the described LSD of described frame buffer is not conducted interviews.
6. the step of the method for claim 1, wherein described pixel being carried out error diffusion operation comprises:
For each source pixel of each planes of color of described view data,, calculate output valve corresponding to the source pixel value of described source pixel according to predetermined consecutive order;
Calculate the error between described output valve and the described source pixel value; And
With two neighborhood pixels of described error diffusion to described source pixel.
7. method as claimed in claim 6, wherein, right pixel and bottom pixel that described two neighborhood pixels are described source pixel.
8. method as claimed in claim 6, wherein, the step of described error diffusion to two neighborhood pixels is comprised that at least a portion of utilizing described error regulates the pixel value of described two neighborhood pixels, wherein, the described part that is diffused into the described error of the described neighborhood pixels in the delegation is temporarily stored in the register, and the part that is diffused into the described error of the described neighborhood pixels in the next line is temporarily stored in the line buffer.
9. method as claimed in claim 6 also comprises: before each pixel being stored in described first section of described frame buffer, reduce each color of pixel position that color depth reduces, to be adapted in described first section of described frame buffer.
10. method as claimed in claim 9, wherein, the step that reduces each color of pixel position of color depth reduction comprises:
For each pixel of planes of color, will be added on the original value of pixel from the error of two neighborhood pixels diffusions with algorithmic approach; And
The highest significant position (MSB) of the predetermined quantity of described output valve is stored in described first section of described frame buffer.
11. the method for claim 1, wherein described error diffusion operation is carried out by scrambler, realizes at least one in the chipset of software, display controller and data handling system of described scrambler.
12. a device that is used for error diffusion comprises:
Directly memory module is used for during normal power state, the pixel of the planes of color of view data directly is stored in first section and second section of frame buffer;
Execution module is used for during low power state, and described pixel is carried out error diffusion operation to reduce the color depth of described pixel, and described normal power state and low power state are separate and can switch mutually; And
Memory module is used for during described low power state, and at least a portion in the described pixel that color depth is reduced is stored in described first section of described frame buffer, and described second section of described frame buffer is not conducted interviews.
13. device as claimed in claim 12, wherein, described execution module also comprises:
The output valve computing module is used for each source pixel for each planes of color of described view data, according to predetermined consecutive order, calculates the output valve corresponding to the source pixel value of described source pixel;
Error calculating module is used to calculate the error between described output valve and the described source pixel value; And
The error diffusion module is used for two neighborhood pixels of described error diffusion to described source pixel.
14. device as claimed in claim 13, also comprise and being used for before each pixel being stored in described first section of described frame buffer, reduce each color of pixel position that color depth reduces, to be adapted at the module in described first section of described frame buffer, this module comprises:
Be used for each pixel, will be added to the module on the original value of pixel with algorithmic approach from the error of two neighborhood pixels diffusion for planes of color; And
Be used for the highest significant position (MSB) of the predetermined quantity of described output valve is stored in the described first section module of described frame buffer.
15. a method that is used for error diffusion comprises:
During the low power state of frame buffer,,, calculate output valve corresponding to the source pixel value of described source pixel according to predetermined consecutive order for each source pixel of each planes of color of view data with first section and second section;
Calculate the error between described output valve and the described source pixel value;
With two neighborhood pixels of described error diffusion to described source pixel; And
During the described low power state of described frame buffer, store the described output valve of described source pixel and the described error that is diffused into two neighborhood pixels into described frame buffer described first section, and described second section of described frame buffer do not conducted interviews.
16. method as claimed in claim 15 also comprises: before being stored in described first section of described frame buffer, reduce the colour bits of each output valve and described two neighborhood pixels, to be adapted in described first section of described frame buffer.
17. method as claimed in claim 16, wherein, the step that reduces colour bits comprises:
For each pixel of planes of color, will be added on the original value of pixel from the error of two neighborhood pixels diffusions with algorithmic approach; And
The highest significant position (MSB) of the predetermined quantity of described output valve is stored in described first section of described frame buffer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2005800517021A CN101273360B (en) | 2005-09-29 | 2005-09-29 | Error diffusion for display frame buffer saving electricity |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2005800517021A CN101273360B (en) | 2005-09-29 | 2005-09-29 | Error diffusion for display frame buffer saving electricity |
| PCT/CN2005/001620 WO2007036070A1 (en) | 2005-09-29 | 2005-09-29 | Error diffusion for display frame buffer power saving |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101273360A CN101273360A (en) | 2008-09-24 |
| CN101273360B true CN101273360B (en) | 2010-05-12 |
Family
ID=40006402
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2005800517021A Expired - Fee Related CN101273360B (en) | 2005-09-29 | 2005-09-29 | Error diffusion for display frame buffer saving electricity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101273360B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103593040B (en) * | 2012-08-17 | 2016-06-01 | 宏碁股份有限公司 | Power-supply management system and method for managing power supply |
| CN110688081B (en) * | 2014-04-28 | 2024-05-17 | 华为终端有限公司 | Method for displaying data on screen and display control device |
| CN115831061B (en) * | 2022-08-23 | 2023-07-28 | 广州文石信息科技有限公司 | Ink screen display method, device, terminal equipment and storage medium |
-
2005
- 2005-09-29 CN CN2005800517021A patent/CN101273360B/en not_active Expired - Fee Related
Non-Patent Citations (3)
| Title |
|---|
| JP特開2002-251173A 2002.09.06 |
| JP特開平8-146905A 1996.06.07 |
| JP特開平9-101771A 1997.04.15 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101273360A (en) | 2008-09-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20090102850A1 (en) | Error Diffusion for Display Frame Buffer Power Saving | |
| US7262776B1 (en) | Incremental updating of animated displays using copy-on-write semantics | |
| US11398195B2 (en) | Backlight brightness processing method and system, backlight brightness adjustment method, storage medium | |
| CN205645213U (en) | Electronic equipment | |
| CN107273301A (en) | Data handling system | |
| US7460136B2 (en) | Efficient scaling of image data in graphics display systems | |
| US20110074800A1 (en) | Method and apparatus for controlling display operations | |
| US10410398B2 (en) | Systems and methods for reducing memory bandwidth using low quality tiles | |
| US11030728B2 (en) | Tone mapping techniques for increased dynamic range | |
| US10636365B2 (en) | Device and method for image correction | |
| US20200234131A1 (en) | Electronic apparatus and control method thereof | |
| CN101273360B (en) | Error diffusion for display frame buffer saving electricity | |
| CN103794182B (en) | Display control unit and data handling system | |
| US8587600B1 (en) | System and method for cache-based compressed display data storage | |
| TW201817232A (en) | Image processing method and related apparatus | |
| CN111311478B (en) | Pre-reading method and device for GPU rendering core data and computer storage medium | |
| CN111080505B (en) | Method and device for improving graphic element assembly efficiency and computer storage medium | |
| CN107479897A (en) | A kind of vector graphics interface Zoom method used on video monitoring equipment | |
| TWI240221B (en) | Memory access device, semiconductor device, memory access method and recording medium | |
| US7327873B2 (en) | Fast software rotation of video for portrait mode displays | |
| US7382376B2 (en) | System and method for effectively utilizing a memory device in a compressed domain | |
| CN102906695A (en) | Client browser image rendering method and system | |
| CN104504640A (en) | Method and device for adding watermark into picture and computation equipment | |
| US10565677B2 (en) | System and method for adaptive z-buffer compression in low power GPUS and improved memory operations with performance tracking | |
| CN115831042B (en) | Image display method and system, display driving device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20100512 Termination date: 20140929 |
|
| EXPY | Termination of patent right or utility model |