Patent Application
20060170712
The present invention relates to color display devices and, more particularly, to arrangements of subpixel elements in such color display devices.
Typical flat panel displays employ pixel patterns with red, green, and blue stripes. A portion of such a display device 2 is shown in
Applying the stripe pixel pattern shown in
It is also known in the art that when relatively large pixels are displayed on a small display or when graphics image regions are likely to be shown that demand a uniform appearance, alternating rows of light emitting subpixel elements may be offset horizontally to reduce the visibility of banding in a display device. Rows of alternating red, green and blue subpixels may be offset, e.g., to form an RGB “delta pattern”. Unlike the stripe pattern, this pattern reduces the visibility of banding and improves the uniform appearance in areas of constant color by shifting the alignment of the red, green, and blue subpixel elements in alternating rows. Unfortunately, this pattern also creates a visible jagged pattern in vertical lines containing primarily green light emitting subpixel elements as the human eye is very sensitive to offsets in light emitting subpixel elements that are high in luminance. Additionally, horizontally offsetting subsequent rows of pixels often results in a subpixel arrangement that does not form a column and typically forces the use of non-linear power and data lines, increasing the length of these lines and therefore the area between subpixels on the display device as well as the resistance of the lines. It is, however, known that in certain display structures, such as top-emitting, active-matrix OLED structures that the electronics may be placed on a different vertical layer, allowing the electronics to reside under the subpixel and reducing the impetus for the pixels to be laid out in a rectilinear grid.
It has been known for many years that the human eye is most sensitive to greenish-yellow light and less sensitive to red and blue light. More importantly, the spatial resolution of the human visual system is driven primarily by the luminance rather than the chrominance of a signal. Since green light provides the preponderance of luminance information in a display device employing red, green and blue subpixels when viewed in typical viewing environments, the spatial resolution of the visual system under normal daylight viewing conditions is highest for green light, lower for red light, and even lower for blue light when viewing images generated by a typical color balanced image capture and display system. This fact has been used in a variety of ways to optimize the frequency response of imaging systems.
It is further known in the art to employ different numbers of red, green, and blue subpixels within a repeating pattern of a display device in order to improve the perceived image quality of the display device for a given number of subpixels. In published papers, Rogowitz in 1988 (The psychophysics of spatial sampling in the Society of Photographic and Instrumentation Engineers, Vol. 901, Image Processing, Analysis Measurement and Quality, pp. 130-138) and later Silverstein and colleagues in 1990 (Effects of spatial sampling and luminance quantization on the image quality of color matrix displays in the Journal of the Optical Society of America, Vol. 7, No. 10, pp. 1955-1968) described the use of a four element pattern having two green, one blue and one red subpixel per pixel as shown in
A particularly noteworthy advantage of this pixel pattern is that because the red and blue subpixels are offset from each other on a diagonal axis, each of the three sets of color emitters has an equal sampling lattice in the horizontal and vertical axes. Therefore, the largest horizontal or vertical separation 32 between any two neighboring subpixels is only one subpixel plus the inactive area between the subpixels. This is important since if this separation is large, banding or dithering-like artifacts may be visible in any flat field within an image.
One disadvantage of the pattern shown in
A second disadvantage of this pixel pattern is that the subpixels are relatively large in both the horizontal and vertical dimensions as compared to other potential pixel patterns, such as the stripe pattern shown in
A third disadvantage is that this pixel pattern may require additional power and/or data lines, if the data and power lines provide only data and power to a single colored subpixel as is the case in traditional displays.
Other pixel patterns with fewer red and blue subpixels have been discussed by Credelle (U.S. patent application 2004/0080479 filed on Jan. 16, 2003 and entitled “Sub-pixel arrangements for striped displays and methods and systems for sub-pixel rendering same”) who discusses an arrangement of stripe pixel patterns having two subpixels of one color (typically green) and one subpixel each of a second and third color (typically red and blue). One such pixel pattern is shown in
It is also worth noting that Credelle (U.S. patent application 2004/0080479) also discusses methods for resampling the input data to the particular subpixel arrangement that is provided. In the approach provided by Credelle, a 3×3 matrix, or filter kernel, is convolved with the input data. A disadvantage of this technique is that it requires 3 rows of data to be buffered in peripheral or external memory or controlling circuitry such that data for the preceding and following rows are available to perform this convolution. For small portable devices, this requirement may add complexity and cost while also increasing the power demands for the final ASIC.
There is a need, therefore, for an improved pixel pattern for color display devices that takes advantage of the eye's relative inability to sense high spatial frequency information in both the red and blue channels in comparison to higher luminance channels, such as the green channel of a display, and to reduce the overall number of subpixels required to obtain a desired display quality wherein regions of uniform color appear uniform and are not degraded by visible banding or dithering-like patterns due to the scarcity of the sampling pattern. Such a pattern should improve the uniformity of a pattern and yet avoid the visibility of jagged vertical or horizontal lines. Further, it would be desirable for such pixel pattern to allow the areas of the subpixels to be adjusted independently of one another while ideally providing a rectangular grid for the routing of select, power, and data lines. Finally, it would be further desirable for such a pixel pattern to allow a less complex approach to resampling that does not require multiple rows of data to be buffered.
In accordance with one embodiment, the invention is directed towards a color display device, comprising: an array of subpixels of three different colors, including subpixels of a relatively high luminance first color and subpixels of relatively lower luminance second and third colors, wherein the subpixels are arranged into rows or columns to form a repeating pattern of alternating lower luminance and high luminance color subpixels in each row or column, with the sequential order of the two lower luminance color subpixels being alternated within each row or column, and wherein the alignment of subpixels of the same colors in adjacent rows or columns is such that the high luminance color subpixels are aligned more closely to perpendicular than are each of the lower luminance color subpixels relative to the direction of the rows or columns in which the subpixels are arranged in a repeating pattern.
In accordance with various embodiments of the invention, the use of pixel patterns with fewer relatively lower luminance color subpixels than relatively high luminance color subpixels (e.g., fewer red and blue subpixels than green subpixels) is enabled while providing a uniform appearance in regions of solid primary or secondary colors. The various embodiments further improve the uniformity of a pattern and yet decrease the visibility of jagged vertical or horizontal lines. Further, in preferred embodiments the pixel pattern will provide the flexibility of resizing the relative areas of the different color subpixels to provide display color balance and/or extend the lifetime of the display device. Further, the pixel patterns in various embodiments allow red and blue subpixels to share data and power lines, simplifying panel layout and potentially improving pixel fill factor. Additionally while data resampling methods known in the art may be applied to avoid certain sampling artifacts that can result with subsampling, these pixel patterns allow the reduction of these artifacts when resampling is not applied in order to simplify the processing of the input signal. Finally, many of the pixel patterns allow the use of a rectilinear grid for routing of data select, data and power lines.
In accordance with various embodiments described herein, the invention is directed towards a color display device, comprising: an array of subpixels of three different colors, including a first relatively high luminance color and two relatively lower luminance colors, wherein the subpixels are arranged into rows or columns to form a repeating pattern of alternating lower luminance and high luminance color subpixels in each row or column, with the sequential order of the two lower luminance color subpixels being alternated within each row or column. Thus, due to the two lower luminance color subpixels being alternated in sequence with the single high luminance color subpixel, there are more high luminance color subpixels than lower luminance subpixels of a single color, and the lower luminance colors are subsampled relative to the high luminance color. Further in accordance with the invention, the alignment of subpixels of the same colors in adjacent rows or columns is such that the high luminance color subpixels are aligned more closely to perpendicular than are each of the lower luminance color subpixels relative to the direction of the rows or columns in which the subpixels are arranged in a repeating pattern. Pixel patterns meeting such requirements are designed to reduce the maximum separation between the lower luminance subpixels while maintaining the high luminance color subpixels in relative perpendicular alignment, by providing pixels having more than one subpixel arrangement in neighboring pixels.
Experiments conducted by the inventors have shown that when displaying patterns with red and blue subsampling, as known in the prior art, on a display device at resolutions typical of manufacturing today, banding or dithering artifacts are readily apparent when primary and/or secondary colors are displayed. Further experiments have demonstrated that if the sampling lattice of the pattern is designed such that neighboring subpixels of any given color are separated by less than one minute of arc, the visibility of the banding or dithering is significantly reduced and, in fact, may be essentially eliminated if the separation is significantly less than a visual angle of one minute of arc. This result is surprising since it is to be expected that a 100 percent contrast white target on a display could be resolved at this resolution, however, the visual system would typically be assumed to be less responsive to targets of lower brightness and therefore lower contrast than the white point of the display. Further, the experiments conducted by the authors have demonstrated that the maximum of the horizontal and vertical distance between neighboring pairs of subsampled, lower luminance subpixels can be reduced through the use of different subpixel arrangements in neighboring pixels. These different arrangements can be achieved by alternating the location of lower luminance subpixels on alternating rows or columns of pixels and/or by offsetting the subpixels in successive rows or columns of pixels. This distance may be further reduced through the use of non-rectilinear subpixel shapes, such as triangles, that produce overlaps in the horizontal and/or vertical dimension. By requiring that the alignment of subpixels of the same colors in adjacent rows or columns is such that the high luminance color subpixels are aligned more closely to perpendicular than are each of the lower luminance color subpixels relative to the direction of the rows or columns in which the subpixels are arranged in a repeating pattern, the maximum of the horizontal and vertical distance between neighboring pairs of subsampled, lower luminance subpixels can be minimized while maintaining the high luminance color pixels in relative perpendicular alignment, thus decreases the visibility of jagged vertical or horizontal lines.
Within this document the term “subpixel” represents the smallest individually addressable element in a display device. The term “pixel” is applied to represent an arrangement of neighboring subpixels containing two, higher-luminance subpixels and two lower-luminance subpixels wherein the two lower-luminance subpixels have different colors. The term “data location” is applied to represent a theoretical location in which a set of input code values would be rendered on a traditional, fully-sampled, three-color display system.
In one embodiment of the present invention, alternating horizontal rows of pixels in the display comprise a first pixel type wherein the subpixels are positioned in a sequence of relatively lower luminance second color, relatively high luminance first color, relatively lower luminance third color, and relatively high luminance first color subpixels, where the sequence of subpixels in the alternating rows of pixels is repeated across the width of the display, while interleaving horizontal rows of pixels between the alternating rows in the display comprise a second pixel type wherein the subpixels are positioned in a sequence of relatively lower luminance third color, relatively high luminance first color, relatively lower luminance second color, and relatively high luminance first color subpixels, where the sequence of subpixels in the interleaving rows of pixels also is repeated across the width of the display, and where the sequences of subpixels in alternating and interleaving rows repeat across the height of the display. In such arrangement, the subpixel types on one row may be arranged to form a repeating pattern of lower luminance and higher luminance subpixel types in a stripe arrangement while the order of the lower luminance subpixels are altered on the successive row as shown in
As shown in
An arrangement such as this may be particularly desirable because the higher luminance elements are aligned vertically, perpendicular to the horizontal rows. This fact is important since vertical lines within text characters and other high-contrast, vertically-oriented edges will appear “jagged” (i.e., have a sawtooth pattern appearance) if these high contrast subpixels are not vertically aligned. The fact that the location of the red and blue subpixels are interchanged with each successive row of subpixels, and that the alignment of the red and blue subpixels of the same colors in adjacent rows or columns is accordingly further from perpendicular than that of the high luminance green subpixels, decreases the maximum of the vertical and horizontal distance between neighboring subpixels of these color channels and improves the overall uniformity when flat fields of red and blue colors are displayed while maintaining the desired relative perpendicular alignment of the green pixels.
It should additionally be noted that a single data 76 and power line 78 may be used to provide connections to both a red 58 and a blue 68 subpixel within this pixel arrangement. As will be discussed later, processing to provide the correct voltage and current to each subpixel is performed on the input data signal to enable the data line 76 to be shared by both red 58 and blue 68 subpixels. The fact that a common data line shares red and blue subpixels while a second data line is used to drive the only high luminance subpixel also allows data to be communicated to the display using two input channels. This has the effect of reducing the number of output channels that must be supported by a display processor (e.g., asic) from three to two. This not only reduces the complexity of the processor but also reduces the power required to drive the analog output channels of the processor.
Within this pixel pattern, it is important to note that the relative areas of the subpixels can be adjusted. The active area of the green subpixel can be adjusted by simply changing the horizontal width of the subpixel. The relative areas of the red and blue subpixels with respect to the active area of the green subpixel may be adjusted using the same method. Note that the relative areas of the red and blue subpixels, however, can not be easily adjusted relative to one another by adjusting their horizontal widths without making one wider than another, resulting in a larger inactive area than necessary or a non-rectilinear grid. With care, however, the relative heights of these subpixels can be adjusted.
While such a pixel pattern may be useful for any display technology, it may be particularly useful in OLED display applications since it is known to be desirable to allocate different areas to the different colored light emitting elements in order to optimize their lifetime. For this reason,
The subpixels are arranged in a matrix of rows and columns. That is, for example, subpixel 58 and subpixel 60 are arranged in a first row. Select line 76a and capacitor line 78a are shared by the subpixels in this first row. Subpixel 68 and subpixel 70 are arranged in a second row. Select line 76b and capacitor line 78b are shared by the subpixels in this second row. Subpixel 58 and subpixel 68 are arranged in a first column. Data line 80 and power line 82 are shared by the subpixels in this first column. While only a limited number of rows and columns are shown, this design can be expanded to provide for a plurality of rows and columns. Alternate arrangements can also be practiced. For example, two adjacent columns may share the power line. Alternately, the power line may be run in the same row direction instead of the column direction and be shared by the subpixels of the row. Also, other more complex subpixel circuits having more transistors in various arrangements are known in the art and can also be applied to the present invention by one skilled in the art.
The drive circuitry operates in a manner well known in the art. Each row of subpixels is selected by applying a voltage signal to the select line, such as select line 76a, which turns on a select transistor, such as select transistor 84, for each subpixel. The brightness level for each subpixel is controlled by a voltage signal, which has been set on the data lines such as data line 80. The storage capacitor, such as storage capacitor 86, for each subpixel is then charged to the voltage level of the data line associated with that subpixel and maintains the data voltage until the row is selected again during the next image frame. The storage capacitor 86 is connected to the gate of the power transistor 88 so that the voltage level held on storage capacitor 86 regulates the current flow through the power transistor 88 to the organic light emitting diode 90 and thereby controls the subpixel's brightness. Each row is then un-selected by applying a voltage signal to the select line, such as 76a, which turns off the select transistors. The data line signal values are then set to the levels desired for the next row and the select line of the next row, for example 76b, is turned on. This is repeated for every row of subpixels.
A layout diagram for the portions of the drive circuitry used to drive subpixels 58, 60, 62, 64, 68, 70, 72, and 74 is shown is shown in
Connections between layers are formed by etching holes (or vias) in the insulating layers such as via 98 connecting data line 80 to the first semiconductor region 92. Similarly, via 100 connects the power transistor gate 108 to first semiconductor region 92, via 104 connects the second semiconductor region 94 to power line 82, and the via 102 connects the second semiconductor region 94 the first electrode 96.
First electrode 96 serves to provide electrical contact to the organic electroluminescent media of the organic light emitting diodes. Over the perimeter edges of the first electrode 96, an intersubpixel dielectric layer (not shown) may be formed to cover the edges of said electrodes and reduce shorting defects as described below. The emitting area of subpixel 58 is defined by the areas of the first electrode 96 which is in electrical contact with the organic electroluminescent media. This emitting area is the area of the first electrode 96 reduced by any area covered by dielectric material.
Each of the differently colored subpixels can have different efficiencies and lifetimes. Therefore, the emitting area for each differently colored subpixel will be optimized differently. Several approaches to optimizing the emitting area are known in the art, examples of which can be found in U.S. Pat. Nos. 6,366,025 and 6,747,618.
The emitting areas of the subpixels can be adjusted without bending of any of the signal lines by adjusting the size of the emitting area in the column direction, or height (H), or adjusting the size of the emitting area in the row direction, or width (W). By disposing select line 76a and select line 76b on the outside of their associated subpixels, different heights of the subpixel emitting areas can be achieved for subpixels in the same row, as shown. That is pixels 58, 60, 62, 64, 68, 70, 72, and 74 are disposed between select line 76a and select line 76b, allowing the select lines to be formed in a straight, unbending, fashion. For pixels in the same column, these pixels may generally have the same width. These heights and widths are thereby balanced so that each different colored pixel has the desired emitting area. It is not necessary that the subpixel emitting areas be perfectly rectangular, as irregularities in the emitting areas, as shown, may be provided to conform to the areas of the circuit components, such as the transistors.
One or more of the subpixels may further include a color filter element (not shown) to alter the spectrum of the emitted light of the subpixel. The color filter elements may be disposed between the organic electroluminescent media and the viewer.
A cross-sectional view illustrating the vertical arrangement of the various layers of the device of
Above the substrate 112, a first semiconductor layer is provided, from which semiconductor region 92 is formed. Above semiconductor region 92, first dielectric layer 114 is formed and patterned by methods such as photolithography and etching. This dielectric layer is preferably silicon dioxide, silicon nitride, or a combination thereof. It may also be formed from several sub-layers of dielectric material. Above first dielectric layer 114, a first conductor layer is provided, from which power transistor gate 108 is formed and patterned by methods such as photolithography and etching. This conductor layer can be, for example, a metal such as Cr, as is known in the art. Above power transistor gate 108, a second dielectric layer 116 is formed. This dielectric layer can be, for example, silicon dioxide, silicon nitride, or a combination thereof. Above second dielectric layer 116, a second conductor layer is provided, from which power line 82 and data line 80 are formed and patterned by methods such as photolithography and etching. This conductor layer can be, for example, a metal such as an Al alloy as is known in the art. Power line 82 makes electrical contact with semiconductor region 92 through a via opened in the dielectric layers. Over the second conductor layer, a third dielectric layer 118 is formed.
Above the third dielectric layer, a first electrode 96 is formed. First electrode 96 is preferably highly transparent for the case of a bottom-emitting configuration and may be constructed of a material such as ITO. Above first electrode 96, an inter-subpixel dielectric 120 layer, such as is described in U.S. Pat. No. 6,246,179, is preferably used to cover the edges of the first electrodes in order to prevent shorts or strong electric fields in this area. While use of the inter-subpixel dielectric 120 layer is preferred, it is not required for successful implementation of the present invention. As described previously, the area of the first electrode 96 not covered by inter-subpixel dielectric 120 constitutes the emitting area and is represented on the this cross-section view as dimension Y.
Each of the subpixels further includes an organic EL media 110. There are numerous configurations of the organic EL media 110 layers wherein the present invention can be successfully practiced. For the organic EL media, a broadband or white light source, which emits light at the wavelengths used by all the subpixels, may be used to avoid the need for patterning the organic EL media between subpixels. In this case, color filters (not shown) may be provided for some of the subpixels in the path of the light to produce the desired light colors from the white or broadband emission for a multi-color display. Some examples of organic EL media layers that emit broadband or white light are described, for example, in U.S. Pat. No. 6696177B1. However, the present invention can also be made to work where each subpixel has one or more of the organic EL media layers separately patterned for each subpixel to emit differing colors for specific subpixels. The organic EL media 110 is constructed of several layers such as; a hole injecting layer 122, a hole transporting layer 124 that is disposed over the hole injecting layer 122, a light-emitting layer 126 disposed over the hole transporting layer 124, and an electron transporting layer 128 disposed over the light-emitting layer 126. Alternate constructions of the organic EL media 110 having fewer or more layers can also be used to successfully practice the present invention. These organic EL media layers are typically comprised of organic materials, either small molecule or polymer materials, as is known in the art. These organic EL media layers can be deposited by several methods known in the art such as, for example, thermal evaporation in a vacuum chamber, laser transfer from a donor substrate, or deposition from a solvent by use of an ink jet print apparatus.
Above the organic EL media 110, a second electrode 130 is formed. For a bottom emitting device, this electrode is preferably highly reflective and may be composed of a metal such as aluminum or silver or magnesium silver alloy. The second electrode may also comprise an electron injecting layer (not shown) composed of a material such as lithium to aid in the injection of electrons. When stimulated by an electrical current between first electrode 96 and second electrode 130, the organic EL media 110 produces light emission 132.
Most OLED devices are sensitive to moisture or oxygen, or both, so they are commonly sealed in an inert atmosphere such as nitrogen or argon, along with a desiccant such as alumina, bauxite, calcium sulfate, clays, silica gel, zeolites, alkaline metal oxides, alkaline earth metal oxides, sulfates, or metal halides and perchlorates. Methods for encapsulation and desiccation include, but are not limited to, those described in U.S. Pat. No. 6,226,890. In addition, barrier layers such as SiOx, Teflon, and alternating inorganic/polymeric layers are known in the art for encapsulation.
OLED devices of this invention can employ various well-known optical effects in order to enhance its properties if desired. This includes but is not limited to optimizing layer thicknesses to yield maximum light transmission, providing dielectric mirror structures, replacing reflective electrodes with light-absorbing electrodes, providing anti-glare or anti-reflection coatings over the display, providing a polarizing media over the display, or providing colored, neutral density, or color conversion filters over the display.
While the embodiments described herein refers to a specific configuration of active matrix drive circuitry and subpixel design, several variations of conventional circuits that are known in the art can also be applied to the present invention by those skilled in the art. For example, one variation in U.S. Pat. No. 5,550,066 connects the capacitors directly to the power line instead of a separate capacitor line. A variation in U.S. Pat. No. 6,476,419 uses two capacitors disposed directly over one and another, wherein the first capacitor is fabricated between the semiconductor layer and the gate conductor layer that forms gate conductor, and the second capacitor is fabricated between the gate conductor layer and the second conductor layer that forms power lines and data lines.
While the drive circuitry described herein requires a select transistor and a power transistor, several variations of these transistor designs are known in the art. For example, single- and multi-gate versions of transistors are known and have been applied to select transistors in prior art. A single-gate transistor includes a gate, a source and a drain. An example of the use of a single-gate type of transistor for the select transistor is shown in U.S. Pat. No. 6,429,599. A multi-gate transistor includes at least two gates electrically connected together and therefore a source, a drain, and at least one intermediate source-drain between the gates. An example of the use of a multi-gate type of transistor for the select transistor is shown in U.S. Pat. No. 6,476,419. This type of transistor can be represented in a circuit schematic by a single transistor or by two or more transistors in series in which the gates are connected and the source of one transistor is connected directly to the drain of the next transistor. While the performance of these designs can differ, both types of transistors serve the same function in the circuit and either type can be applied to the present invention by those skilled in the art. The example of the preferred embodiment of the present invention is shown with a multi-gate type select transistor 84 which is represented by a single transistor symbol in the circuit schematic diagrams
Also known in the art is the use multiple parallel transistors, which are typically applied to power transistor 106. Multiple parallel transistors are described in U.S. Pat. No. 6,501,448. Multiple parallel transistors consist of two or more transistors in which their sources connected together, their drains connected together, and their gates connected together. The multiple transistors are separated within the subpixels so as to provide multiple parallel paths for current flow. The use of multiple parallel transistors has the advantage of providing robustness against variability and defects in the semiconductor layer manufacturing process. While the power transistors described in the various embodiments of the present invention are shown as single transistors, multiple parallel transistors can be used by those skilled in the art and are understood to be within the spirit of the invention.
Turning again to other pixel patterns that are designed for application within any known display technology, another method for reducing the maximum of the vertical and horizontal spacing between subpixels within the red and blue channels is to shift rows or columns of pixels in order to reduce the gap between neighboring red or blue subpixels, as shown in
Within the particular layout of
Each of the embodiments shown has used rectilinear shaped subpixels. However, this is not required, and, in fact, the effective distance between neighboring subpixels of a single color may be reduced through the use of non-rectilinear shaped subpixels. This is especially useful in displays with large fill factors. One such pixel pattern is shown within a small portion of the display device 180 in
By arranging these subpixels as shown in
It should further be noted that by applying this arrangement of triangularly-shaped subpixels that the red and blue subpixels can once again share data and power lines, thereby providing the advantages of the pattern shown in
While it is not necessary that pairs of subpixels form a rectangle, the fact that each pair of triangles in
It should also be noted, that the rectangular arrangement that is formed from pairs of triangles can be maintained even when different sized, red, green and blue subpixels are required.
Having a display device with these pixel patterns, the input three-color data stream that is input into a display device must be converted to a signal capable of driving such a display device. Such a rendering method is dependent upon numerous parameters.
Before processing the data, one will determine 254 the ratio of the output luminance for each of the color channels to the luminance required from each channel to form the white point of the display device. Depending upon the technology of the display device, this ratio may be affected by a number of factors. For instance, in an LCD, the relative area of each liquid crystal element, the color filter and/or the spectrum of the backlight can affect the output luminance of each color channel when spatially integrated across several pixels. In emissive devices, such as OLEDs, the drive current, the size of the emitter, the spectrum of the emitter, and/or the spectrum of any color filter that is applied may affect this same ratio. However, in a preferred embodiment, this ratio will be unity. Calculating or measuring the average luminance of each color channel at its maximum drive value may determine this value. This value is used as a numerator and the denominator is determined by calculating the luminance required from each color channel to form the desired display white point.
Following this step, the sampling area will be determined 256 for each subpixel on the display. Since the image resolution is sampled to the same number of data locations as the display device has higher luminance (e.g., green) subpixels, each green subpixel will typically represent the same area as the input image data. However, each lower luminance (e.g., red and blue) subpixel will typically represent a larger area in the original input since there are fewer of these subpixels.
While the previous discussion showed a two-dimensional sampling area, subsequent image processing steps may be simplified if the sampling area is thought of in a single dimension.
This sampling area is used as an input to calculate 258 an initial resampling matrix. This resampling matrix is typically defined such that the denominator of entries in this matrix are a function of the area of each input pixel that lies within the sampling area. For example, the two matrices as shown below can be formed by simply taking the inverse of the area of each input data location that lies within the sampling area for the two-dimensional resampling matrix as well as the one-dimensional resampling matrix.
In this example, each entry in these matrices represents the proportion of each data location that lies within the sampling area. While other approaches may be used to create these matrices, these values will be a function of the proportion of each data location that lies within the sampling area.
These matrices are then normalized 260 such that the sums of the matrix elements are equal to the ratio of the output luminance for each of the color channels to the luminance required from each channel to form the white point of the display device as defined in step 254. Assuming that this ratio is 1.0 for the target display, the final matrices are formed with the values shown as:
These matrices are then convolved 262 with the input signal, which, ideally, will be expressed in a metric that is linearly related to the desired luminance output of the display device. Those skilled in the art will recognize that this process is a prefiltering process that is well known in the art as digital prefiltering, a process that is applied when downsampling a digital image to a lower resolution digital image to avoid aliasing as discussed by W. K. Pratt in Digital Image Processing, John Wiley and Sons, New York, 1978 on pages 104-111. Those skilled in the art will recognize the matrices as filter kernels. For the pixel patterns shown in
This same approach may be applied for each pattern of subpixels disclosed herein. However, the preferred matrices will be different for each of the remaining subpixel patterns. Other simpler approaches may be applied. For example, a single interpolation step may be performed in which the green channel is interpolated to the number of green subpixels while the red and blue channels are interpolated to the number of red and blue subpixels. The resulting values may then be used to directly drive the display device. It may also be noted that the resampling implemented as steps 256 through 262 in
The invention has been described in detail with particular reference to certain preferred embodiments thereof, but it will be understood that variations and modifications can be effected within the spirit and scope of the invention.
