Patent Application
20090303248
1. Field of the Invention
The present invention relates generally to processes for driving image display devices, and more particularly to an improved system and method for dithering video data. Even more particularly, the present invention relates to a system and method for dithering video data to be displayed on a display including an array of individual pixel cells.
2. Description of the Background Art
In recent years the demand for flat panel image/video displays has drastically increased, mainly because the overall volume and weight is significantly less than that of traditional CRT (cathode ray tube) displays of equivalent screen area. In addition, flat panel display devices are used in other applications unsuitable for conventional CRTs, for example in high resolution video projection systems. Examples of flat panel displays used in video projection systems include, but are not limited to, liquid crystal on silicon (LCOS) and deformable mirror devices (DMDs).
Today digital displays (e.g., LCDs) are common. When driving digital LCDs, the pixel is driven in one of two states: an “on” state or an “off” state. During the “on” state a saturation voltage potential is applied across the liquid crystal layer which results in the maximum light output (i.e., a light pixel or “on”). Conversely, the “off” state is obtained by applying a threshold voltage potential across the liquid crystal layer which results in the minimum light output (i.e., a dark pixel or “off”). Thus, at any given instant in time, a pixel is either on or off.
Because a digital LCD pixel only has two states, on or off, PWM (pulse width modulation) techniques have been employed so that a single pixel can display what appears to be other intermediate intensities. PWM involves modulating a pixel back and forth between two different states at such a rate the human eye integrates the two intensities to perceive a single intensity. For example, to display what appears to be a single intensity of 10% maximum brightness the “off” state is asserted 90% of the time frame while the “on” state is asserted the other 10% of the time frame. Similarly, to display what appears to be a single intensity of 75% maximum brightness the “off” state is asserted 25% of the time frame while the “on” state is asserted the other 75% of the time frame.
In a similar fashion, a method commonly referred to as dithering is used to display intensities unobtainable by single frame PWM. As an example, a particular type of dithering called temporal dithering is used to display intensity levels that are between the intensity levels that are attainable by PWM. Temporal dithering works similarly to PWM, except that temporal dithering modulates the values attained by PWM. In other words, PWM intensities are attained by modulating 0% and 100% intensities between time slices of a single frame while temporal dithering intensities are attained by modulating these PWM intensities over several frames of data. For example, to display the intermediate pixel value 127.25 on a single pixel, the value 127 is obtained from PWM and displayed three out of every four frames while the value 128 (also obtained from PWM) is displayed once every four frames. As a result, a greater number of intensity levels than defined by the PWM scheme can be achieved.
One problem associated with temporal dithering is that the number of displayable intermediate intensities between the PWM intensities are limited to the number of frames over which the data is dithered. For example, if a cycle includes a series of 10 frames, the only attainable intermediate intensities are tenths. Likewise, if the cycle includes a series of 4 frames, the only attainable intermediate intensities are fourths. For example, if the cycle includes 4 frames, the displayable intermediate intensities between N and N+1 are 1.25N, 1.5N, and 1.75N, N being an arbitrary intensity value defined by the PWM scheme, and N+1 being the next intensity value defined by the PWM scheme. Note that cycle refers to the sequence of frames needed to display a particular intensities.
Another dithering method, commonly known as spatial dithering, involves combining the simultaneous output of a plurality of pixels to achieve intermediate intensity levels. For example, a group of four pixels will appear to have a uniform value of 127.75 if three pixels are illuminated with a value of 128 and the other pixel is illuminated with a value of 127. Similarly, a group of four pixels will appear to have a uniform intensity value of 127.5 if two pixels are illuminated with a value of 127 and the other two pixels are illuminated with a value of 128.
One problem commonly associated with Spatial Dithering is that image resolution is sacrificed for the increase in intensity resolution. This is because it takes multiple pixels to make a single intensity value, rather than just modulating a single pixel to render a single intensity as described for pure temporal dithering. As an example, if an LCD includes groups consisting of four adjacent pixels that render what appears to be a single intensity, the resolution of the entire display will be four times less than it would be if each individual pixel were responsible for a single intensity.
One problem with prior art circuit 100 is that the number of displayable pixel values are limited by the size of the data word received by the dithering logic. For example, if display driver circuit 100 is driven by 8-bit data words, then only 256 different values can be defined, before modulation techniques are applied. So, the smallest increments between intensity values is limited to the value of data word's LSB (least significant bit). For example, if a dithering logic process adds a bit value to an 8-bit data word, the original value is increased by a value of 1/256 which is approximately 0.3906% of the maximum value.
Another problem is that the electro-optical response curve of the some displays (e.g., LCDs) is not linear. As a result, even if display data can be dithered to precisely achieve an intermediate root-mean-square (RMS) voltage, that RMS voltage may not produce the desired intensity output.
Other known methods for displaying intermediate intensity values involve estimation techniques. However, estimating values leads to noticeable image problems such as the appearance of “steps” or “lines” in contoured images. The appearance of such “steps” is a result of a an estimated intensity value being more different than it's true value than that of an adjacent intensity value being displayed on adjacent pixels.
What is needed, therefore, is a display driving circuit and method capable of more accurately displaying intensity values on a pixel or group of pixels. What is also needed is a display driving circuit and method that eliminates visual artifacts from displayed images.
The present invention overcomes the problems associated with the prior art by providing a system and method for dithering video data. Video data is converted to a second data type that defines a greater number of intensity levels than the original data and includes dither bits that identify one of a plurality of dithering schemes to be applied to that particular data. The converted data is temporally dithered, and the phase of the temporally dithered data stream is shifted based on the relative location of the pixels to which the data is to be written. The invention facilitates greater accuracy in the reproduction of intensity levels and substantially reduces visual artifacts in displayed data including, but not limited to, flicker and contouring.
A disclosed example display driver circuit includes an input for receiving video data, a data converter coupled to receive the video data and operative to convert the video data into pixel data to be written to pixels of a display, and a ditherer operative to receive the pixel data and to dither the pixel data to generate dithered pixel data. The video data is data of a first type, and the pixel data is data of a second type, different from the first type. In the disclosed example, the first type of data includes a binary data word, and the second type of data includes a compound data word. The compound data word includes a first set of binary weighted bits, including at least one bit, and a second set of arbitrarily weighted bits, also including at least one bit. Optionally, at least some of the arbitrarily weighted bits are equally weighted.
The video data is capable of defining a first number of values, and the pixel data is capable of defining a second number of values, the second number of values being greater than the first number of values. In a disclosed example, the video data includes data words having a first number of bits, and the converted pixel data includes data words having a second number of bits, the second number of bits being greater than the first number of bits. More particularly, in a disclosed example, the video data is binary-weighted video data, and the pixel data includes data words having a group of equally weighted bits. The data words of the pixel data further include a group of binary weighted bits.
The ditherer performs a predetermined dithering function based on at least a portion of the pixel data. For example, the data converter (e.g., a look-up-table) inserts dither bits into the converted pixel data. The dither bits identify a particular one of a plurality of different dither schemes that is to be performed on that particular data word.
A method for driving a display device is also disclosed. An example method includes receiving video data of a first type, converting the first type of video data to data of a second type, dithering the data of said second type to form dithered pixel data, and outputting the dithered pixel data. The step of receiving the video data includes receiving a binary data word indicative of an optical intensity level.
The first type of data is defined by a first data word, and the second type of data is defined by a second data word. The first data word has a least significant bit, and the second data word has a least significant bit. The least significant bit of said second data word is less significant than the least significant bit of the first data word. This facilitates dithering at a finer scale.
Optionally, the step of converting the video data to the data of a second type includes converting the video data to the data of the second type via a lookup table. The second type of data includes more bits and defines more values than the first type of data. In addition, the step of converting the first type of data to data of a second type includes adding a set of dither bits to each data word of the second type, and the step of dithering the second type of data includes dithering the data word of the second type according to one of a plurality of predetermined dithering logic functions depending on the value of the dither bits.
Optionally, the step of converting the video data to the second data type includes converting the video data to compound data words. The compound data words each include a first set of binary bits and a second set of arbitrarily weighted bits, the first set of binary bits and the second set of arbitrarily weighted bits each including at least one bit. 22. In the example method, the arbitrarily weighted bits include a set of equally weighted bits.
A disclosed example method can also be described as including the steps of providing a display with an array of individual pixels, defining a group of said pixels of said display, temporally dithering data to be written to each pixel of said group to generate a series of values to be asserted on each pixel of said group, and changing the order of at least one of said series of values depending on the location of a pixel of said group upon which said reordered series of values is to be asserted. In other words, the series of values is written to each pixel of the group out of phase with the other pixels of the group, thereby reducing flicker which can sometimes be caused by prior art temporal dithering methods.
The present invention is described with reference to the following drawings, wherein like reference numbers denote substantially similar elements:
a is time sequenced top view of a pixel group showing 1.25 ×N dithering;
b is a timing diagram corresponding to the pixel group of
a is a time sequenced top view of pixel group showing 1.5 ×N dithering;
b is a timing diagram corresponding to the pixel group of
a is a time sequenced top view of a pixel group showing 1.75 ×N dithering;
b is a timing diagram corresponding to
The present invention overcomes the problems associated with the prior art by providing a system and method for driving an image display that more accurately displays intensity values and reduces visual artifacts including, but not limited to, contouring. In the following description, numerous specific details are set forth (e.g., number of pixels in a pixel group, specific data schemes, etc.,) in order to provide a thorough understanding of the invention. Those skilled in the art will recognize, however, that the invention may be practiced apart from these specific details. In other instances, details of well known electronics manufacturing practices (e.g., specific device programming, circuitry layout, timing signals, etc.) and components have been omitted, so as not to unnecessarily obscure the present invention.
Greater accuracy with respect to displayed intensities is achievable, because the incoming video data is converted to a higher resolution data scheme. The particular intensity values are then mapped to particular intensity values of the display data scheme that provide the closest correlation between the actual intensity displayed and the value of the original video data. The primary reason for mapping the video data to a higher resolution data scheme is not to increase the color bit depth of display 204. Rather, increasing the intensity resolution of the display data 220 facilitates a closer matching between the values of the original video data and the actual intensities displayed.
Dithering of the display data 220 (as opposed to dithering of the original video data 218) provides even closer matching between the values of the video data words and the intensities displayed. Because each video data word is converted into a display data word of greater resolution, the LSB (least significant bit) of the display data has a smaller value than the LSB of the video data word. The smaller valued LSBs allow finer adjustments via dithering.
For example, an 8-bit binary data word can define 256 intensity levels, each level corresponding to 1/256 (0.39%) of the full intensity. Temporal dithering data over four frames would facilitate an adjustment of ¼ of 0.39%, or about 0.98%. On the other hand, adding just two additional binary bits to the data word results in a ten-bit data word that can define 1,024 intensity levels, each corresponding to 1/1,024, or about 0.098%, of the full intensity. Temporal dithering of the 10-bit data over four frames would then facilitate an adjustment of ¼ of 0.098%, or about 0.024%.
Although the foregoing example uses data words with binary weighted data bits, it should be understood that the technique can be used with data words including other bit-weighting schemes. For example, data words can include binary-weighted bits, equally-weighted bits, arbitrarily-weighted bits, thermometer bits (sequentially set bits), or any combination thereof. As long as the converted display data defines more intensity values than the original video data, the dithering process can provide finer adjustment of the intensity levels.
In addition to the data conversion that facilitates finer adjustment of intensity values by a dithering process, display artifacts such as contouring can be significantly reduced by a novel dithering technique. The novel dithering technique combines aspects of temporal and spatial dithering, and achieves good results without sacrificing spatial resolution. The new technique, therefore, provides an important advantage over the dithering techniques of the prior art. The new dithering technique will be explained with reference to
a shows data values asserted on the pixels of group 302 during four successive frames, during a dithering process intended to display a intensity of 1.25N. In this particular example, N represents an arbitrary value defined by display data 220 of
Note that the values N and N+1 are asserted on each pixel to properly achieve 1.25N dithering, but not at the same time. During the first frame, N+1 is applied to pixel 00 while N is applied to adjacent pixels 01, 11, and 10. During the second frame N+1 is applied to pixel 01 while N is applied to adjacent pixels 11, 10, and 00. During the third frame, N+1 is applied to pixel 11 while N is applied to adjacent pixels 10, 01, and 01. During the fourth frame, N+1 is applied to pixel 10 while N is applied to adjacent pixels 00, 01, and 11. As a result, each pixel receives the temporally dithered data, so there is no loss of spatial resolution.
This new type of dithering can be considered spatially phase-shifted, temporal dithering. As shown, each pixel receives the same temporally dithered data. However, the sequence in which the data values are asserted on each pixel is offset with respect to the other pixels. The offset is determined by the relative location of the individual pixel.
Diagram 400B includes four rows, each showing corresponding to a different pixel address 00, 01, 11, or 10. During each frame, either value N or N+1 is asserted on each pixel. During the first frame, N+1 is applied to pixel 00, and N is applied to pixels 01, 11, and 10. During the second frame, N+1 is applied to pixel 01, and N is applied to pixels 11, 10, and 00. During the third frame , N+1 is applied to pixel 11, and N is applied to pixels 10, 00, and 01. Finally, during the fourth frame, N+1 is applied to pixel 10, and N is applied to neighboring pixels 00, 01, and 11.
It should be apparent from the view of
Pixel address/counter 702 receives timing signals 723 (e.g., Vsynch, Hsynch, pclk, etc.) and uses the timing signals 724 to keep track of the pixel address for which each incoming 8-bit data word is destined and provides a group sub-address 728 (00, 01, 10, or 11) to distinguish that pixel from the other three pixels in a four pixel group. The Vsynch signal indicates the start of a new frame of data, the Hsynch signal indicates the start of a new row of data, and the pclk signal indicates each new 8-bit data word. The group sub-address 728 corresponds to the 2-bit pixel addresses shown in
Frame count XY remapper 708 receives pre-frame count 726 and group sub-address 728, and then remaps the pre-frame count to a frame count 730, depending on the value of the group sub-address. Thus, remapper 708 facilitates the phase shifting of the temporal dithering depending on the location of a particular pixel within a four-pixel group, as illustrated in
Dithering logic 710, responsive to the values of both frame count 730 and dither bits 718, outputs a bit to be added to compound data word 736. In particular, dither bits 718 can have one of four possible values, each of which causes dithering logic 710 to implement a respective one of four logic operations. If dither bits 718 have the value 00, dithering logic 710 will output a single bit with a value of 0. If dither bits 718 have the value 01, dithering logic 710 will perform a logical “AND” operation on the bits of frame count 730, then output the single bit result as output bit 732. If dither bits 718 have the value 10, output bit 732 will be set equal to the inverse (i.e., logical “NOT”) of the LSB of the frame count 730. If dither bits 718 have the value of 11, dithering logic 710 will perform a logical “AND” operation on the bits of frame count 730 and output the inverse of the result. Thus, if frame count 730 has the value 00, 01, or 10, output bit 732 will be set to 1. If the frame count 730 has the value 11, output bit 732 will be set to 0. The results of the logical operations performed by dithering logic 710 are summarized in the following table, where the frame count values are listed in the top row and the D-bit values are listed in the left most column. A value of N indicates that the value of output bit 732 is 0, and a value of N+1 indicates that the value of output bit 732 is 1.
Output bit 732 is added to compound data word via adder 712 and SHL 714. In particular, adder 712 adds single bit value of 1 or 0 to the six bit binary word defined by B-bits 720. If the summing of B-bits 720 and output bit 732 generates a carry bit 734, then carry bit 734 is added to the thermometer bits via shift left register (SHL) 714. The resulting binary and thermometer bits are then output to subsequent processing circuitry such as a data planarizer.
The description of particular embodiments of the present invention is now complete. Many of the described features may be substituted, altered or omitted without departing from the scope of the invention. For example, pixel groups of different sizes may be substituted for 2×2 pixel group 302. As another example, data types different than those described can be used with the present invention. As yet another example, the present invention can be implemented with a programmable logic device including a computer-readable storage medium having code embodied therein for causing an electronic device to perform the methods disclosed herein. These and other deviations from the particular embodiments shown will be apparent to those skilled in the art, particularly in view of the foregoing disclosure.
