JP2002543473A - Method, apparatus, and data structure for maintaining a consistent baseline position in a system for rendering text - Google Patents

Abstract

(57) Summary The resolution of text rendered on a display device with sub-pixel elements, such as an RGB LCD, specifically a display device with horizontal striping (812 '), is (i) vertical (or (1610) Overscaling (or oversampling) (1610) the character outline information in the Y) direction and (ii) filtering (1610) information (1650) displaced from the overscaled (or oversampled) character outline information. To improve. The metric associated with the character outline information is adjusted accordingly (1620). The vertical (or Y) position of the baseline of an adjacent character can be limited (1640) by forcing the first pixel above the baseline to consist of all N number of scan conversion source samples. Where N corresponds to the overscaling (or oversampling) factor. The group of scan conversion source samples can be converted to a compressed pixel index value (1660). Color values can be selectively filtered when the intensity difference between adjacent sub-pixel elements would otherwise be difficult to see. Finally, the gamma of the pixel values can be corrected so that the gamma of the display device is taken into account so that the intensity values of the sub-pixel elements fall within the range of intensities for which gamma correction is more useful.

Description

DETAILED DESCRIPTION OF THE INVENTION

(§1. Background of the Invention) (§1.1 Field of the Invention) The present invention relates to a liquid crystal display (or an LCD) having horizontal striping.
CD) For producing more legible text on video displays such as flat panel video monitors, including video monitors.

(§1.2 Related Art) The present invention can be used in the context of flat panel video monitors, such as LCD video monitors. In particular, the present invention can be used as part of a process for generating more legible text on an LCD video monitor with horizontal striping.

[0003] Color display devices have become a major option for most computer users. The color is rendered on the display monitor by manipulating the display monitor to emit light (eg, a combination of red, green, and blue light), and the light is rendered by one or more human eyes. Perceived as color.

[0004] Generally, color video monitors, specifically LCD video monitors, are well known to those skilled in the art, but they are introduced below for the convenience of the reader. A cathode ray tube (or CRT) is first introduced in §1.2.1 below. Next, an LCD video monitor is introduced in §1.2.2 below.

[0005] CRT Video Monitors [0005] CRT display devices include a screen having a phosphor coating that can be applied as a sequence of dots. Different phosphor coatings are usually associated with producing different colors, for example, red, green, and blue. Therefore,
A repeated sequence of phosphor dots is defined on the screen of the video monitor.
One or more electron guns generate an electron beam that is swept across the screen, generally from left to right, top to bottom. When the phosphor dots are illuminated by the electron beam, they cause the phosphor dots to render and emit their associated colors, such as, for example, red, green, and blue.

The term “pixel” is generally used to refer to a spot in a group of spots, for example, thousands of spots in a rectangular grid. This spot is selectively activated to form an image on the display device. Most color CRTs cannot uniquely select a single triple of red, green, and blue phosphor dots. Therefore, the smallest possible pixel size will depend on the focus, alignment, and bandwidth of the electron gun used to excite the phosphor dots. Light emitted from one or more triads of red, green, and blue phosphor dots in various arrangements known for CRT displays tends to mix, and is monochromatic at a distance. Seen as a light source.

In a color display, an additive primary color (additive primary c) is used.
color (e.g., red, green, and blue) can be varied in intensity to achieve almost any desired color pixel appearance. By not adding any color, i.e. emitting no light, a black pixel is created. By adding 100 percent of all three colors, a white pixel is created.

Now that a color CRT video monitor has been introduced, a color LCD video monitor will be introduced in § 1.2.2 below.

(§1.2.2 LCD Video Monitor) In portable computing devices (commonly also referred to as computing or untethered computing appliances), rather than CRT displays, liquid crystal displays (LCDs) or other Flat panel display devices are often used. This is because flat panel displays tend to be smaller and lighter than CRT displays. In addition, flat panel displays are generally well suited for battery powered applications because they consume less power than comparable sized CRT displays.

A color LCD display is an example of a display device that explicitly addresses pixel elements to represent each pixel of the displayed image.
Typically, each pixel element of a color LCD display has three non-square elements ("
Sub-pixel element "or" sub-pixel component "). More specifically, each pixel element can include adjacent red, green, and blue (RGB) sub-pixel elements. Thus, a set of RGB sub-pixel elements together define a single pixel element. Some LCD displays have non-square pixels and / or pixels defined by four or more sub-pixel elements.

[0011] Known LCD displays generally include a series of RGB sub-pixel elements that are typically arranged to form a stripe along the display. The RGB stripes typically extend in one direction throughout the length of the display. The resulting RGB stripe is sometimes referred to as "RGB striping." Many LDC monitors used for computer applications are longer in width than in height and require
They tend to have vertical stripes. On the other hand, many LDC monitors used in unconstrained or handheld computing devices tend to be longer in height than width and have RGB horizontal stripes. The present invention
It can be used when rendering text on a monitor having horizontal stripes, such as an LCD RGB monitor.

FIG. 1 shows a known LCD screen 100 arranged in a plurality of rows (R1 to R8) and a plurality of columns (C1 to C6). That is, a pixel is defined at each row-column intersection. Each pixel includes a red subpixel element indicated by hatching, a green subpixel element indicated by cross hatching, and a blue subpixel element indicated by no hatching.

FIG. 2 shows the upper portion of the known display 100 in more detail. How each pixel element, eg, the (R1, C6) pixel element, comprises three distinct subpixel elements or subpixel elements: a red subpixel element 210, a green subpixel element 220, and a blue subpixel element 230. Note if you have. The height of each known sub-pixel 210, 220, 230 is one-third or about one-third of the height of the pixel, and its width is equal to or approximately equal to the width of the pixel. Thus, when combined, three sub-pixels 210, 220, 230 that are 1/3 height and full width define a single pixel element.

Returning to FIG. 1, one known arrangement of RGB pixel sub-components 210, 220, 230 defines a horizontal color stripe on display 100. Thus, the arrangement of 1/3 height color sub-pixel elements 210, 220, 230 in the known manner shown in FIGS. 1 and 2 represents an arrangement sometimes referred to as "horizontal striping".

In known systems, RGB sub-pixel elements are generally addressed and used as a group to generate a single color pixel corresponding to a single sample of the image to be represented. More specifically, in known systems, intensity values for all sub-pixel elements of a pixel element are generated from a single sample of the image to be represented. For example, referring to FIG. 3, the image section 300 is segmented by the grid 310 into 12 squares. Each square of the grid 310 defined by the segmented image section 300 represents an area of the image section 300 to be represented by a single pixel element. In FIG. 3, a hatched circle 320 is used to represent a single image sample from which intensity values associated with the associated red, green, and blue sub-pixel elements 330, 332, and 334 of the pixel are generated.

Having introduced the general structure and operation of known LCD displays, known techniques for rendering text on such LCD displays, as well as disadvantages of such known techniques, are described below in §1. Introduced in 2.2.1.

§ 1.2.2.1 Rendering Text on LCD Displays In addition to pure image or video information, LCD displays are often used to render text information. For example, a personal information manager can be used to render contact information such as, for example, a person's address, phone number, fax number, and email address.

Expressions of text information using a font set are introduced in § 1.2.2.1.1 below. Then, the so-called pixel precision
Rendering text information using, and the perceived shortcomings of doing so, are introduced in § 1.2.2.1.2 below.

(§1.2.2.1.1.1 Font Set) “Font” is a set of the same typeface (Times Rome)
an, Courier New, etc.), the same style (italic, etc.), and the same weight (bold, etc., strictly speaking, the same size). Character,
For example, Microsoft Corporation of Redmond, Washington
Word ™ word processors “Parties MT”, “Webd
ings "and" Windings "symbol groups. “Typeface” is oblique (ie, inclined (
degree of slant) and stoke weight (
That is, a design of a specific name of a set of print characters (for example, Helvetica Bold Oblique) specifying the thickness of the line). Strictly speaking, a typeface is not the same as a font, and a font is a specific size of a specific typeface (1
2 point Helvetica Bold Oblique). However, some fonts are "scalable", so the terms "font" and "typeface"
Are sometimes used interchangeably. A "typeface family" is a group of related typefaces. For example, the Helvetica family is Helvetica, He
lvetica Bold, Helvetica Oblique, and He
lvetica Bold Oblique.

Many modern computer systems use font outline technology, for example, scalable fonts, to facilitate rendering and display of text. Microsoft Copora, Redmond, Washington
Tion's TrueType ™ is an example of such a technology. Such systems include, for example, "Times New Roman", "Onyx",
Various font sets such as “Courier New” can be provided. This font set typically includes a high resolution outline display, such as a series of outlines, for each character that can be displayed using the provided font set. This contour can be, for example, a straight line or a curve. A curve is defined, for example, by a series of points that draw a quadratic Bezier spline. The points that define the curve are generally numbered in sequential order. The ordering of this point can be important. For example, when following a curve in the ascending direction of point numbers, this character outline can be "filled" to the right of the curve. Thus, a high-resolution character outline display can be defined by a set of points and mathematical expressions.

The position of this point can be described in, for example, a “font unit”. "Font unit" can be defined as the smallest measurable unit in the "em" square. This "em" square is a virtual square used to resize and align glyphs (glyphs can be considered characters). FIG.
The "em" square 910 around the character outline 920 of the alphabet Q
Is shown. Historically, "em" was approximately equal to the width of the capital letter M. More historically, glyphs could not extend beyond the em square. More generally, however, the dimensions of the "em" square are the dimensions of the total body height 940 of the font plus some extra spacing. This extra spacing prevents lines of text from colliding when a typeset without leading is used. More generally, some of the glyphs can extend outside the em square. Lines and curves (
Or, the coordinates of the point defining the contour) can be positioned relative to the baseline 930 (Y coordinate = 0). The portion of the character outline 920 above the baseline 930 is referred to as the "ascent" 942 of the glyph. The portion of the character outline 920 below the baseline 930 is referred to as the "decent" 944 of the glyph. Note that there are languages, such as Japanese, where the characters lie above the baseline and there are no portions of the characters extending below the baseline.

The stored outline character representation is based on the maximum horizontal and vertical boundaries of the character (
"White space" or "side bearing"
) Is not usually represented. Therefore, the stored character outline portion of the character font is often black body (or BB).
). A font generator is a program that converts a character outline into a bitmap of the style and size required by the application. Font generators (also called "rasterizers") generally operate by scaling a character outline to the required size, and often can expand or compress the generated characters.

In addition to black body character outline information, character fonts typically include black body size information, black body positioning information, and full character width information. Black body size information may be represented by the dimensions of the bounding box used to define the vertical and horizontal borders of the black body.

Next, some terms used to define characters are defined with reference to FIG. FIG. 4 shows a letter outline 400 of the alphabets A and I. Box 408 is a bounding box that defines the size of black body 407 of character (A). The full width of character (A), including the white space to be associated with character (A), is indicated by advance width (or AW) value 402. The advance width generally starts at the point to the left of the bounding box 408. This point 404 is called the left side bearing point (or LSBP). Left side bearing point 404 defines a horizontal start point for positioning character (A) relative to the current display position. Left end of bounding box 408 and bearing point 4
The horizontal distance 410 between the left and right sides is called the left side bearing (or LSB). The left side bearing 410 is the bounding box 40 of the current character (A).
8 shows the amount of white space between the left edge of 8 and the right bearing point of the preceding character (not shown). Bounding box 408 at the end of the advance width 402
Is called the right side bearing point (or RSBP). The right side bearing point 406 defines the end of the current character (A) and the point where the left side bearing point 404 'of the next character (I) should be located. The horizontal distance 412 between the edge of the bounding box 408 and the right side bearing point 406 is
Called the right side bearing (or RSB). The right side bearing 412 indicates the amount of white space between the right end of the bounding box 408 of the current character (A) and the left side bearing point 404 'of the next character (I). Note that the left and right side bearings can have zero (0) or negative values. For characters used in Japanese and other Far Eastern languages,
It should also be noted that similar metrics for the advance width, left side bearing, and right side bearing can be used, ie, advance height (AH), top side bearing (TSB), and bottom side bearing (BSB).

As discussed above, a scalable font file typically contains black body size information, black body positioning information,
And character width information. Black body size information may include horizontal and vertical size information expressed in the form of bounding box 408 dimensions. The black body positioning information can be represented as a left side bearing value 410. The entire character width information can be represented as an advance width 402.

§ 1.2.2.1.2 Rendering Text to Pixel Accuracy Recall that the font generator converts a black body character outline to a bitmap. This conversion can take into account the point size of the font to be rendered and the resolution (eg, dots / inch, pixels / inch, etc.) of the device that will ultimately render the text (eg, video display, printer, etc.). Most computer systems force the start and end points of the character to be rendered (eg, left and right side bearing points, respectively) on pixel boundaries. In addition, such computer systems force the black body width and left side bearing to be converted to integer multiples of the pixel size. In known embodiments, these constraints are based on (i) scaling the size and positioning information contained in the character font according to the point size and device resolution described above, and (i)
i) It is then performed by rounding the size and positioning values to an integer multiple of the pixel size used in that particular display device. By using the pixel size unit as the smallest (or "atomic") distance unit,
What is called "pixel accuracy" is generated. This is because that value is exactly the size of one pixel.

By rounding the character font size and positioning values to pixel precision, changes, or errors, are introduced into the displayed image. Each of these errors can be up to 1/2 pixel in size (assuming that values less than 1/2 pixel are truncated and values greater than 1/2 pixel are rounded up). Thus, because the AW of the character is (potentially) rounded, the overall width of the character may be less accurate than desired. In addition, the positioning of the character's black body within the total horizontal space allocated to the character may be sub-optimal because the left side bearing is rounded (possibly). At small point sizes, the changes introduced by rounding with pixel precision can be significant.

(§1.2.3 Unmet Needs) In view of the error introduced when character values are rounded to pixel precision, as introduced in §1.2.2.1.2 above. What is needed is a method and apparatus for improving text resolution. The improvement in readability and perceived quality of text should continue to function not only on displays such as RBG LCDs with vertical striping, but also on displays such as RBG LCDs with horizontal striping.

The present invention improves the resolution of text rendered on a display device having sub-pixel elements, such as an RGB LCD, specifically a display device having horizontal striping. In the present invention, (i) the character outline information is overscaled (or oversampled) in the vertical (or Y) direction, and (ii) the character outline information is displaced (directly) from the overscaled (or oversampled) character outline information. This can be done by filtering (e.g., averaging) scan conversion source samples, either adjacent or overlapping, which overlap.

The present invention may also appropriately adjust the metrics associated with character outline information (such as left side bearing, advance width, vertical character size, ascent, decent, etc.).

The present invention enforces that the first pixel above the baseline be composed of all N number of scan conversion source samples, thereby allowing the vertical (or Y) position of the baseline of adjacent characters Can be constrained. Where N corresponds to the overscaling (or oversampling) factor. This prevents "jumping" or "bouncing" of the baseline.

The present invention can also convert groups of scan conversion source samples into compressed pixel index values.

The present invention may also selectively filter color values when differences in the intensity of adjacent sub-pixel elements would otherwise obstruct viewing.

Finally, the present invention corrects the gamma of the pixel values so that the gamma of the display device is taken into account and the intensity values of the sub-pixel elements fall within the range of intensities for which gamma correction is more useful. (Or an effect similar to gamma correction can be achieved).

The present invention is a novel method, apparatus, and method that has horizontal striping and is used to increase the resolution of fonts to be rendered on a display, such as an RGB LCD display. Regarding data structure. The following description is presented to enable one of ordinary skill in the art to make and use the invention, and is provided in the context of a particular application and its requirements. Various modifications to the disclosed embodiments will be apparent to those skilled in the art. The general principles described below can be applied to other embodiments and application examples. Accordingly, the present invention is not intended to be limited to the embodiments shown.

The method, apparatus and data structure of the present invention will be described in the context of a system for improving the readability of text on a flat panel screen. Section 4.1 introduces functions that can be performed by the present invention. Section 4.2 describes exemplary processes, methods, and data structures for implementing the present invention. Throughout section 4.2, the operations and intermediate data structures of the exemplary embodiments of the present invention are provided. Finally, a summary of the present invention is presented in §4.3.

(§4.1 Features of the Present Invention) The present invention increases the resolution of text rendered on a display device with sub-pixel elements, such as an RGB LCD, specifically a display device with horizontal striping. Work to make it work. In the present invention, (i) the character outline information is overscaled (or oversampled) in the vertical (or Y) direction, and (ii) the character outline information is displaced (directly) from the overscaled (or oversampled) character outline information. This can be done by filtering (e.g., averaging) scan conversion source samples, either adjacent or overlapping, which overlap.

The present invention also serves to properly adjust the metrics associated with character outline information (such as left side bearing, advance width, vertical character size, ascent, decent, etc.).

The present invention enforces that the first pixel above the baseline is composed of all N number of scan conversion source samples, thereby allowing the vertical (or Y) position of the baseline of adjacent characters Can be constrained. Where N corresponds to the overscaling (or oversampling) factor.

The present invention can also convert groups of scan conversion source samples into compressed pixel index values. Although the scan conversion process is described below as operating on a “scan conversion source sample”, the scan conversion process, and the overscaling (or oversampling) and hinting processes, are analogous rather than discrete samples. It can be an analog operation that operates on information.

The present invention also serves to selectively filter color values when differences in the intensity of adjacent sub-pixel elements would otherwise interfere with viewing.

Finally, the present invention considers the gamma of the display device and corrects the gamma of the pixel values such that the intensity values of the sub-pixel elements fall within a range of intensities for which gamma correction is more useful. Can be.

(§4.2 Exemplary Devices, Processes, Methods, and Data Structures for Implementing Various Aspects of the Invention) Exemplary Methods that Can Be Used to Implement Various Aspects of the Invention , Apparatus, and data structures are described in the context of a system for improving the readability of text on a flat panel screen with horizontal striping (also referred to as a text enhancement system). An overview of this system is provided in §4.2.2 below. However, before that, exemplary devices and text enhancement systems that can be used to implement at least some aspects of the present invention are described below in §4.2.
1 will be described.

§4.2.1 Exemplary Device FIG. 5A and the following discussion provide a brief, general description of an exemplary device capable of implementing at least some aspects of the present invention. . The various methods of the invention will be described under the general context of computer-executable instructions, such as program modules and / or routines, executed by a computing device, such as a personal computer. Other aspects of the invention are described by way of physical hardware, such as a display device component and a display screen.

Of course, the method of the present invention can be performed by other devices than the devices described here. Program modules are routines, programs, objects, components, that perform tasks or implement particular abstract data types.
A data structure (eg, a lookup table, etc.) can be included. further,
At least some aspects of the present invention may be applied to handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network computers, minicomputers, set-top boxes, mainframe computers, and One of ordinary skill in the art will appreciate that other configurations may be implemented, including displays and the like used in the field of application. At least some aspects of the present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in local and / or remote memory storage devices.

FIG. 5A is a block diagram of an exemplary apparatus 500 that can be used to implement at least some aspects of the present invention. The personal computer 520 includes a processing device 521, a system memory 522, a system memory 5
A system bus 523 may be included that couples various system components, including 22, to the processing device 521. System bus 523 may be any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. System memory 522 may include a read-only memory (ROM) 524 and / or a random access memory (RAM) 525. A basic input / output system 526 (BIOS), containing the basic routines that help to transfer information between elements within the personal computer 520, such as during startup, can be stored in the ROM 524. The personal computer 520 includes a hard disk drive 527 for reading and writing on a hard disk (not shown), and a magnetic disk drive 528 for reading and writing on a (for example, removable) magnetic disk 529.
And an optical disk drive 530 for reading from and writing to (magnetic) optical disks 531 such as compact disks or other (magnetic) optical media.
Hard disk drive 527, magnetic disk drive 528, and (magnetic)
The optical disk drive 530 can be coupled to the system bus 523 by a hard disk drive interface 532, a magnetic disk drive interface 533, and a (magnetic) optical drive interface 534, respectively. The drive and its associated storage medium allow the personal computer 5
Non-volatile storage of machine readable instructions, data structures, program modules, and other data for the device 20 is provided. The exemplary environment described herein uses a hard disk, a removable magnetic disk 529 and a removable optical disk 531, but with a magnetic cassette, flash memory card, digital video disk, Bernoulli cartridge, random access memory (RAM), readout. Those skilled in the art will appreciate that other types of storage media, such as dedicated memory (ROM), can be used in place of or in addition to the storage devices introduced above.

The hard disk 523, the magnetic disk 529, the (magnetic) optical disk 531, R
On the OM 524 or the RAM 525, for example, an operating system 535,
One or more application programs 536, other program modules 537, display driver 732 (described in § 4.2.2.2 below).
And / or some program modules, such as program data 538.

The RAM 525 can also be used to store data used in the rendered image to be displayed, as described below. A user can enter commands and information into personal computer 520 via input devices such as keyboard 540 and pointing device 542, for example. Other input devices (not shown), such as a microphone, joystick, game pad, satellite dish, scanner, etc., may also be included. These and other input devices are often connected to the processing device 521 via a serial port interface 546 that couples to the system bus. However, the input devices can also be connected by other interfaces, such as a parallel port, a game port, or a universal serial bus (USB). A monitor 547 or other type of display device may also be connected to the system bus 523 via an interface, such as a display adapter 548. In addition to the monitor 547, the personal computer 520 can include other peripheral output devices (not shown) such as, for example, speakers and a printer.

Personal computer 520 can operate in a networked environment that defines logical connections to one or more remote computers, such as remote computer 549. Remote computer 549 can be another personal computer, server, router, network PC, peer device, or other common network node, and includes many or all of the above-described elements associated with personal computer 520. Can be. 5A include a local area network (LAN) 551 and a wide area network (WAN) 552 (eg, an intranet and the Internet).

When used on a LAN, the personal computer 520 is connected to the LAN 551 via a network interface adapter card (or “NIC”) 553.
Can be connected to When used in a WAN such as the Internet, personal computer 520 may include a modem 554 or other means for establishing communication over wide area network 552. Modem 554 can be internal or external and has a serial port interface 546.
Can be connected to the system bus 523. In a network environment,
At least some of the program modules shown in connection with personal computer 520 may be stored in a remote memory storage device. The network connections shown are exemplary and other means of establishing a communications link between the computers may be used.

FIG. 5B is a more general machine 500 ′ that can implement at least some aspects of the present invention. Machine 500 'basically includes a processor 502, an input / output interface unit 504, a storage device 506, and a system bus or network 508 for facilitating data and control communications between the coupled components. Processor 502 may execute machine-executable instructions for implementing one or more aspects of the present invention. At least a portion of the machine-executable instructions and data structures may be stored (temporarily or more permanently) on storage device 506 and / or may be received from an external source via input interface unit 504. .

Having described an exemplary device that can be used to implement at least some aspects of the present invention, an overview of a text enhancement system is now provided in §4.2 below.
. Presented in 2.

(§4.2.2 Text Enhancement System) From FIG. 3 described in §1.2.2 above, most conventional systems can map the corresponding part of the image there. Recall treating pixels as individual units. Thus, in conventional systems, the same portion of the image, eg, the pixel element size, determines the light intensity value to be used in each of the RGB sub-pixel elements of the pixel element onto which the scaled image part is mapped. Used for

For a scaling font, the font unit coordinates used to define the locations of the points defining the outline of the character outline are scaled to device-specific pixel coordinates. That is, when the resolution of the em square is used to define the character outline before the character can be displayed, the character outline is converted to the size of the output device to be rendered, the conversion,
And must be scaled to reflect the features. The scaled outline describes the character outline in units that reflect the absolute units of the measurements used to measure the pixels of the output device, rather than the relative system of font unit measurements per em. Specifically, in the past, em
The values in square units were converted to values in pixel coordinate system units according to the following formula:

Size in pixel units = (character outline size / point size / resolution of output device) / (72 points per inch · number of font units per em)

However, the character size is in units of font units, and the resolution of the output device is in units of pixels / inch.

The resolution of the output device is the number of dots or pixels per inch (dpi)
Can be specified by For example, a VGA video monitor can be treated as a 96 dpi device, a laser printer can be treated as a 300 dpi device, and an EGA video monitor can be treated as a 96 dpi device in the horizontal (X) direction but in the vertical (Y) direction. Is a 72 dpi device. The font units per em can be chosen to be a power of 2, such as, for example, 2048 (= 2 ¹¹ ), but that is not necessary.

In the text enhancement system of the present invention, the image is overscaled or oversampled, specifically, in the vertical (or Y) direction. The RGB sub-pixel elements of a pixel are treated as independent intensity elements to which different parts to be overscaled or oversampled can be mapped. This overscaling operation or oversampling operation is described in §4.2.
This will be described in more detail in 2.1.1.1.

FIG. 7 is a high-level diagram of a process that can be performed by the text enhancement system. As shown in FIG. 7, an application process 710, such as a word processor or contact manager, can request to display text and specify a point size for the text. Although not shown in FIG.
Can also request the font name, background and foreground colors, and the screen location where the text should be rendered. The text and, if possible, the point size
A graphics display interface (or GDI) process (or more generally, a graphics display interface) 722 is provided. The GDI process 722 includes display information 724 (which may include display resolution information such as pixels per inch on the display) and text information 728 (which may be straight and curved, advance width information). , As well as character outline information, which can be represented as points defining a sequence of contours, such as left side bearing information), to generate a glyph (or to a cached glyph that has already been generated). to access). The glyph may include a bitmap of the scaled character outline (or bounding box 408 containing black body 407 information), advance width 402 information, and left side bearing 410 information. Each bit of the bitmap can have an associated red, green, and blue intensity value.

As described below, these values can be combined into a single, eg, 8-bit, value that can be referred to as a “compressed pixel value”. The graphics display interface process 722 is described in more detail below in § 4.2.2.1. Graphics display interface process 722, display information 724, and glyph cache 726 are, for example, Windows (from Microsoft Corporation, Lendmond, Wash.)
It can be a part of and can be implemented by an operating system 535 ', such as the CE® or Windows NT® operating system.

The glyph from either glyph cache 726 or graphics display interface process 722 is then provided to a display management process (more generally, a display driver manager) 735. Display driver management process 735 may be part of display (or video) driver 732. In general, display driver 732 may be software that allows a computer operating system to communicate with a particular video display device. Basically, the display driver management process 735 can invoke a color compensation (or color filtering) process 736, a gamma correction process 737, and a color palette selection process 738. These processes 735, 736, 737, and 738 serve to convert character glyph information to actual RGB pixel sub-component intensity values. One or more of processes 736, 737, and 738 may be implemented by a set of pre-computed look-up tables that may be used to perform multiple image processing operations. The display driver management process 735 can receive glyph and display information 724 'as input. Display information 724 'may include, for example, foreground / background color information, gamma values for display device 547, color palette information, and pixel value font information. The display information 724 'can be used to select one of the lookup tables in the set of lookup tables to be used.

The processes that can be executed by the display driver are described in §4.2.2 below.
. This will be described in more detail in Section 2.

The processed pixel values are then sent to a display (video) adapter 548 ′ as a video frame portion, along with screen positioning information (eg, from the application process 710 and / or operating system 535 ′). Will be transferred. Display adapter 5
48 'may include electronics for generating a video signal that is sent to display 547. The frame buffer process 740 can be used to store the received video frame portion in the screen frame buffer 745 of the display adapter 548. Using the screen frame buffer 745 allows, for example, a single image of a text string to be generated from glyphs representing several different characters. The video frames from the screen frame buffer 745 are then provided to a display adaptation process 750 that adapts the video to a particular display device. The display adaptation process 750 can also be performed by the display adapter 548 '.

Finally, the adapted video is presented for rendering on a display device 547, eg, an LCD display.

Having provided an overview of the text enhancement system, the graphics display interface process 722 will now be described in more detail in § 4.2.2.1 below. Next, the processes that can be executed by the display driver are described in §4.2.2.2 below.

FIG. 8 shows a process that can be performed by a graphics display interface (or GDI) process 722, as well as a GDI process 7.
22 shows data that can be used. As shown in FIG. 8, GDI process 722 includes a glyph cache management process (or, more generally, a glyph cache manager) 802 that accepts text or, more specifically, requests that text 820 be displayed. be able to. This request may include the point size of the text. The glyph cache management process 802
This request is forwarded to the glyph cache 726. If glyph cache 726 contains the glyph corresponding to the requested text character, glyph cache 726 provides it for downstream processing. If, on the other hand, glyph cache 726 does not have a glyph corresponding to the requested text character, glyph cache 726 will so notify glyph cache management process 802, which will generate the required glyph. A request for submission is submitted to a type rasterization process (or, more generally, a type rasterizer) 804. Basically, the type rasterization process 804 can be implemented in hardware and / or software, and can include character outlines (the character outlines recall include points that define contours such as straight lines and curves based on mathematical formulas). Can be converted to a raster (ie, bitmapped) image. Each pixel of the bitmap image can have, for example, a color value and intensity. The type rasterization process is described in §4.2.2.1.1.1 below.

(§4.2.2.1.1.1 Rasterizer) In repetition, the type rasterization process 804 basically converts a character outline into a bitmapped image. The scale of the bitmap can be based on the point size of the font and the resolution of the display device 547 (eg, pixels per inch). Text, font, and point size information can be obtained from the application 710, and the resolution of the display device 547 can be obtained from a system configuration file or a display driver file, or monitor settings stored in the memory 522 by the operating system 535. . Display information 724 may include foreground / background color information, gamma values, color palette information, and / or display adapter /
Display device pixel value format information may be included. Again, this information can be provided from graphics display interface 722 in response to a request from application process 710. However, if the background of the requested text should be transparent (as opposed to opaque), this background color information will be the one rendered on the display (eg, a bitmap image or other text). And provided from the display device 547 or the video frame buffer 745.

Basically, this rasterization process can include two or three sub-steps or sub-processes. First, the character outline is overscaled (or oversampled) using an overscaling / oversampling process 806. This process is described in § 4.2.2.1.1.1 below. Next, the overscaled / oversampled image generated in overscaling / oversampling process 806 can be positioned on a grid and have stretched portions using hinting process 808. This process is described in §4.2.2.1.1.2 below. The displaced (eg, immediately adjacent, overlapping, or spaced) samples of the scan conversion source samples of the overscaled / oversampled (and optionally hinted) image are then scanned. Combined by transformation process 812 (eg, filtering, averaging, etc.)
, Values corresponding to the sub-pixel elements 210, 220, 230 of the display 547 are generated. The scan conversion process 812 is described in § 4.2.2.1.1.3 below. The data stored in the obtained glyph cache 726 is described in §4.
This will be described in 2.2.1.1.4.

(§4.2.2.1.1.1 Scaling / Oversampling) From §4.2.2 above, Microsoft, Lendmond, WA
In scaling fonts in conventional systems, such as the TrueType (TM) of the Corporation, the font unit coordinates used to define the location of the points that define the outline of the character outline are scaled to device-specific pixel coordinates. Was. That is, before the characters could be displayed, the em square resolution was used to define the character outline, thus reflecting the size, transformation, and characteristics of the output device on which the characters should be rendered Had to be scaled to The scaled outline is em
Recall that the character outline is described in units that reflect the absolute units of measurement used to measure the measured pixels of the output device, rather than the relative system of font unit measurements per. Therefore, recall that the values in em square units were converted to values in pixel coordinate system units according to the following formula:

(1) Size in pixel units = (character outline size / point size / resolution of output device) / (72 points / inch font unit number / em)

Here, the character size is a font unit, and the resolution of the output device is a pixel / inch unit.

The resolution of the output device is the number of dots or pixels per inch (dpi)
Recall that can be specified by

In the text enhancement system of the present invention, with the above background in mind, the image is over-scaled or over-sampled, specifically over-scaled or over-sampled in the vertical (or Y) direction. The RGB pixel sub-component of a pixel is treated as a separate intensity element to which an overscaled or oversampled image can be mapped. Thus, a higher degree of resolution is provided than is possible with known conversion techniques. Therefore, equation (1) discussed above can be written as:

(2) Size in pixel = (N · character outline size / point size / resolution of output device) / (72 points / inch font unit number / em)

Where the character size is in font units, the output device resolution is in pixels / inch, and N is an overscaling or oversampling factor.

For an RGB horizontal stripe display device, the overscaling and oversampling factors may be zero (0) in the X direction and three in the Y direction. Such an overscaling or oversampling factor is
Scale the character outline from the font unit to sub-pixel elements (recall 210, 220, 230 in FIG. 2). In practice, it is useful to further overscale or oversample. In this way, several scan conversion source samples can be used to derive the value of each sub-pixel component. For example, if the overscaling or oversampling factor is 9 in the Y direction, using three scan conversion source samples (eg, via an averaging operation) to define the intensity of the red subpixel component of the pixel. Can be used to define the intensity of the green subpixel element of the pixel using another three scan conversion source samples, and use the last three scan conversion source samples to determine the intensity of the blue subpixel element of the pixel. The intensity can be defined. Alternative sampling and filtering techniques are also possible. For example, use more scan conversion source samples to define the green sub-pixel element,
The samples can be weighted so that fewer scan conversion source samples are used to define the blue sub-pixel element.

In one embodiment, the red subpixel element is assigned a weight of 5 and is derived from 5 scan conversion source samples, and the green subpixel element is assigned a weight of 9 and is derived from 9 scan conversion source samples. , The blue sub-pixel element is derived from the two scan conversion source samples, assigned a weight of 2, and the overscaling (or oversampling) is performed once in the X direction and 16 times in the Y direction. This is called "weighted sampling." The samples may include some overlapping scan conversion source samples such that some of the scan conversion source samples can be used to determine multiple sub-pixel elements. Alternatively, some scan conversion source samples can be ignored. Similarly, alternative filtering techniques for averaging can be used.

FIG. 6 illustrates an exemplary scan conversion operation performed by the exemplary text enhancement system. In the illustrated embodiment, different image samples 630, 632, 634 of the image 610 segmented by the grid 620 are used, and corresponding portions 640, 642 of the generated bitmap image 650.
, 644 are generated. In the example shown in FIG. 6, the image samples for red and blue are each -−１ of the distance from the green sample.
And + ／ pixel width. Therefore, the placement error introduced when rounded to the pixel precision described in § 1.2.2.1.2, encountered with the well-known sampling / image display method shown in FIG. Decreases to a barely perceptible level. To obtain sub-pixel resolution, the rasterized characters, ie, the bitmaps generated by the rasterization process, are typically overscaled and / or oversampled in a direction perpendicular to the striping of the display device. Oversampling can be viewed as compressing the grid of samples while maintaining the image on which the grid is located. On the other hand, overscaling can be viewed as maintaining a grid of samples while stretching the image on which the grid lies.

In general, the overscaling (or oversampling) process 806 can perform non-square scaling (or sampling) based on the direction and / or number of sub-pixel elements included in each pixel element. Specifically, it is recalled that high resolution character information 728 (this information can include contours defined by a sequence of points defining straight lines and curves).
) Can be overscaled (or oversampled) in a direction perpendicular to the striping at a rate greater than the direction of the striping. Thus, for example, when a vertically striped display is used as a device on which to display data, scaling is performed in the horizontal (or X) direction at a greater rate than is performed in the vertical direction. You.

On the other hand, when using a horizontally striped display of the type as shown in FIG. 1 as a device to display data, the scaling is performed at a greater rate than vertically ( Or in the Y) direction. Such non-uniform scaling allows subsequent image processing operations to take advantage of a higher degree of resolution that can be achieved by using individual sub-pixel elements as independent light intensity sources in accordance with the present invention. It becomes possible. Thus, in most cases, overscaling (or oversampling) of a character or image will further divide the red, green, and blue stripes, thereby allowing subsequent scan conversion operations to be supported. And, although not required, in the direction perpendicular to striping.

Next, the first overscaling (or oversampling) technique will be described with reference to FIGS. 10 and 11, and the second overscaling (or oversampling) technique will be described with reference to FIGS. 12 and 13. .

FIG. 10 illustrates an overscaling (or oversampling) process 806.
Is a high-level flow diagram of a first method 806 'for implementing. FIG.
Shows an example of the operation of the method 806 'of FIG. First, as shown in step 1010, font vector graphics (eg, character outlines), point size, and display resolution are accepted. This information is shown in FIG.
As shown. Left side bearing (recall 410), advance width (4
Font metrics such as 0 (recall 02), ascent (recall 942) and decent (recall 944) can also be accepted.
This information is shown in FIG. 11 as 1112 and 1114. As shown in step 1020, an overscaling factor or oversampling rate (recall N in Equation 2) is accepted. Then, as shown in step 1030, the font vector graphics (e.g., character outline) 1110 has a point size,
Rasterized based on display resolution and overscaling factor (or oversampling rate).

As shown in the example of FIG. 11, the X coordinate of the character outline (in font units), the advance width (also in font units) and the left side bearing are scaled as shown in 1120 (Eq. 1), rounded to the nearest integer pixel value. On the other hand, the Y-coordinate value of the character outline (in units of font units), and also (in units of font units)
The ascent, decent, and other vertical font feature values are overscaled (recall Equation 2) as shown at 1130 and rounded to the nearest integer scan conversion source sample value. The resulting data 1140 is a character outline in pixels in the X direction and in scan conversion source samples in the Y direction. The method 806 'then ends via RETURN node 1040.

FIG. 12 illustrates an overscaling (or oversampling) process 806.
FIG. 13 is a high level flow diagram of a second method 806 "for implementing
An example of the operation of the method 806 "is shown below. First, as shown in step 1210, font vector graphics (eg, character outlines), point size, and display resolution are accepted.
As shown. Font metrics such as left side bearing and advance width can also be accepted. As shown in step 1220, an overscaling factor or oversampling rate (recall N in Equation 2) is accepted. Then, as shown in step 1230, the sub-pixel element of the display is for example R
It is determined whether they are arranged in the order of -GB or BGR. If the subpixel elements are arranged in RGB order, the font vector graphics (eg, character outline) 1310 is rotated 90 degrees counterclockwise. FIG.
Referring to FIG. 3, this transformation can be performed by changing the original X coordinate value to the new Y coordinate value and inverting the original Y coordinate value to generate a new X coordinate value as shown at 1312. .

Next, the process proceeds to Step 1260. Returning to decision step 1230, if the sub-pixel elements are arranged in BGR order, the font vector graphics (e.g., character outline 1310 rotates 90 degrees clockwise).
This transformation can be performed by changing the original X coordinate to the new Y coordinate and changing the original Y coordinate to the X coordinate. Note that in the above example, it was assumed that there were no negative X values-ie the Y axis should be to the left of the font vector graphics (e.g., 1310). In an example where there may be a negative X coordinate value, when rotated 90 degrees clockwise, the X value is inverted and a Y value is generated. Then, the process proceeds to step 1260.

In step 1260, the rotated font vector graphics (eg, character outline) 1110 is rasterized based on the point size, display resolution, and overscaling factor (or oversampling rate).
As shown in the example of FIG. 13, the new Y coordinate value of the character outline (in units of font units), and the advance width and left side bearing (in units of font units) are scaled as shown at 1330 (expression 1). These values are then rounded to the nearest integer pixel value. On the other hand, (
The new X coordinate value of the character outline (in font units) and the ascent and decent values (also in font units) are 1320
Is over-scaled (or over-sampled) (recall Equation 2) and rounded to the nearest integer scan conversion source sample value. The resulting data 1340 is a character outline in pixels in the new Y direction and scan conversion source samples in the new X direction. Returning to FIG. 12, method 806 ″ ends via RETURN node 1270.

Although the first method 806 ′ is easy to implement, the second method 806 ″ allows for some data access when accessing data 1340 during the scan conversion process 812, in some embodiments. Benefits are provided: these benefits are illustrated in FIG.
This will be described in §4.2.2.1.1.3 below with reference to FIG.

Returning to FIG. 8, the overscaling (or oversampling) process 8
After performing 06, an optional hinting process 808 may be performed. This process 808 is described in § 4.2.2.1.1.2 below.

(§4.2.2.1.1.2 Hinting) The purpose of hinting (also called “glyph instructions”) is when the glyph is rendered at different sizes and on different devices. It is to ensure that important features of the font design are preserved. Consistent stem weight, consistent "color" (ie, black and white balance on a page or screen under this circumstance), as well as avoidance of spacing and pixel dropout, are common objectives of hinting is there. In the past, undirected or non-hinting fonts have generally produced good quality results at sufficiently high resolutions and point sizes. However, for many fonts, smaller point sizes on lower resolution displays can compromise readability. For example, at lower resolutions where fewer pixels are available to describe a character shape, features such as weights, crossbar widths, and serif details can be irregular, inconsistent, and even completely erroneous. is there. Although the readability is improved by the higher resolution provided by the present invention, hinting is still useful.

Basically, hinting includes “grid arrangement” and “grid fitting”. Grid placement is used to arrange the scaled (or over-scaled) characters in a grid, and is used by subsequent scan conversion operations to optimize the accurate display of characters using the available sub-pixels. Can be Grid fitting involves deforming the character outline so that the character better matches the shape of the grid.

The grid fitting ensures that some features of the glyph are adjusted. Since the outline is deformed to a size smaller than the specified size, the outline of the font at high resolution remains unchanged and undeformed.

In a grid arrangement, the boundaries of sub-pixel elements are treated as boundaries along which characters can be aligned or should be aligned, or outlines of characters should be adjusted for. be able to.

In many cases, hinting involves aliasing the left edge of the character stem at the left pixel or sub-pixel boundary, and aligning the bottom of the base of the character along the pixel component or sub-pixel element boundary. including. Experimental results show that in the case of vertical striping, letters with stems arranged with a blue or green left edge tend to be more readable than letters with stems arranged with a red left edge. It was shown that there is.

Thus, in at least some embodiments, when hinting characters are to be displayed on a screen with vertical striping, the blue or green left edge relative to the stem is part of the hinting process, Preferred over the red left edge. In the case of horizontal striping, generally the bottom of the character features, especially horizontal character features (such as the crossbars of the letters A and H and the bottom features of the letters E and Z) are arranged to have a red or blue bottom edge. Letters are generally easier to see than letters having a base arranged to have a green bottom edge. Thus, when hinting characters are to be displayed on a screen with horizontal striping, in at least some embodiments, a red or blue bottom edge is preferred over a green bottom edge as part of the hinting process. .

FIG. 14 is a high-level flowchart of an exemplary method 808 ′ for performing at least a portion of the hinting process 808. FIGS. 15A and 15B illustrate the operation of the method 808 'of FIG. First, as shown in step 1410,
An overscaled or oversampled character bitmap is accepted. As described above, if the bottom of a character feature is on a scan conversion source sample that is scan converted to a green subpixel element, then a scan conversion source sample that is scan converted to a blue or red subpixel element, or blue or red It should be shifted (or transformed) to a scan conversion source sample adjacent to the scan conversion source sample that is scan converted to a subpixel element. Thus, as shown in decision step 1420, if the bottom of the character feature is not on a scan-converted source sample that is scan-converted to a green subpixel element, method 808 'ends via RETURN node 1460. On the other hand, if the bottom of the character feature is on a scan-converted source sample that is scan-converted to a green sub-pixel element, processing branches to decision step 1430, where the bottom of the character feature is scan-converted to a red sub-pixel element. Is determined to be close to the scan-converted source sample to be converted or to a scan-converted source sample that is scan-converted to a blue sub-pixel element. In the former case, as shown in step 1440,
The character feature outline is shifted (compressed) such that the bottom of the character feature outline is on a scan-converted source sample that is scan-converted to a red subpixel element. In the latter case, as shown in step 1450, the character feature outline is placed so that the bottom of the character feature outline is on (or directly adjacent to) the scan-converted source sample that is scan-converted to a blue subpixel element. Shift (expansion)
Is done. In either case, method 808 'ends via RETURN node 808'.

FIGS. 15A and 15B show that five scan conversion source samples are used to derive the red sub-pixel element, nine scan conversion source samples are used to derive the green sub-pixel element, and 2 Operation of method 808 'of FIG. 14 for a character outline in an overscaled or oversampled character outline to be subjected to a weighted scan conversion in which one scan conversion source sample is used to derive the green subpixel element. Is shown. FIG. 15A
In the example, the bottom 1512 of the overscaling or oversampled character feature outline 1510 is at a scan conversion source sample to be scan converted to a green sub-pixel element, but a scan conversion to be converted to a blue sub-pixel element. Close to the source sample. Thus, the overscaled or oversampled character feature outline is such that the bottom 1512 'of the resulting overscaled or oversampled character feature outline 1510' is immediately adjacent to the scan conversion source sample to be scan converted to a blue subpixel element. Are shifted down (or stretched) so that they are on the corresponding scan conversion source sample.

In the example of FIG. 15B, the bottom 1522 of the overscaling or oversampled character feature outline 1520 is at the scan conversion source sample to scan convert to a green subpixel element, but scans to a red subpixel element. Close to the scan conversion source sample to be converted. Thus, the overscaling or oversampled character feature outline is the bottom 1522 of the resulting overscaled or oversampled character feature outline 1520 '.
′ Is shifted upward (or compressed) so that it is on the scan conversion source sample immediately adjacent to the scan conversion source sample to be scan converted to a red subpixel element.

[0098] Other hinting instructions well known to those skilled in the art may also be implemented for overscaling or oversampled character outlines. However, it should be noted that due to the additional vertical resolution provided by considering sub-pixel elements, certain hinting instructions may be unnecessarily executed.

After completing hinting process 808 or hinting process 80
If 8 was not performed, overscaling (or oversampling)
After process 806 is completed, scan conversion process 812 is performed. This scan conversion process is described in §4.2.2.1.1.3 below.

(§4.2.2.1.1.3 Scan Conversion) Basically, the scan conversion process 812 converts an overscaled (or oversampled) geometry representing a character into a bitmap image. I do. Conventional scan conversion operations treat pixels as individual units to which corresponding portions of the image to be scaled can be mapped. Thus, in conventional scan conversion operations, the same portion of the image determines the light intensity value to be used at each of the red, green, and blue sub-pixel elements of the pixel element to which the portion of the image to be scaled is mapped. Used for FIG.
Recall that an example of a well-known scan conversion process involves sampling an image to be represented as a bitmap and generating intensity values from the sampled values.

In the scan conversion process 812 of the text enhancement system, the red, green, and blue sub-pixel components of a pixel are treated as independent intensity components. Therefore,
Each sub-pixel element is treated as a separate intensity component to which different parts of the overscaled (or oversampled) image can be mapped. By allowing different parts of the image to be overscaled (or oversampled) to be mapped to different subpixel elements,
A higher degree of resolution is provided than known scan conversion techniques. In other words, different portions of the image that are overscaled (or oversampled) are used to independently determine the light intensity value to use for each subpixel element. Thus, the scan conversion process involves filtering (eg, averaging) the displaced samples.
Can be considered. The displaced samples can be overlapped, then directly adjacent, and / or then spaced.

Recall that FIG. 6 illustrates an exemplary scan conversion process 812 that can be used in a text enhancement system. In the illustrated embodiment, different image samples 630, 6 of images 610 segmented by grid 620 are shown.
32, 634 are used to generate red, green, and blue intensity values associated with corresponding portions 640, 642, 644 of the generated bitmap image 650.
In the example shown in FIG. 6, the image samples for red and blue are displaced by-／ and + ／ pixel height, respectively, at a distance from the green sample.

The scan conversion process 812 produces red, green, and blue (R, G, B) intensity values for each pixel sub-component. These values can be expressed in the form of separate red, green, and blue light intensity levels.

Having provided a brief overview of the scan conversion process 812, two exemplary methods for implementing the scan conversion process 8.12 will now be described with reference to FIGS.

FIG. 16 is a high-level flowchart of an exemplary method 812 ′ for performing the scan conversion process 812. As shown at step 1610, a hinted and overscaled (or oversampled) character outline bitmap is accepted. Similarly, as shown in step 1620, the overscaled (or oversampled) glyph metric is also accepted. Then, step 16
As shown at 30, the expected size glyph metric is determined from the overscaled (or oversampled) glyph metric. More specifically, some of the metrics accepted by scan conversion process 812 (
For example, ascent and decent) are oversampled by N (eg, N = 16) in the Y direction (recall FIGS. 11 and 13) so that their overscaled (or oversampled) glyph metrics are When converted to a glyph metric of the expected size (ie, scaled for the particular output device on which the text is to be rendered), the ascent (
The body or size of the left side bearing (also called the Y component) and the glyph (ie, the sum of the ascent and the absolute value of the decent) can be determined as follows.

Ascent @ Ceiling (Overscaled Ascent / N) (3)

Body Height≡ Ceiling (overscaled ascent / N) + | Ceiling (overscaled decent / N) | (4)

Note that “sealing” is an operator that rounds a non-integer to the next largest integer.
This step is used to ensure that there are complete pixels for the scan-converted source samples of the overscaled (or oversampled) character outline bitmap.

Next, as shown in step 1640, the remaining scan conversion source samples in the ascent and decent of the character outline are zero padded. More specifically, if the number of scan conversion source samples in the ascent is not evenly divisible by the overscaling (or oversampling) factor N, then the number of scan conversion source samples in the ascent is overscaling (or oversampling) factor N Additional scan conversion source samples with the value zero (0) are added to the character outline ascent until is evenly divisible by. Similarly, if the scan conversion source samples during decent are not evenly divisible by the overscaling (or oversampling) factor N, then the number of scan conversion source samples during decent is evenly divisible by the overscaling (or oversampling) factor N. Until,
An additional scan conversion source sample having the value zero (0) is added to the decent of the character outline. As will be shown later with reference to FIGS. 18A and 18B,
This step ensures that the baseline of adjacent characters does not "jump" or "bounce". Note that other techniques that are functionally equivalent to zero padding can be used instead. Such functionally equivalent techniques include, for example, having the scan conversion process access an integer multiple of the overscaling factor (or oversampling rate) of the scan conversion source samples above the baseline, and ascenting the character outline. May include ignoring (eg, using a masking operation) the scan conversion source samples above.

Next, as shown in step 1650, sub-pixel element values are determined based on the scan conversion source samples. As discussed above, this determination is made by filtering the displaced samples. be able to. FIG. 19 shows an example of the weighted scan conversion process. The intensity value of the red subpixel element is based on five scan conversion source samples, the intensity value of the green subpixel element is based on nine scan conversion source samples, and the intensity value of the blue subpixel element is two scan conversion source samples. Because it is sample based, this exemplary scan conversion process is called "weighting."
This weighting can be used because the human eye perceives the luminosity from different colored light sources at different rates.

For the perceived brightness of white pixels obtained by setting the red, green, and blue sub-pixel elements to their maximum intensity output, green is about 60% and red is about 3%.
0% and blue contribute about 10%. Thus, when allocating resources such as, for example, luminosity levels, a higher level than blue or red can be allocated to green.
Similarly, higher intensity levels can be assigned to red and then to blue. However, in some embodiments, intensity levels equal to red, green, and blue are assigned. The overscaling (or oversampling) factor in this example was 16.

In FIG. 19, the dashed line 1910 indicates overscaling (or oversampling).
Shows the part of the character outline that has been drawn. Reference numeral 1912 indicates a region inside the character outline 1910, and reference numeral 1914 indicates a region outside the character outline 1910. FIG. 19 shows a scan conversion source sample corresponding to two pixels having six sub-pixel elements of a display on which the character 1910 is to be rendered. In this example, the filtering operation is simply that the center 1922 of the scan conversion source sample 1920 therein is within the character outline 1910 or character outline 1910 with the overscaled (or oversampled).
Add the number of samples that are above 10. Alternatively, the number of scan conversion source samples that are at least 50% within the character outline 1910 can be used.
Note that the highest source subpixel is zero padded so that the ascent of the character outline 1910 is evenly divisible by 16 (step 16).
Remember 40. ). In a first set of scan conversion source samples 1930a, the red value 1932a is 3, the green value 1934a is 8, and the blue value 1936
a is zero (0). In a second set of scan conversion source samples 1930b,
The red value 1932b is zero (0), the green value 1934b is zero (0),
The blue value 1936b is 2.

Returning to FIG. 16, in optional step 1660, the sub-pixel element values are “compressed”. In the exemplary embodiment, the sub-pixel element values are compressed into a single 8-bit value according to the following equation:

Pixel value to be compressed≡3 × (10 × red + green) + blue (5)

Thus, for example, the compressed pixel value corresponding to the first set of scan conversion source samples 1930a is 114 (= 3 × (10 × 3 + 8) +0), and the compressed pixel value corresponding to the second set of scan conversion source samples 1930b. The pixel value is 2 (= 3 × (
10 × 0 + 0) +2). The compressed pixel value is zero (0) (ie, red,
Between green and blue values are all zero (0)) and 179 (ie, when the red value is 5, the green value is 9 and the blue value is 2). Will have a value.

Next, in step 1670, the character bitmap and glyph metrics are stored in glyph cache 726. An exemplary data structure for data stored in glyph cache 726 is described in § 4.2.2.1.1.4 below. Method 8
12 'ends via RETURN node 1680.

FIG. 17 is a high level flow diagram of an alternative method 812 ″ for implementing the scan conversion process 812. The method 812 ″ of FIG. 17 uses the overscaling (or oversampling) method 806 ″ of FIG. Steps 1240 and 1250 in FIG. 12 and glyph 1314 in FIG.
Recall that the character outline 1310 rotates 90 degrees. Accordingly, in the method of claim 17, the scan conversion glyph is rotated back as shown in step 1770. More specifically, for a video display device having sub-pixels arranged in the order RGB, the Y coordinate is mapped to the final X coordinate and the negative X coordinate is mapped to the positive final Y coordinate. And the positive X coordinate is mapped to the negative final Y coordinate. Such coordinate mapping can be performed automatically such that the scan conversion source samples are accessed (as can be seen from the description of FIG. 20B below). Otherwise, the method 812 ″ of FIG. 17 is similar to FIG.

As discussed in § 4.2.2.1.1.1 above, the overscaling (or oversampling) method 80611 of FIG. 12 is equivalent to the overscaling (or oversampling) method 806 ′ of FIG. Scan conversion process 8 such as scan conversion method 812 "of FIG. 17 than when storing scan conversion source sample information.
12 to store scan conversion source sample information for easy access. 20A and 20B illustrate this difference in the ease with which the scan conversion source sample information can be accessed by the scan conversion process 812. As shown in FIG. 20A, the scan conversion source sample information from the overscaled (or oversampled) character outline 1140 is divided into a number of bytes from left to right, top to bottom, and denoted by reference numeral 2010. It can be stored as a series of bytes as indicated. The scan conversion accesses several (eg, 16) consecutive scan conversion source samples in the Y direction, so that, for example, byte 200
One bit must be accessed from 0 _{1,1 '} , i bytes must be skipped, and one bit must be accessed from bytes 2000 _2,1' .

On the other hand, as shown in FIG. 20B, the scan conversion source information from the rotated and overscaled (oversampled) character outline 1340 is divided into several bytes from left to right and top to bottom. , 2010 ′, can be stored as a series of bytes. However, in this case, since the scan conversion accesses several (eg, 16) consecutive bits in the x-direction (originally in the Y-direction), it only has two consecutive bytes (eg, 2000 _{1,1 ′} , 2000 _{1). , 2 '} ). If there is a scan conversion source sample corresponding to the space between the body 940 of the character outline and the em box 910 (see FIG. 9), an access method that compensates for such an offset can be used.

In step 1640 of FIG. 16 and step 1740 of FIG. 17, the remaining scan conversion source samples in the ascent and decent of the character outline were zero-padded. More specifically, if the number of scan conversion source samples in the ascent is not evenly divisible by the overscaling (or oversampling) factor N, then the number of scan conversion source samples in the ascent is overscaling (or oversampling) factor N Recall that additional scan conversion source samples with the value zero (0) are added to the ascent of the character outline until they are evenly divisible by.

Similarly, if the number of scan conversion source samples during decent is not evenly divisible by an overscaling (or oversampling) factor N, then the number of scan conversion source samples during decent is overscaling (or oversampling). Additional sub-pixel source samples having the value zero (0) are added to the decent of the character outline until divisible by a factor N. FIG. 18A
18B and FIG. 18B serve to show how these steps prevent "jumping" or "bouncing" of the baseline.

In FIG. 18A, assume that source sample 1850a is the highest subpixel source sample in the overscaled (or oversampled) character outline. Reference numeral 1810b indicates the first 16 scan conversion source samples above the baseline 1820 (if the overscaling or oversampling factor is 16), and reference numeral 1810a indicates the last 16 scan conversion source samples above the baseline 1820. Fig. 4 shows scan conversion source samples. If you do not use zero padding,
The first set of samples will take scan conversion source samples from set 1810a and one scan conversion source sample from the next set (not shown). As more sets of samples are processed by the scan conversion process 812, this offset is due to the fact that the set of samples takes the scan conversion source samples from the set 1810 above the baseline 1820 and the first set below the baseline 1820. From one to one scan conversion source sample. In effect, the result of not zero-filling the top set of scan conversion source samples 1810a is that the baseline 1820 moves down one scan conversion source sample. By zero padding the top set of scan conversion source samples 1810a, the position of the baseline 1820 is maintained.

Referring now to FIG. 18B, it is assumed that scan conversion source sample 1850c is the best scan conversion source of the overscaled (or oversampled) character outline. Reference numeral 1810d indicates the first 16 scan conversion source samples above the baseline 1820 (if the overscaling or oversampling factor is 16), and reference numeral 1810c indicates the last 16 scan conversion source samples above the baseline 1820. Fig. 4 shows scan conversion source samples. Without zero padding, the first set of samples would take one scan conversion source sample from set 1810c and 15 scan conversion source samples from the next set (not shown). As more sets of samples are processed by the scan conversion process 812, this offset is due to the set of samples taking one scan conversion source sample from the set 1810d above the baseline 1820 and the second sample below the baseline 1820. It will propagate down to take 15 scan conversion source samples from a set. Effectively, the result of not zero-filling the top set of scan conversion source samples 1810c is that the baseline 1820 moves down the scan conversion source samples. By zero padding the top set of scan conversion source samples 1810c, the position of the baseline 1820 is maintained. As can be seen from a comparison of FIG. 18A and FIG. 18B, if the scan conversion source samples in the top set of N (eg, N = 16) scan conversion source samples are not zero padded, the baseline 1820 The position can move or "jump" from one character to the next.

Having described the scan conversion process 812, an exemplary data structure for storing the resulting glyphs and glyph metrics is described in § 4.2.2.1.1.4 below.

(§4.2.1.1.4 Exemplary Compressed Pixel Value Data Structure) In many systems, the R, G, and B light intensity values are specified and stored as three discrete quantities. ,It is processed. Each of the three discrete quantities has a number corresponding to the number used to specify a sub-pixel element intensity value for display adapter 548 and / or display device 547. For example, many systems use 8-bit quantities, each representing an R, G, or B light intensity value. In such an embodiment, the R, G, and B light intensity values require 24 bits of storage, processing, and transfer per pixel.

For some devices where memory, processing, and even bus bandwidth are a limited resource (eg, portable computers and handheld computing devices), each R, G, and B light intensity throughout the rendering process Using eight bits to represent a value can be a significant burden on available resources. To reduce the resources, including memory, required to process and store glyphs, separate R, G, and B intensity level values can be converted to a single number, eg, a single 16 and FIG.
7 Steps 1660 and 1760 and their associated §4.2 above.
. Recall the explanation in 1.1.3. ).

Again, this number is referred to as a "compressed pixel value" because it represents compressing the R, G, and B intensity values associated with the pixel into a single value. The range of numbers used to represent the light intensity levels of pixels R, G, and B, for example, the range of compressed pixel values, can uniquely identify each possible combination of R, G, and B light intensity levels. It is chosen large enough to allow. Therefore, R
, G, and B light intensity levels, the total number of compressed pixel values used to represent the total number of supported red intensity levels, the total number of green intensity levels, and the number of blue intensity levels Should be at least the same as the product of the total numbers. The product should be able to be specified as an 8-bit quantity or a multiple thereof, as it is often convenient to work with bytes, ie, 8-bit quantities, in terms of memory access, processing, and data transfer operations. In a handheld computing device, a significant savings in terms of memory and the like (compared to an embodiment using 8 bits per luminosity value of a sub-pixel element and requiring a total of 24 bits per pixel) A single 8-bit representation of the product of the intensity values of R, G, and B is particularly desirable.

In general, to limit the resource load associated with a rendered image, specifically a text image, the scan conversion process 812 compresses the separate R, G, and B intensity values associated with the pixel. Can be converted to pixel values.
In such an embodiment, the glyphs are represented and stored using compressed pixel values rather than, for example, as separate 8-bit R, G, and B intensity values. The compressed pixel representation can be converted to separate R, G, and B intensity values in the form used by display device 547 before the intensity values are provided to display adapter 548.

The conversion of the separate R, G, and B light intensity levels to compressed pixel values can be performed as part of the scan conversion process 812 or as a post process to the scan conversion process 812. (Recall steps 1660 and 1760 of FIGS. 16 and 17, respectively, above.) Using shift operations or arithmetic expressions, separate R, G, and B light intensity levels and compressed pixel values associated with pixels Can be converted between. Such an operation may generate a total of M (0 to M-1) distinct compressed pixel value entries. Where M is a possible R, G,
And B are the total number of combinations of light intensity levels. Corresponding R, G, and B
Are associated with each compressed pixel value.

If RP is the maximum possible number of red light intensity levels, the light intensity value of R varies from 0 to RP-1. Let GP be the maximum possible number of green light intensity levels,
The luminosity value of G varies from 0 to GP-1. If BP is the maximum possible number of blue light intensity levels, the light intensity value of B will vary from 0 to BP-1. Or,
The compressed pixel values are separated using the look-up table by using the compressed pixel values as an index into the look-up table and outputting the individual R, G, and B intensity level entries associated with the compressed pixel values. It can also be converted to R, G, and B luminosity values.

Recall that the exemplary scan conversion processes 812 ′ and 812 ″ described above supported 16 times overscaling (or oversampling) in the direction perpendicular to RGB striping. Luminosity levels, for example levels 0-5, were used, ten green luminosity levels, for example levels 0-9, and three blue luminosity levels, for example levels 0-2, were used. A total of 180 (= 6 × 10 × 3) possible R, G, and B intensity level combinations are obtained, In such an embodiment, N is equal to 180, and the look-up table is compressed pixel values 0 to 0. 179.

The number of R, G, and B intensity levels supported is typically the number of scan conversion source samples used to determine the R, G, and B intensity levels during the scan conversion process 812. Note that it is a function of numbers. 180 possible R
, G, and B intensity levels In an exemplary embodiment in which a combination of intensity levels is supported, each of the compressed pixel values 0-179 can also be represented using an 8-bit quantity. This significantly reduces storage requirements compared to embodiments where separate 8-bit values are used for a total of 24 bits per combination of R, G, and B.

FIG. 21 shows an example of information that can be stored in glyph cache 726 ′ by scan conversion process 812. Glyph cache 726 'may include several glyph files 2110a. Each of the glyphs 2110a has a number of glyph metrics 2112a (eg, left side bearing, advance width, ascent, decent, etc.) and a number of pixel records 2120.
Can be included. Each of the pixel records 2120 may include a display screen pixel coordinate 2122 and a compressed pixel 2124.

Having described the graphics display interface process 722, the display driver 723 and its associated processes are described below in §4.2.
This will be described in 2.2.

(§4.2.2.2 Display Driver Components) The display driver 732 allows the computer system to communicate information to the video display device 547 or video adapter 740 so that the video display device 547 can interpret it. Software instructions that enable the Display driver 732 is operating system block 5
Although shown outside 35 ', the display driver 732 can be considered part of the operating system 535'. As shown in FIG. 7, the display driver 732 can accept the display information 724 ′ and perform a color compensation (or color filtering) process 736, a gamma correction process 737.
, And a display driver management process 735 that coordinates the color palette selection process 738. Display information 724 'can include the gamma of display device 547 and the color palette of display device 547.

FIG. 22 is a high-level flowchart of a method 735 ′ that can be used to implement the display driver management process 735. As shown in steps 2210 and 2220, the display driver management method uses the glyph cache 7
Accepts 26-735 'glyphs and accepts display information from display 547 (or accepts display information about display device 547 from system configuration file). The method 735 'then proceeds to step 223.
At 0, the color filter processing process 736, described in more detail below in § 4.2.2.1, can be invoked. The method 735 'may also invoke, at step 2240, a gamma correction process 737, which is described in more detail below in § 4.2.2.2.2. Finally, the method 735 ′ includes, in step 2250, the following §4
. A color palette selection process 738, described in more detail in 2.2.2.3, can be invoked. Method 735 'ends via RETURN node 2260.

§ 4.2.2.2.2.1 Color Compensation (Filtering) The red 210, green 220, and blue 230 subpixels are considered separately by the overscaling (or oversampling) and scan conversion processes. This effectively increases the resolution of the display device in the vertical direction, but if the intensity values of adjacent sub-pixels are very different, the resulting display may be visually unpleasant for the user. If the intensity differences are too large, a color compensation (color filtering) process 736 can be used to reduce the intensity differences between certain adjacent sub-pixel elements.

FIG. 23 is a flowchart of an exemplary method 736 ′ that may be used to implement the color filter processing process 736. First, as shown in step 2310, the filter parameters are accepted. In this method 736 ', two filters are provided, a red filter and a blue filter. In this case, the red filter uses the threshold, the red factor, and the green factor. The blue filter uses a threshold, a green factor, and a red factor. As shown, the loop defined by steps 2320 and 2380 is performed for each compressed pixel value of the glyph being processed. Within loops 2320-2380, normalized red, green, and blue intensity values are determined from the compressed pixel values, as shown in step 2330. More specifically, for example, color spaces 0 to 255 can be divided into equal segments based on weighted colors. Thus, for example, 5
If there are two red colors, the color space is divided into five segments, each separated by 255/5. This results in six unique red colors, which are normalized to the color space. Therefore, the normalized color can be determined using the following equation:

(Total number of colors−desired color index) * 255 / total number of colors (6)

Then, in decision step 2340, it is determined whether the absolute value (ie, magnitude) of the difference between the red and green intensities is greater than the red filter threshold value. If it is greater than the red filter threshold, as shown in step 2350, the red and green intensities are
It is adjusted so that the magnitude of the difference is reduced. For example, this part of the step can be performed according to the following equation:

When (RG)> red filter threshold, R ′ = R − ((RG) * red factor of red filter) / 10 G ′ = G + ((RG) * Green factor of red filter) / 10

Where R is the original red intensity, G is the original green intensity, R 'is the new red intensity, and G' is the new green intensity. Then, the process proceeds to the determination step 2360
Proceed to. Returning to decision step 2340, if the absolute value (ie, magnitude) of the difference between the red and green intensities is not greater than the red filter threshold, processing proceeds directly to decision step 2360.

At decision step 2360, it is determined whether the absolute value (ie, magnitude) of the difference between the green (which may have been corrected at step 2350) and the blue intensity is greater than the blue filter threshold value. Is determined. If so, the green intensity (modified intensity if modified in step 2350), blue intensity, and / or red intensity (modified in step 2350), as shown in step 2370. Corrected intensity) is adjusted to reduce the magnitude of the difference. For example, this part of the step can be performed according to the following equation:

If (GB)> blue filter threshold, then G '= G-((GB) * green factor of blue filter) / 10 B' = B + ((GB) * R '= R-((GB) * red factor of blue filter) / 10

Here, R is the original red intensity (modified intensity if modified in step 2350), and G is the original green intensity (modified intensity if modified in step 2350). Yes, B is the original blue intensity, R 'is the new red intensity, G' is the new green intensity, and B 'is the new blue intensity. Then the process is Step 2
Proceed to 380. Returning to decision step 2360, the absolute value of the difference between the green and blue intensities (
If the size is not greater than the threshold value of the blue filter, the process proceeds directly to step 2380. After all compressed pixels of the glyph have been processed, method 73
6 'ends via RETURN node 2390.

Some sample values for the filter variable values are as follows: Red filter threshold = 100 red filter red factor = 3 red filter green factor = 2 blue filter threshold = 128 blue filter red factor = 2 blue filter green factor = 1 blue filter Filter blue factor = 3

Although the color filtering process is shown as part of the display driver 732, instead of applying scan converted sub-pixel element intensity values to these values before they are combined into compressed pixel values. Please note that it was also possible.

(§4.2.2.2 Gamma Correction) Many display devices 547 exhibit non-linear behavior between an input signal and its output. More specifically, when the input signal to the display device 547 is i and the output of the display device 547 is o, the relationship between i and o is
Generally, it can be expressed as follows.

O = ci ^γq (7)

Here, c is a constant, and 7 is an index that is generally not 1 (generally called “gamma” of the device). For LCD displays, gamma is generally less than one. To ensure that the perceived grayscale in the image is correctly rendered on the display 547, an additional compensated non-linear device, commonly referred to as a "gamma is a collector", is often used. Therefore, referring to step 2240 of FIG. 22, the gamma correction process is activated.

FIG. 24 is a high-level flowchart of an exemplary gamma correction method 2240 ′ that can be used to implement the gamma correction process 2240. This method 2240
'Can optionally normalize the intensity values to the extent that gamma correction is most effective. First, in optional step 2410, a "useful" upper and lower limit is accepted. These boundaries reflect the range of intensity where gamma correction is most effective. When the intensity range is from 0 to 255, for example, the lower boundary may be set to 30 and the upper boundary may be set to 250. At step 2420, the intensity value is accepted.
Then, in optional step 2430, the intensity values are normalized to a "useful" range of intensity. This normalization can be performed by simply setting the intensity below the lower limit to the lower limit and clamping the intensity above the upper limit to the upper limit. Alternatively, normalization can be performed by shifting and clamping the intensity. In this alternative technique, assuming an upper limit of 250 and a lower limit of 30, 0 to 255
Is shifted up from 15 to 270 by, for example, 15
The intensity between 5 and 29 is clamped at 30 and the intensity between 251 and 270 is 25
It can be clamped to clamp to zero. In another embodiment, the intensity is scaled and rounded to be within a "useful" range of intensity. In this case, assuming a lower bound of 30 and an upper bound of 250, 256 possible values between 0 and 255 are scaled (and rounded) to 221 possible values between 30 and 250. Become. Of course, other techniques for normalizing the intensity to the "useful" range can also be used. In step 2440, the gamma of device 547 is accepted, and in step 2450, the (normalized) intensity value is adjusted based on the gamma of device 547. Method 2240 'then returns to RETURN node 246.
Exit through 0.

(§4.2.2.2.3 Color Palette Selection) The display device has a generally available color palette. Thus, the color-filtered and gamma-corrected red-green-blue intensity value triplet is mapped to the closest color of the display device 547.

As described above, the display device 724 'can be used to select one of the lookup tables in the set of lookup tables to be used. The look-up tables in the set of look-up tables are used to convert compressed pixel values used to represent glyphs into pixel values. The processed pixel values are in the form of, for example, 8-bit R, G, and B light intensity values, and are used by display adapter 548 'and / or display device 547. Each look-up table contains one entry for each potential compressed pixel value and corresponding output value. In the exemplary case of 180 possible compressed pixel values, the look-up table includes 180 compressed pixel values and 18D corresponding processed (eg, output) pixel values.
Each output pixel value can be a pre-computed value generated by performing a display driver processing operation performed using the compressed pixel value to which the output pixel value corresponds as input.

A set containing R, G, and B light intensity values in the form used by the attached display adapter or display device by using the compressed pixel values representing the glyphs as indices in a look-up table. Is obtained.

The processed pixels included in the look-up table may have been pre-calculated. That is, it can be calculated before it is used by the display driver 732. By pre-computing the pixel values to be processed, real-time gamma correction, color compensation, and / or palette selection operations during image rendering are avoided.

An alternative to the pixel value look-up table approach can also be used. Gamma correction, color compensation, and / or palette selection operations are performed using the compressed pixel values as input, and processing the processed pixel values in a format suitable for use with display adapter 548 'and / or display 547. Can be generated.

(§4.2.2.3 Display Adapter) A conventional display adapter 548 ′ operates in a conventional manner to buffer videograms and adapt video frames for output to a display device 547. can do.

(§4.3 Conclusion) In view of the above description, the present invention provides a resolution of text rendered on a display device having sub-pixel elements, such as an RGB LCD, in particular, a display device having horizontal striping. A technique for improving is disclosed.

The present invention has disclosed techniques for appropriately adjusting the metrics associated with character outline information (left side bearing, advance width, vertical character size 1, etc.). The present invention also disclosed techniques for preventing "jumping" or "bouncing" of the baseline.

The present invention also disclosed techniques for compressing red, green, and blue intensity values.

The present invention also disclosed a technique for selectively filtering color values when differences between adjacent sub-pixel elements would otherwise be difficult to see.

Finally, the present invention takes into account the gamma of the display device and for correcting the gamma of the pixel such that the intensity values of the sub-pixel elements are in the range of the intensity where the gamma correction is most useful. Techniques have been disclosed.

[Brief description of the drawings]

FIG. 1 shows a known arrangement of sub-pixel elements of an LCD display with horizontal striping.

FIG. 2 is a diagram showing the portion of FIG. 1 in more detail;

FIG. 3 illustrates a known image sampling operation.

FIG. 4 is a diagram showing a known method of representing character information.

FIG. 5A is a block diagram of a computer system that can be used to implement at least some aspects of the present invention.

FIG. 5B is a high-level block diagram of a machine that can be used to implement at least some aspects of the present invention.

FIG. 6 illustrates an image sampling technique with which the present invention can be used.

FIG. 7 is a high level process diagram of an environment in which at least some aspects of the present invention can operate.

FIG. 8 is a diagram of a graphic display interface process of an environment in which at least some aspects of the present invention can operate.

FIG. 9 illustrates letterpress printing terms used in describing some aspects of the present invention.

FIG. 10 shows a first for performing an overscaling or oversampling process.
3 is a high-level flowchart of the method.

11 is an example showing the operation of the method shown in FIG.

FIG. 12 illustrates a second method for performing an overscaling or oversampling process.
3 is a high-level flowchart of the method.

13 is an example showing the operation of the method shown in FIG.

FIG. 14 is a high-level flowchart for performing a hinting process.

FIG. 15A is an example showing the operation of the hinting method of FIG. 14;

FIG. 15B is an example showing the operation of the hinting method of FIG. 14;

FIG. 16 is a high level flowchart of a first method for performing a scan conversion process.

FIG. 17 is a high-level flowchart of a second method for performing a scan conversion process.

FIG. 18A illustrates the usefulness of the zero padding step in the scan conversion methods of FIGS. 16 and 17.

FIG. 18B illustrates the usefulness of the zero padding step in the scan conversion methods of FIGS. 16 and 17.

FIG. 19 is an example illustrating an exemplary scan conversion process.

FIG. 20A illustrates storage and retrieval of scan conversion source samples (or, more generally, information).

FIG. 20B illustrates storage and retrieval of scan conversion source samples (or, more generally, information).

FIG. 21 is an exemplary data structure that can be used to store glyph information in a glyph cache.

FIG. 22 is a high-level flowchart of an exemplary method for performing a display driver management process.

FIG. 23 is a high-level flowchart of an exemplary method for implementing a color compensation (or color filtering) process.

FIG. 24 is a high level flowchart of an exemplary method for implementing a gamma correction method.

──────────────────────────────────────────────────続 き Continuation of front page (81) Designated country EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE ), OA (BF, BJ, CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG), AP (GH, GM, KE, LS, MW, SD, SL, SZ, TZ, UG, ZW), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, CA, CH, CN, CU, CZ, DE, DK, EE, ES, FI, GB, GD, GE, GH, GM, HR, HU, ID, IL, IN, IS , JP, KE, KG, KP, KR, KZ, LC, LK, LR, LS, LT, LU, LV, MD, MG, MK, MN, MW, MX, NO, NZ, PL, PT, RO, RU, SD, SE, SG, SI, SK, SL, TJ, TM, TR, TT, UA, UG, UZ, VN, YU, ZW (72) Inventor Leroy B. Keely Jr. United States 94028 Portra Valley, California Gabada Way 210 (72) Gregory Sea. Hitchcock United States 98072-9236 Woodinville, Washington Northeast 159 Avenue 17828 (72) Inventor Ryan E. Kukiahman United States 98052 Redmond, Washington 156 Avenue North East 4850 Apartment 72 F-term (Reference) 5B080 FA00 FA02 FA05 GA00 5C080 AA10 BB05 CC03 DD01 DD07 EE01 EE29 EE30 JJ01 JJ02 JJ07 [Continued when filtering] can do. Finally, the gamma of the pixel value can be corrected so that the gamma of the display device is taken into account and the intensity values of the sub-pixel elements fall within the range of intensities for which gamma correction is more useful.

Claims (24)

[Claims] 1. A method for improving the resolution of text to be rendered on a display device having horizontally striped sub-pixel elements, comprising: a) determining the intensity value of a first sub-pixel element of a pixel Determining based on second information; and b) determining a luminosity value of a second sub-pixel element of the pixel based on second information, wherein the second information is determined from the first information. Determining the luminosity value of a first sub-pixel element of the pixel; and determining the luminosity value of a second sub-pixel element of the pixel, i) accepting textual information and display device resolution information Ii) accepting a resolution enhancement overscaling factor; and iii) displaying. Method characterized by including the sub-steps of scaling the character information based on the device resolution information and resolution enhancement over scaling factor. 2. The method of claim 1, further comprising: determining a light intensity value of a first sub-pixel element of the pixel; and determining a light intensity value of a second sub-pixel element of the pixel, the sub-step of accepting a point size. The method of claim 1, wherein the sub-step of scaling textual information is further based on point size. 3. The character information defines a vertical direction and a horizontal direction, and in the sub-step of scaling the character information, the character information is scaled to be larger in the vertical direction than in the horizontal direction. Item 7. The method according to Item 1. 4. The method of claim 1, wherein said resolution enhancing overscaling factor is at least three. 5. The method of claim 1, further comprising: determining a light intensity value of a first sub-pixel element of the pixel; and determining a light intensity value of a second sub-pixel element of the pixel. The method of claim 1, further comprising the sub-step of rotating the character information, which is performed before. 6. A method for improving the resolution of text to be rendered on a display device having horizontally striped sub-pixel elements, comprising: a) determining the intensity value of the first sub-pixel element of the pixel Determining based on second information, and b) determining a luminosity value of a second subpixel element of the pixel based on second information, wherein the second information is displaced from the first information. Determining the luminosity value of a first sub-pixel element of the pixel; and determining the luminosity value of a second sub-pixel element of the pixel, each comprising: i) (a) scanning conversion source samples. An oversampled and scaled character bitmap consisting of: Accepting one of the mapped character bitmaps; ii) accepting the overscaled glyph metric; and iii) determining the expected size glyph metric from the overscaled glyph metric. Iv) determining a luminosity value of a first sub-pixel element of a pixel from the first information, wherein the first information includes at least one scan conversion source sample; v) determining a light intensity value of a second sub-pixel element of the pixel from the second information, wherein the second information includes at least one other scan conversion source sample. And how. 7. The first information comprises a first group of scan conversion source samples, the second information comprises a second group of scan conversion source samples, and the first group of scan conversion source samples comprises a scan conversion source sample. 7. The method of claim 6, wherein the method is displaced from a second group of source samples. 8. The method of claim 7, wherein at least one of the first group of scan conversion source samples also belongs to a second group. 9. One of the first group of scan conversion source samples is adjacent to one of the second group of scan conversion source samples, and the first group and the second group are: The method of claim 7, wherein no common scan conversion source samples are included. 10. The glyph metric includes a left side bearing vertical component, and the sub-step of determining an expected size glyph metric from the over-scaled glyph metric is divided by an over-scaling factor. The method of claim 6, including determining a ceiling of the scaled left side bearing vertical component. 11. The glyph metric includes ascent and decent, and from the overscaled glyph metric, determining a glyph metric of an expected size comprises: (i) dividing by an overscaling factor. Determining a glyph size by taking the sum of the overscaled ascent ceiling and (ii) the absolute value of the overscaled decent ceiling divided by the overscaling factor. 7. The method of claim 6, wherein 12. A method for improving the resolution of text to be rendered on a display device having horizontally striped sub-pixel elements, comprising: a) determining the intensity value of the first sub-pixel element of the pixel Determining based on second information; and b) determining a luminosity value of a second sub-pixel element of the pixel based on second information, wherein the second information is determined from the first information. Determining the luminosity value of a first sub-pixel element of the pixel; and determining the luminosity value of a second sub-pixel element of the pixel, i) accepting textual information and display device resolution information Ii) accepting a resolution enhancement overscaling factor; and iii) displaying. Scaling the character information based on the device resolution information to generate scaled character information; and iv) oversampling the scaled character information based on a resolution enhancement overscaling factor. Features method. 13. The step of determining a light intensity value of a first sub-pixel element of the pixel and the step of determining a light intensity value of a second sub-pixel element of the pixel further include a sub-step of accepting a point size. The method of claim 12, wherein the sub-step of scaling textual information is further based on a point size. 14. The method of claim 14, wherein the character information defines a vertical direction and a horizontal direction, and in the sub-step of scaling the character information, the character information is sampled larger in the vertical direction than in the horizontal direction. Item 13. The method according to Item 12. 15. The resolution enhancing overscaling factor may be at least 3
13. The method according to claim 12, wherein 16. The method of claim 1, further comprising: determining a light intensity value of a first sub-pixel element of the pixel; and determining a light intensity value of a second sub-pixel element of the pixel. 13. The method according to claim 12, further comprising the sub-step of rotating the character information performed before. 17. A method for improving the resolution of text to be rendered on a display device having horizontally striped sub-pixel elements, comprising: a) determining the intensity value of the first sub-pixel element of the pixel 1) determining based on the information; b) determining a light intensity value of a second sub-pixel element of the pixel based on the second information; c) determining a light intensity value for each sub-pixel element; Ii) normalizing the intensity values of the sub-pixel elements based on the useful intensities of the intensity values to produce a normalized intensity value; iii) Accepting the gamma of the display device; iv) adjusting the normalized luminosity value of the sub-pixel element based on the gamma, Wherein the equipped and adjusting based on the gamma. 18. The method of claim 1, wherein the step of normalizing the intensity values of the sub-pixel elements comprises: clamping the intensity values of the sub-pixel elements such that each of the normalized intensity values is within a useful range of intensity values. 18. The method according to claim 17, comprising: 19. The sub-step of normalizing the luminosity value of the sub-pixel element, comprising: a) shifting the luminosity value of the sub-pixel element to generate a shifted luminosity value; and b) the shifted luminosity value. Clamping the resulting luminosity values such that each normalized luminosity value is within a useful range of intensity values.
7. The method according to 7. 20. The method of claim 20, wherein the sub-step of normalizing the light intensity value of the sub-pixel element comprises scaling the light intensity value such that the normalized light intensity value is within a useful range of intensities. The method of claim 17. 21. A method for improving the resolution of text to be rendered on a display device having horizontally striped sub-pixel elements, comprising: a) determining the intensity value of a first sub-pixel element of a pixel to a second one; Determining based on the second information, and b) determining the luminous intensity value of the second sub-pixel element of the pixel based on the second information, wherein the second information is displaced from the first information. Determining the luminosity value of a first sub-pixel element of the pixel; and determining the luminosity value of a second sub-pixel element of the pixel, each comprising: i) (a) scanning conversion source samples. An oversampled and scaled character bitmap consisting of: Accepting one of the scaled character bitmaps; and ii) accepting an overscaled metric, wherein the overscaled character bitmap defines a vertical distance away from the baseline of the character. Iii) substeps of determining the expected size integer and remainder values from the overscaled metric by dividing the overscaled metric by an oversampling factor; and iv) a non-zero remainder value. If accepts some scan conversion source samples that exceed the vertically overscaled character bitmap, but ignores its value; and v) the luminosity of the first subpixel element of the pixel from the first information Sub-step to determine the value Wherein the first information comprises at least one scan conversion source sample; and vi) determining a light intensity value of a second subpixel element of the pixel from the second information, wherein the second information comprises at least one. Including one of the other scan conversion source samples. 22. The number of scan conversion source samples to be accepted that exceed the vertically overscaled character bitmap and have a value to ignore is equal to the overscaling factor and less than the remainder value. The method according to claim 21, characterized in that: 23. The method of claim 21, wherein the overscaled metric has an overscaled ascent value. 24. The method of claim 21, wherein the overscaled metric has an overscaled decent value.