RU2405205C1 - Method for generation of sprite - Google Patents

Abstract

FIELD: information technologies.

SUBSTANCE: sprite represents images of irregular shape overlaid over background images, preferably online. Generated sprite has specified first multitude of possible values of non-transparent colours of pixels and at least one value of pixel colours identifying transparent colour of pixel. Sprite is generated using initial data, including the first data structure specifying colour image with specified second multitude of possible colours of pixels, and the second data structure specifying non-transparent part of specified sprite. Compliance of colours of the first data structure and possible non-transparent colours of sprite is partial projection of the specified second multitude into the specified first multitude, where augmentation of area of identification of specified partial projection to the specified second multitude is the multitude, capacity of which is considerably lower than capacity of the specified second capacity, besides to identify the value of colours included into the specified first multitude and complying with colour of pixel from the first data structure, the following actions are taken: if pixel of the first data structure has colour that does not belong to the specified area of identification, then value of this pixel is established as equal to colour being one of nearest to initial colour of pixel and included in the specified area of identification; specified partial projection is applied to colour of pixel of the first data structure.

EFFECT: improving sprite generation reliability.

5 cl, 2 dwg

Description

The invention relates to methods for creating sprites that are irregularly shaped images and superimposed on background images, preferably in real time. The invention can be used in the formation of overlay images with a given color key in real time.

In [http://www.xserver.ru/computer/langprogr/razn/36/10.shtml] flat raster images of irregular shape are called sprites. They are called decades in the DirectX 2 SDK documentation. The sprite can be imagined as a figure cut out of cardboard and installed at a certain point in the layout.

In [http://www.delphisources.ru/pages/faq/base/sprites_work.html] with reference to the book by Andre La Mota, sprites are called "small objects that are on the playing field and can move. This term took root with the light hand of Atari and Apple programmers in the mid-70s. Sprites are characters in PC games that can easily move around the screen, change color and size. ” In other words, a sprite is a character in a game. We can say that a sprite is an array of flowers; for simplicity, imagine it as a BMP file or TBitmap, especially since this format is supported by Windows and does not contain compression. What we need: from the sprite - make it appear on the screen and form an animation. Animation is not only a change in the coordinate of a sprite, but also a change in the image itself. Therefore, a sprite can have not one image, but several. Changing them leads to animation.

In [http://durus.ru/content/view/292/59/] a sprite refers to an image in a rectangular area filled with a color key, where the color key is one or more colors intended for the image of a "transparent" color. Pixels painted with this color are not copied when blitting. This allows you to create the illusion of a non-rectangular image, although only rectangular areas are copied when blitting.

In [http: /www.mobilab.ru/artictes/53/] a sprite is a partially transparent image or a set of images that can be superimposed and moved in the background or in another sprite without the need to redraw the entire main window. In other words, a sprite is a roaming game object.

In [http://silencer.times.lv/tools_articles_1.htm] sprites are essentially the same as bitmaps, but with one distinctive feature - they have areas with a transparent color, which is simply ignored during rendering and not is displayed. The term "sprite" is still sometimes used to indicate an animated bitmap. Modern sprites usually consist of several frames of animation, this is done in order, for example, to achieve the effect of a flying ship, a walking person, etc.

In [http://computers.plib.ru/programming/SVGA/Glava%203/Index13.htm] sprites are called patterns for which the mask is formed dynamically, depending on the values of the code points of the picture, but in any case, the mask has the same purpose . Regardless of the shape of the drawing itself, for example, a triangle, an arrow, an hourglass, etc., its image is always supplemented to a rectangle so that all lines have the same size. This greatly simplifies the storage and reproduction of drawings and at the same time forces them to apply masking, which excludes the display on the screen of that part of the rectangular region that is not related to the picture. By applying a mask, for example, the image of the cursor arrow is shown on the screen, and not the black rectangle against which it is drawn.

In accordance with [http://www.255.ru/index.php?newsid=636], a sprite is a small image that moves freely around the monitor. In the original sense of the word, this term was applied only to hardware-output images. Actually, only one real sprite can be found on an IBM PC - a hardware mouse cursor. With the architecture of × 86, it is customary to understand a sprite as a programmatically displayed image that can have a complex shape and move over the background without overwriting it.

In US Pat. No. 6,337,701, sprites are called graphic objects that move across the display screen in front of any other objects.

In the case when a 4-byte representation of the colors of the pixels is used, one of the bytes is allocated to the so-called alpha channel, indicating the degree of transparency of the pixel. In this case, there are no problems with creating sprites. However, color representation formats that do not contain an alpha channel are widely used. In this case, for coding transparency regions, colors specially allocated for this purpose are used, the combination of which determines the color key.

In [http://www.xserver.ru/computer/langprogr/razn/36/10.shtml] a method for creating a sprite is proposed, in which all extraneous details are deleted to leave a clean image of the figure, and the surrounding background is filled with black (transparent) color, and if the original image contains black areas, replace the black color in the image with a different shade (for example, instead of RGB: 0, 0, 0, you can use RGB: 0, 0, 1). Then they use Microsoft Imager to reduce the number of colors from 256 to 8. The disadvantage of this method is that it is intended only for creating sprites in a palette display mode, and also that it is not intended for use in real time.

In [http://durus.ru/content/view/292/59/] a method for creating a sprite is proposed, in which an image limited to a rectangular area is created, and all the pixels of this rectangular area that are not related to the image are painted with a special color, which was not used in the image itself, and this particular color is assigned as the color key. During blitting, it will not be copied, and instead of pixels colored with this color, background pixels will remain - an illusion of transparency will appear. The disadvantage of this method is that determining the color that was not used in the image itself is a very time-consuming operation, and therefore it is very problematic to use this method to create sprites in real time, especially on low-power computers.

US Pat. No. 6,337,701 (prototype) proposes a method for generating a sprite with a given first set of possible pixel colors from a first data structure that defines a color image with a given second set of possible pixel colors and a second data structure that determines the shape of the specified sprite. To obtain the value of each pixel of the sprite, it is determined by the second data structure whether the specified pixel is included in the opaque part of the specified sprite, and if so, then set the color value of the current sprite pixel to the value from the specified first set of colors, corresponding to the color of the pixel from the first data structure, and if not, set the color value of the current pixel of the sprite to a value that defines the transparent color of the pixel. The determination of whether a pixel is included in the opaque part of the specified sprite is made by determining whether the corresponding element of the second data structure has a predetermined value. In this case, the color cursor is considered as a sprite, and the formation and output of the sprite is performed only if the number of different colors in the structure of the color data of the image does not exceed the number of available colors in the hardware sprite. The disadvantage of this method is its low reliability, due to the fact that the output of the sprite occurs only when the number of different colors in the structure of the color data of the image does not exceed the number of available colors in the hardware sprite. This is a consequence of the fact that in the prototype method, no attempt is made to adjust the colors in the original image to satisfy the hardware limitations on the number of available colors.

In view of the foregoing, it is urgent to develop a method for generating a sprite having a predetermined first set of possible values of opaque pixel colors and at least one predetermined pixel color value defining a transparent pixel color, in which the sprite is generated from the source data including the first data structure, defining a color image with a given second set of possible pixel colors, and a second data structure defining an opaque part of the specified sprite. In this case, the method should ensure that the sprite is generated even when the number of different colors in the structure of the color image data exceeds the number of available colors due to hardware limitations so as not to lead to a situation where it is impossible to create a sprite.

This is achieved due to the fact that to obtain the values of each pixel of the sprite, it is determined from the second data structure whether the specified pixel is included in the opaque part of the specified sprite, and if so, then set the color value of the current sprite pixel to the value from the specified first set of colors corresponding to the color pixels from the first data structure, and if not, then set the color value of the current sprite pixel to a value that defines the transparent color of the pixel. Moreover, unlike the prototype, the correspondence of the colors of the first data structure and the possible opaque colors of the sprite is a partial mapping of the specified second set to the specified first set. Supplementing the definition domain of said partial mapping to said second set is a set whose power is substantially less than the power of said second set. To determine the value included in the specified first set of colors and corresponding to the color of the pixel from the first data structure, perform the following steps:

if the pixel of the first data structure has a color that does not belong to the specified definition area, then the value of this pixel is set to a color that is one of the pixels closest to the original color and is included in the specified definition area;

the specified partial mapping is applied to the pixel color of the first data structure.

By correcting the pixel values of the first data structure and using a partial mapping of the second set to the specified first set, i.e. the set of all colors of the image into the set of acceptable opaque colors of the sprite, the main technical result is achieved, which consists in increasing the reliability of the method of generating the sprite, because, unlike the prototype, the generation of the sprite is possible in any case. In this case, a slight decrease in the quality of the sprite is possible due to the distortion of the values of some colors of the original image. However, in most cases, the partial display definition area is chosen so that the maximum distance from the points of addition to the definition area is very small, and thus the specified distortion is almost imperceptible.

The invention is illustrated by drawings, where

Figure 1. Formation of a sprite using two input data structures.

Figure 2. Partial mapping structure φ ₀ .

The implementation of the invention

1. General description

The main requirement for the generated sprite is that the colors of the sprite pixels must belong to a given first set of possible values of opaque pixel colors, from which the values defining the transparent color of the pixels are excluded. Moreover, both the indicated first set of possible values of the opaque pixel colors and the values defining the transparent color of the pixels are a subset of the space of possible hardware-implemented colors. When overlaying a sprite on a background image, the computer graphics system will replace the pixel values of the background image with the corresponding pixel values of the sprite, if they correspond to opaque pixel colors, and leave the pixel values of the background image unchanged if the corresponding sprite pixels have values that define the transparent color of the pixel. It is very important that no sprite pixel inside the area corresponding to the opaque part has a value corresponding to a transparent color. Otherwise, in such pixels, instead of the color of the sprite image, the observer will see the background color on which the sprite is superimposed, and the sprite itself will appear to have a “hole”.

When a sprite is displayed on the computer screen as an overlay, many transparent colors are set with a color key. For the overlay surface, set the color key of the source, which sets the range of transparent colors. According to [http://www.bringyou.to/games/ddraw.PDF, p.36], the color key is used when calling the IDirectDrawSurface :: UpdateOverlay method to control which part of the overlay surface will be visible above the primary surface and which part of the overlay surface will be transparent. Overlay color keys can be associated with either a source or a receiver, which is set using the IDirectDrawSurface :: SetColorKey method. When displaying sprites in the form of an overlay, the source color key is used, for which the DDCKEY_SRCOVERLAY flag is set in the first dwFlags parameter of this method. The IDirectDrawSurface :: UpdateOverlay method uses the source color key to determine which sprite pixels on the overlay surface should be considered transparent, making it possible to see the primary surface through them. The formation of the color key can be done in different ways, which are discussed in paragraph 2.

The initial data for generating the sprite includes two data structures, where the first data structure defines a color image with a given second set of possible pixel colors, and the second data structure defines the opaque part of the specified sprite (see FIG. 1).

As the first data structure that defines a color image with a given second set of possible pixel colors, any image in one of the standard formats, such as BMP, TIFF, etc., can be used. At the same time, the space V of possible pixel colors of such an initial image can both coincide with the space W of possible hardware-implemented sprite colors and be different from it. In any case, when blitting the original image into the sprite surface, the hardware converts the colors of the original image into hardware-implemented colors. This transformation defines a map φ: V → W. The space of possible opaque sprite colors is a subset of W ₀ ⊂ W of the set of all hardware-implemented colors, from which all colors that define the transparent color of the sprite pixels are removed. These colors are usually determined by the value of the color key and constitute the box K in the space W of all hardware-implemented colors. Thus, there is a partial mapping φ ₀ : V → W ₀ having a domain of definition φ ₀ ^-1 (W ₀ ) = φ ^-1 (W ₀ ) = V / φ ^-1 (K) (see Fig. 2). Thus, the correspondence of the colors of the first data structure and the possible opaque colors of the sprite is a partial mapping of the specified second set to the specified first set.

The complement φ ^-1 (K) of the domain of the indicated partial mapping to the specified second set is a set whose power is substantially less than the power of the specified second set. Indeed, as will be seen from Section 2, the power of the set K is significantly less than the power of the set W of all hardware-implemented colors. In addition, in real situations when blitting an original image into a sprite, the number of colors in the original image that can turn into a single sprite color is very small, otherwise the image quality will inevitably be corrupted when blitting. It follows that the cardinality of the set φ ^-1 (K) is significantly less than the cardinality of the specified second set, and for each point from the set φ ^-1 (K) there is a point close to it from the domain V / φ ^-1 (K) (see figure 2).

There are several options for the second data structure that defines the opaque part of the generated sprite. In any case, such a data structure should allow for each pixel of the generated sprite to determine whether this pixel should be part of the opaque part of the sprite and have a color that matches the pixel color of the original image, or should have a transparent color. At least three types of data structures are widely used for defining an opaque part of a sprite: a bit mask, a set of horizontal segments of pixels, and vector defining the border of an opaque part of a sprite.

Sprite generation is as follows. First of all, for each color C from the set φ ^-1 (K), the color C 'replacing it is determined, which is the closest to the color C from all the colors not included in the set φ ^-1 (K). After that, make a list of all such pairs (C, C ') and arrange the set of all such pairs in ascending order C. These actions can be performed once before starting an application using sprite generation. If necessary, build a sprite from the above source data to obtain the value of each pixel of the sprite is determined by the second data structure, whether the specified pixel is included in the opaque part of the specified sprite. If the pixel is included in the opaque part of the sprite, then check whether the pixel color matches one of the colors C from the existing list of pairs (C, C '), and if so, then replace the color of the pixel C with the corresponding color C'. Otherwise, no pixel color change is performed. After that, set the color value of the current pixel of the sprite equal to the value included in the specified first set of colors and the corresponding color of the pixel from the first data structure, using the mapping φ. If the pixel is not included in the opaque part of the sprite, then set the color value of the current pixel of the sprite to a value that defines the transparent color of the pixel.

2. Formation of a color key

It is most convenient to form the overlay color key for sprite output so that the value of the upper border of the color key range coincides with the value of the lower border of the color key range. Thus, the set of colors belonging to the color key color range will consist of only one color, called the key color. The choice of key color depends on the color depth of the pixels. Most often, a fixed color is chosen as the key color, which is least likely to be found in real images. Acceptable are, for example, black [http://progma.narod.ru/directdraw.html], blue [http://jack.kiev.ua/linuxjournal/LJ/0081/4401.html], red-blue color [http://mini-stalker.googlecode.com/svn-history/r29/trunk/Mini_STALKER/animation.h], fuchsin [http://www.gamasutra.com/features/20010629/geczy_02.htm].

Sometimes a situation occurs when not one color is used as a color key, but many colors representing a parallelepiped in the color space. For example, in the well-known blue screen technology, in order to form a sprite, a real scene, for example, an announcer of a television program, is shot against a blue backdrop, after which only that part of the image that corresponds to the real scene is cut out. In this case, it is convenient to use the minimum box in the color space as the color key, which contains the colors of all image pixels corresponding to the backdrop.

In any case, the set of colors included in the color key has a power significantly less than the power of all possible colors of the image.

If the original image specified by the first data structure has a color depth that matches the color depth of the sprites, then the key color value can be chosen so that when converting the pixel color format of the source image, at most one pixel color of any such source image when converted takes the key color value . In fact, we consider the mapping φ of the color space of the original image into the color space of sprites. Since these two color spaces have the same color depth, the powers of the two color spaces coincide. Therefore, either the map φ is bijective or not surjective. In the first case, at any choice of the key color value when converting the pixel color format of the source image, at most one pixel color of any such source image during conversion can take the key color value. In the second case, in the color space of the overlay surface, there is a color value that lies outside the image of the mapping φ. If this value is taken as the key color, then when converting the pixel color format of the source image, no pixel color of any such source image will be able to take the key color value when converting. The way to implement this choice of key color is obvious: it is enough to convert an array of all the colors in the color space of the source images into the color space of the sprites, arrange the obtained values in ascending order and select any value that is missing in the resulting color array as the key color. If there are no missing values, then the mapping φ is bijective, and any color can be taken as the key color.

Claims (5)

1. A method of generating a sprite having a predetermined first set of possible values of opaque pixel colors and at least one predetermined pixel color value defining a transparent pixel color, in which a sprite is generated from source data including a first data structure defining a color image with a given the second set of possible pixel colors, and a second data structure that defines the opaque part of the specified sprite, in which to obtain the values of each pixel of the sprite is determined by W In the next data structure, does the specified pixel enter the opaque part of the specified sprite, and if so, set the color value of the current pixel of the sprite to the value from the specified first set of colors, the corresponding color of the pixel from the first data structure, and if not, set the color value of the current pixel sprite equal to the value that defines the transparent color of the pixel, characterized in that the correspondence of the colors of the first data structure and the possible opaque colors of the sprite is a partial display of of said second set to said first set, where the addition of the definition region of said partial display to said second set is a set whose power is substantially less than the power of said second set, and for determining a value included in said first set of colors and corresponding to a pixel color from the first data structure , perform the following steps: if the pixel of the first data structure has a color that does not belong to the specified definition area, then the value of the pixel is set equal to the color, which is one of the closest to the original color of the pixel and included in the specified domain; the specified partial mapping is applied to the pixel color of the first data structure. 2. The method according to claim 1, characterized in that the set of transparent colors consists of only one element. 3. The method according to claim 1, characterized in that the set of transparent colors is a parallelepiped in the color space of the sprite. 4. The method according to claim 1, characterized in that the addition of the definition area of the specified partial display to the specified second set is a prototype of the color key during the blitting operation of the specified first data structure into the sprite space. 5. The method according to claim 1, characterized in that the addition of the definition area of the specified partial display to the specified second set is a set of at most one element.

Images