KR20130002780A - Ray tracing core and ray tracing method - Google Patents

Abstract

PURPOSE: A ray tracing core and a ray tracing method thereof are provided to generate a realistic scene by applying a ray tracing environment. CONSTITUTION: A start level determination unit(910) determines a start level of a inputted ray. A ray movement unit(930) reads an altitude from a cache(940). The ray movement unit performs ray movement computation. When the ray movement unit moves the ray to a corresponding level, a level selecting unit selects the next level. A crossing test unit(960) checks the reach of the ray to the lowest level and the generation of a cross spot between the ray and the altitude. [Reference numerals] (910) Start level determination unit; (920) Multiplexer; (930) Ray movement unit; (940) Altitude from a cache; (950) Ray selecting unit; (960) Crossing test unit; (AA) Ray; (BB) Cross spot

Description

RAY TRACING CORE AND RAY TRACING METHOD}

The disclosed technique relates to ray tracing technology, and more particularly, to a ray tracing core and a ray tracing method capable of accelerating displacement mapping.

3D graphics uses a mapping technique to increase realism. For example, mapping techniques include texture mapping, bump mapping, and displacement mapping. Bump mapping is implemented by perturbing the normal vectors of a surface by applying bumps and wrinkles to the surface of the object [2]. However, bump mapping can have the limitation that it cannot change the shape of the object. Displacement mapping shows the effect by actually moving the geometry by a height-field value. Displacement mapping is implemented by deforming the sample points in a direction perpendicular to the plane by the displacement obtained from the height map [15]. Recently, GPU (Graphics Processing Unit) performance has advanced dramatically, and GPU-based displacement mapping algorithms are continuously announced. However, displacement mapping is expensive and still requires a great cost.

 Parallax mapping [3], parallax occlusion mapping [16, 12], relief mapping [11], and pyramidal displacement mapping methods to speed up the displacement mapping [7]. Inverse displacement mapping [7] has been proposed, such as displacement mapping [9, 14, 17]. Inverse displacement mapping can find the intersection of a ray and a height field by projecting a ray or view vector onto a height-field without changing geometry data.

In Reference [15], state-of-the-art methods can be classified into approximation algorithms and accurate algorithms. The first algorithm is fast, but it can be difficult to find the correct intersection. The second algorithm may require pre-computed data for fast searching and accurate intersection calculation. An approximation algorithm, POM and relief mapping, finds intersections by using at least one of linear search or binary search. However, a trade-off occurs between the step spacing and the processing speed during the finding of the intersection. For example, the shorter the step spacing, the higher the accuracy of the intersection but the slower the processing speed. Moreover, missed intersections may occur at steep elevation fields and at steep surface angles.

As a solution to this problem, a per-pixel ray tracing technique using a displacement map based on the image pyramid [4] has been presented [9, 14, 17]. The pyramid displacement map is a mip-map image pyramid composed of several levels of sub-images. With the pyramid displacement map, you can quickly find the intersection of the ray and altitude fields, starting from the root level and gradually descending to the leaf level. Pyramid displacement mapping may be possible for sub-node visits through one level down and neighbor node visits through node crossing when searching for an image pyramid [9]. On the other hand, the reference [17] can reduce the iteration step for the search compared to the reference [9] by presenting a one level up technique to effectively perform empty space skipping.

The disclosed technique provides a ray tracing core that can reduce the search time of a pyramid displacement map.

The disclosed technique provides a ray tracing core that can produce realistic scenes.

Among the embodiments, the ray tracing core checks node crossing for a start level determiner that determines a start level based on a ray defined as a start point and an end point, and node crossing for the determined start level, so that node crossing occurs. And a level selector for performing a selective level up after performing node crossing.

In one embodiment, if no node crossing occurs, the level selector may perform multi-level down by a predefined multi-level based on the current level. In an embodiment, the ray tracing core may further include a cross test unit configured to perform a cross test after the multi level is down. In an exemplary embodiment, the ray tracing core may further include a ray moving unit configured to perform an altitude read operation and a ray movement operation when the ray reaches the lowest level in the crossover test and the intersection with the ray does not occur. .

In an example embodiment, the start level determiner may calculate a start level by performing an exclusive logic operation on the start point and the end point of the ray.

In one embodiment, when the node crossing occurs, the level selector may perform node crossing to visit neighboring nodes and selectively increase the level.

Among embodiments, a ray tracing method executed by a ray tracing core includes determining a start level based on a ray defined as a start point and an end point, checking node crossing for the determined start level, and If node crossing occurs, performing an optional level up after performing node crossing.

In one embodiment, the ray tracing method may further include performing a multi-level down by a predefined multi-level based on the current level if no node crossing occurs. In one embodiment, the ray tracing method may further comprise performing a cross test after the multi-level down. In an embodiment, the ray tracing method may further include performing an altitude read operation and a ray movement operation when the intersection point with the ray does not occur in the cross test.

In one embodiment, determining the start level based on a ray defined as a start point and an end point may include calculating a start level by performing an exclusive logic operation on the start point and the end point of the ray.

In one embodiment, if the node crossing occurs, performing the selective level up may include performing node crossing to visit the neighboring node and selectively increasing the level if the node crossing occurs.

Among the embodiments, the ray tracing core performs XOR bit operation on the X coordinate of the start point and the end point of the ray based on the ray defined as the start point and the end point, and at the same time the Y coordinate and the end point of the start point of the ray. An XOR bit operation is performed on the Y coordinate to select a larger number of the respective result values, and a start level determination unit for calculating a start level based on the selected result value, and node crossing for the determined start level. Checking includes performing a node crossing if node crossing occurs, visiting the neighboring node and selectively increasing the level, and if the node crossing does not occur, includes a level selector to down one or more levels.

The disclosed technique can reduce the search time of the pyramid displacement map from one configuration by means of solving the problem.

The disclosed technique can be applied to a ray tracing environment from one configuration by means of solving the problem to create a realistic scene.

1 is a diagram illustrating an example of a pyramid displacement map.
2 is a view for explaining an example of a search process of a pyramid displacement map.
3 is a view for explaining another example of the search process of the pyramid displacement map.
4 is a flowchart illustrating a ray tracing method of the present invention.
FIG. 5 is a diagram for describing a start level determining step of FIG. 4.
6 is a view for explaining the multi-level down step of FIG.
FIG. 7 is a diagram for describing an example of the optional level up process of FIG. 4.
FIG. 8 is a diagram illustrating an optional level up process of FIG. 4.
9 is a flowchart for explaining a ray tracing core.
FIG. 10 is a diagram for explaining a start level determining unit of FIG. 9.
FIG. 11 is a view for explaining a level selector in FIG. 9.
12 and 13 illustrate an image generated by a ray tracing method.
FIG. 14 is a table illustrating a node crossing occurrence probability for the kitchen images of FIG. 13.
15 is a table for explaining repetition steps for multi-level down.
FIG. 16 is a table for explaining a repeating step required for searching for an image pyramid.

The description of the disclosed technique is merely an example for structural or functional explanation and the scope of the disclosed technology should not be construed as being limited by the embodiments described in the text. That is, the embodiments may be variously modified and may have various forms, and thus the scope of the disclosed technology should be understood to include equivalents capable of realizing the technical idea. In addition, the objects or effects presented in the disclosed technology does not mean that a specific embodiment should include all or only such effects, and thus the scope of the disclosed technology should not be understood as being limited thereto.

Meanwhile, the meaning of the terms described in the present application should be understood as follows.

The terms "first "," second ", and the like are intended to distinguish one element from another, and the scope of the right should not be limited by these terms. For example, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

It is to be understood that when an element is referred to as being "connected" to another element, it may be directly connected to the other element, but there may be other elements in between. On the other hand, when an element is referred to as being "directly connected" to another element, it should be understood that there are no other elements in between. On the other hand, other expressions describing the relationship between the components, such as "between" and "immediately between" or "neighboring to" and "directly neighboring to", should be interpreted as well.

It should be understood that the singular " include "or" have "are to be construed as including a stated feature, number, step, operation, component, It is to be understood that the combination is intended to specify that it does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

In each step, the identification code (e.g., a, b, c, etc.) is used for convenience of explanation, the identification code does not describe the order of each step, Unless otherwise stated, it may occur differently from the stated order. That is, each step may occur in the same order as described, may be performed substantially concurrently, or may be performed in reverse order.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Commonly used predefined terms should be interpreted to be consistent with the meanings in the context of the related art and can not be interpreted as having ideal or overly formal meaning unless explicitly defined in the present application.

 1 is a diagram illustrating an example of a pyramid displacement map, FIG. 2 is a diagram illustrating an example of a search process of a pyramid displacement map, and FIG. 3 is a diagram illustrating another example of a search process of a pyramid displacement map.

The pyramidal displacement map is a quad-tree image pyramid generated through a computing process. The limb tree image pyramid is a set of hierarchical images from the 0 level with 2 ⁿ × 2 ⁿ pixels to the n level with 1 × 1 pixels.

Referring to FIG. 1, the original displacement map data may have a mip-map level of 0 and each node may represent a displacement value of an actual surface. Each node ( i , j ) in the high-level image has four nodes (2 i , 2 j ), ( 2i +1, 2j ), ( 2i , 2 j +1), (2 i +1, 2 j +1) can store the largest height-field value (height-field). The root node can store the maximum altitude value as a whole. The inner node can store the maximum altitude value locally [9, 15].

Referring to FIG. 2, the pyramid displacement mapping may search the image pyramid from root level n to leaf level 0 to find the correct intersection point between Ray and altitude field. Search for a Pyramids displacement mapping process is to first, the current texture pixel (texture) coordinates P coming to read the altitude fields d ₁ of the root level of the image pyramid from, the current position P of Ray and d ₁ meeting where P ₁ Move and reduce the image pyramid by one level. The next process reads the altitude field d ₂ of the current mipmap level from P ₁ , and if d ₂ is greater than d ₁ , moves P ₁ to the point P ₂ where ray and d ₂ meet and decrements the image pyramid by one level. . If it is not large, reduce the image pyramid by one level without moving the ray. You can repeat this process until you reach the leaf level.

When the ray is long when searching the image pyramid to cross the boundary of the current level node, the search method described with reference to FIG. 2 may lead to inaccurate results. Node crossing can check whether the position of the moved ray is at the same node as before the move, and if the current node is crossed, the position of the ray can be moved to the boundary of the node. As a result of the node crossing, Ray visits the neighboring node.

Referring to FIG. 3, reference document [17] proposes a search algorithm that can reduce the number of node crossings of [9] by leveling up an image pyramid. The level up method increases the mip-map level by one step when the ray exists where the cell boundary can be divided into two. Among the lines vertically up and down, the solid line represents the boundary of each node to the k level of the image pyramid, and the boundary of each node to the k -1 level is twice the k level, so both the dashed line and the solid line can correspond. . In reference [17], the node crossing can be raised to the k level if the node crossing occurs at the solid line position with respect to the k −1 level. For example, in reference [9], the node may be visited four times in the order of 1 → 2 → 3 → 4 for the k −1 level. In the case of reference [17], it is possible to visit C after ascending to k level at position A, then descending to k −1 level and visiting C. Therefore, the node may be visited three times in the order A → B → C.

4 is a flowchart illustrating a ray tracing method of the present invention.

Referring to FIG. 4, the ray tracing method may include a start level calculation step, a multi level down step, and a selective level up step.

The start level determination step may determine a start level for the input ray (step S401). In one embodiment, the ray tracing method may determine the start level based on the search method described in FIGS.

The ray tracing method reads the altitude information from the cache storing the altitude information with the ray information and the corresponding mip-map level value (step S402), and moves the ray to the point where the altitude information and the ray meet. (Step S403).

The ray tracing method may begin a search to find the intersection of the ray and the altitude field from the start level. The ray tracing method may determine whether node crossing occurs (step S404). If node crossing occurs, after performing node crossing (step S405), node crossing is performed through an optional level up process (step S406). If no node crossing occurs, it is possible to check whether the current level is greater than 2 (step S408). . If greater than 2, the multi-level down process is performed (step S409), and if not greater than 2, the level is decreased by one (step S410). A detailed description of the selective level up process and the multi level down process will be described later with reference to FIGS. 7 and 6, respectively.

The ray tracing method checks if the ray has reached the lowest level and there is an intersection (step S411), and if there is an intersection, performs an illumination process (step S412) with the color value and normal vector information of the intersection. This completes the whole process for one ray. If there is no intersection point, steps S402 to S411 are repeated.

5A and 5B are diagrams for describing a start level determination step of FIG. 4.

Existing search methods for image pyramids, such as references [9, 17], start searching from the top level. The ray tracing method of the present invention may calculate the minimum level at which no node crossing occurs for a given ray and set it as the start level. Mipmaps at levels higher than the start level may not be searched and may not require information on a separate altitude field. The ray tracing method can calculate the start level using only the start and end points of the ray.

Referring to FIG. 5A, it is assumed that a mipmap image has a mipmap level of 0, a ray start point P and an end point A. The coordinate of P is (409, 60) and the coordinate of A is (420, 50). Referring to FIG. 5B, a mipmap image of level 7 is shown for the case of FIG. 5A. The ray presented in FIG. 5A may be included in the node of coordinates (3, 0) in the case of FIG. 5B. This may mean that node crossing does not occur in mipmaps of level 7 or higher.

In the case of FIG. 5A, the search may be started by setting the start level to 7. Since the existing method starts searching at level 9, the new search method can reduce two iteration steps compared to the existing search method.

For example, suppose two points P and A are defined as P = (P _X , P _Y ) and A = (A _X , A _Y ), and the image size of mipmap level 0 is 512 × 512. P _X , P _{Y ,} A _X , and A _Y are each 9 bits. At this time, the position of _X on the image P of the mipmap level K corresponds to the value of the upper 9 bits of the P-K _X.

For example, in FIG. 5A, the position of P _X is 110011001 (2), and in FIG. 5B, the position of P _X is 11 (2). The same may be applied to the case of A _X. If the upper r bits of P _X and A _X are the same, P _X and A _X may exist in the same node up to the upper r level. The same applies to P _Y and A _Y , and the starting level can be calculated using this.

The method for calculating the start level is as follows. Assuming two points, P = (P _X , P _Y ) and A = (A _X , A _Y ), first perform an XOR (exclusive-or) operation on the bits for P _X and A _X , and simultaneously P Performs an XOR operation on bits _Y and A _Y. Compare the result of the first operation with the result of the second operation to determine the larger number. It may be determined whether there is a leading zero of the bit for the determined value, and if so, the prezero may be calculated. Subtracting the calculated value with respect to the total number of levels can finally calculate the position of the starting level.

For example, in FIG. 5A, the results of P _X XOR A _X and P _Y XOR A _Y are 000111101 (2) and 000001110 (2), respectively, and the larger of the two values is 000111101 (2). The leading zero of 000111101 (2) is three. If you subtract 3 from the total number of 10, the starting level is 7.

6 is a view for explaining the multi-level down step of FIG.

Referring to FIG. 6, the conventional discovery method visits a lower node through one level down and visits a neighbor node through node crossing. Conventional search methods can cause the frequency of level down to be significantly higher than node crossing. This is because there is coherence in the search of the image pyramid.

The ray tracing method may descend one or more levels of the image pyramid at a time through multi level down. Multi-level down may reduce the number of repetitive steps compared to sequential level down.

It is assumed that the ray tracing method uses two level down among the multi level down.

The multi-level down process determines whether node crossing occurs at the current level. If so, node crossing can be performed to visit neighboring nodes. If not, check if the current level is greater than 2. If it is greater than 2, level 2 may be performed. If not greater than 2, level 1 may be performed.

For example, suppose the start level is five. Sequential level down may begin the search from level 5 to level 0 sequentially. Multi-level down can start at level 5 and sequentially navigate through level 3, level 1, and level 0. Here, sequential level down repeats the search six times, and multi-level down repeats the search four times.

FIG. 7 is a diagram illustrating an example of a selective level up process of FIG. 4, and FIG. 8 is a diagram illustrating an optional level up process of FIG. 4.

Referring to FIG. 7, a search method for inverse displacement mapping is illustrated for an example of a search process presented in Ref. [17]. In reference [9], only a visit to a neighboring node is performed through node crossing, whereas in reference [17], an unconditional level up can be performed in node crossing at positions of Cells 7 and 11. have. Here the cell is the same as the lowest level node. Reference [17] can reduce the number of node crossings as compared with [9].

When the level up is performed in the cell 11 position, the user can level down to the left node. In this case, a repetition process may increase by 1 compared with visiting a neighbor node at cell 11. This means that leveling down to the left node after leveling up is disadvantageous in terms of repetitive staff compared to visiting neighboring nodes.

The ray tracing method can reduce the repetitive step compared to the reference [17]. The ray tracing method may select whether to level up at a position where level up is possible.

Referring to FIG. 8, if the position after node crossing is P, the ray tracing method may calculate the intersection point P ₁ through a linear equation with the altitude field value d of the ray and the current m-node position. The next step is to calculate the difference between the coordinates of the current node and the intersected node. If the calculated difference is greater than 1, you can choose level up. If it is not greater than 1, you can choose not to level up. In FIG. 8, since the coordinate difference between P and P ₁ is 1, no level up is selected.

9 is a flowchart for explaining a ray tracing core.

Referring to FIG. 9, the ray tracing core may include a start level determiner 910, a ray mover 930, a level selector 950, and a cross tester 960.

The start level determiner 910 may determine a start level of the input ray. The ray mover 930 may read altitude from a cache 940 that stores altitude information and perform a ray move operation. When the ray mover 930 moves the ray to the level, the level selector 950 selects the next level.

The intersection test unit 960 checks whether a point where the ray reaches the lowest level and where the ray and the altitude cross each other occurs. If it occurs, the search can be terminated and the intersection point can be output. If not, the altitude read operation, the ray shift operation, and the level selection process can be repeated sequentially for the next search.

FIG. 10 is a diagram for explaining a start level determiner of FIG. 9, and FIG. 11 is a diagram for explaining a level selector of FIG. 9.

Referring to FIG. 10, the start level determiner 910 performs an XOR operation on the x coordinate of the input start point and the x coordinate of the end point (step S1001), and simultaneously the y coordinate of the start point of the ray and the y of the end point. An XOR operation may be performed on the coordinates (step S1002). The start level determiner 910 may select a large number for two result values through a multiplexer Mux (step S1003), and output a calculated value by calculating a leading zero with respect to the selected value. (Step S1004).

Referring to FIG. 11, the level selector 950 may perform node crossing with the x and y coordinate values of the ray's elevation before moving and the x and y coordinate values of the moved ray's elevation (see FIG. 11). Step S1101). The level selector 950 may perform subtraction to perform selective level up at the same time (step S1102). If no node crossing occurs, multi-level down is performed (step S1103), and if it occurs, node leveling may be performed (step S1104) and then optional level up may be performed (step S1105). The level selector 950 may selectively output a level value (step S1106).

12 and 13 illustrate an image generated by a ray tracing method.

12A shows the displacement map for marble and FIG. 12B shows the displacement map for brick. 13A and 13B show a kitchen image to which a displacement map for marble is applied, and FIGS. 13C and 13D show a kitchen image to which a displacement map for brick is applied.

Hardware environment is Intel Core2 Duo E8400 3.20GHz, 4GB DDR2 SDRAM, Nvidia 9600GT. For performance evaluation criteria, kitchen was used in BART [8], which is widely used for performance evaluation of ray tracing animation. Kitchen images have many reflections and refractions, and are made up of more than 110,000 triangles and textures. The experimental environment was constructed by additionally implementing the method proposed based on ray tracing environment in Ref. [10]. Up to 10 secondary rays due to reflection and refraction can be generated during ray tracing.

FIG. 14 is a table illustrating a node crossing occurrence probability for the kitchen images of FIG. 13.

14 shows the probability of node crossing when searching for an image pyramid with reference [9] and reference [17], respectively. Experimental results showed that the probability of node crossing was about 24.5% in Ref. [9] and 20.6% in Ref. [17]. As low. This may mean that the frequency of level down in the search is high. During the ray tracing process, the start level determination step and the multi-level down step may use a high level down frequency during the search. In addition, since the reference [17] includes a level up process, the frequency of node crossing may be lower than that of the reference [9].

15 is a table for explaining repetition steps for multi-level down.

FIG. 15 shows, for each level, a repeating step for searching for an image pyramid due to n level down. The image selects scene 4 and assumes that the start level determination step and the optional level up step do not apply. One level is the same as that of Ref. [9] because it is sequentially lowered. In the case of marble and brick, the minimum repetition staff is 10 because the image pyramid consists of 10 levels in total. Experimental results show the best performance when level 3 is down, and there are more repetitive steps for level 2 down than level 3 down.

FIG. 16 is a table for explaining a repeating step required for searching for an image pyramid.

Referring to FIG. 16, the repetitive steps required for searching the image pyramid were compared with the existing tracing process and the ray tracing process of the present invention. For marble, up to 21%, 17%, and 5% repetitive steps were reduced for the start level determination, 2 level down, and optional level up steps, respectively. As a result of applying all the search processes, the number of repetitive staff was reduced by up to 35% in scene 4.

In the case of bricks, the number of repetitive staff decreased by up to 22%, 27%, and 5% for the start level determination stage, the 2 level down stage, and the optional level up stage, respectively. As a result of applying all the search processes, the number of repetitive staff was reduced by up to 38% in scene 4.

For both displacement maps, the reduction rate of the repeat steps was measured highest in scene 4. This is because the effect of the 2 level down is applied because of the lowest node crossing ratio in scene 4. On the other hand, in scene 3, the node tracing ratio is the highest, so the effect of ray tracing is the lowest.

Although described above with reference to a preferred embodiment of the present application, those skilled in the art various modifications and changes to the present application without departing from the spirit and scope of the present application described in the claims below I can understand that you can.

[Reference literature]

[1] Baboud, Lionel and Decoret, Xavier. Rendering geometry with relief textures. Proceedings of Graphics Interface, pages 195--201, 2006.

[2] James F. Blinn. Simulation of wrinkled surfaces. Proceedings of the ACM annual conference on Computer graphics and interactive techniques, pages 286--292, 1978.

: Advanced Rendering with DirextX and OpenGL. Cengage Learning, 2005. [3] Zoe Brawley and Natalya Tatarchu. Shader

Advanced Rendering with DirextX and OpenGL. Cengage Learning, 2005.

[4] Daniel Cohen and Amit Shaked. Photo-realistic Imaging of Digital Terrain. Computer Graphics Forum, 12 (3): 363-373, 1993.

[5] William Donnelly. GPU Gems 2, chapter 8, pages 123-136. Addison Weasley, 2005.

[6] Stefan Jeschke and Stephan Mantler and Michael Wimmer. Interactive Smooth and Curved Shell Mapping. Proceedings of the Eurographics Symposium on Rendering, pages 351--360, 2007.

[7] Tomomichi Kaneko and Toshiyuki Takahei and Masahiko Inami and Naoki Kawakami and Yasuyuki Yanagida and Taro Maeda and Susumu Tachi. Detailed shape representation with parallax mapping. Proceedings of the International Conference on Artificial Reality and Telexistence, pages 205--208, 2001.

[8] Jonas Lext and Ulf Assarsson and Tomas Moller. A Benchmark for Animated Ray Tracing. IEEE Computer Graphics and Applications, 21: 22--31, 2001.

[9] Kyoungsu Oh and Hyunwoo Ki and Cheol-Hi Lee. Pyramidal Displacement Mapping: a GPU based Artifacts-Free Ray Tracing through an Image Pyramid. Proceedings of the ACM symposium on Virtual reality software and technology, pages 75--82, 2006.

[10] Woochan Park and Jaeho Nah and Jeongsoo Park and Kyungho Lee and Dongseok Kim and Sangduk Kim and Jinhong Park and Yoonsig Kang and Sungbong Yang and Tackdon Han. An FPGA Implementation of Whitted-style Ray Tracing Accelerator. Proceedings of the IEEE Symposium on Interactive Ray Tracing, pages 187, 2008.

[11] John W. Patterson and S. G. Hoggar and J. R. Logie. Inverse Displacement Mapping. Computer Graphics Forum, 10 (2): 129-139, 1991.

[12] Policarpo, Fabio and Oliveira, Manuel M. and Comba, Joao L. D. Real-time relief mapping on arbitrary polygonal surfaces. Proceedings of the ACM Symposium on Interactive 3D Graphics and Games, pages 155--162, 2005.

[13] Fabio Policarpo and Manuel M. Oliveira. GPU Gems 3, chapter 18, pages 409-428. Addison Weasley, 2008.

[14] Schroders, M. F. A. and Gulik, R. v. Quadtree relief mapping. Proceedings of the ACM SIGGRAPH / EUROGRAPHICS Symposium on Graphics hardware, pages 61--66, 2006.

[15] Szirmay-Kalos, L. and Umenhoffer, T. Displacement Mapping on the GPU-State of the Art. Computer Graphics Forum, 27 (1), 2008.

[16] Natalya Tatarchuk. Dynamic parallax occlusion mapping with approximate soft shadows. Proceedings of the ACM Symposium on Interactive 3D Graphics and Games, pages 63--69, 2006.

[17] Art Tevs and Ivo Ihrke and Hans-Peter Seidel. Maximum Mipmaps for Fast, Accurate, and Scalable Dynamic Height Field Rendering. Proceedings of the ACM Symposium on Interactive 3D Graphics and Games, pages 183--190, 2008.

Claims (13)

A start level determiner configured to determine a start level based on a ray defined as a start point and an end point;
And a level selector configured to check node crossing with respect to the determined start level, and perform a selective level up after performing node crossing if node crossing occurs. The method of claim 1, wherein the level selector
And if no node crossing occurs, performing a multi-level down by a predefined multi-level based on the current level. The method of claim 2,
And a cross test unit configured to perform a cross test after the multi-level down. The method of claim 3,
And a ray moving unit configured to perform an altitude read operation and a ray movement operation when the ray reaches the lowest level in the crossover test and the intersection point with the ray does not occur. The method of claim 1, wherein the start level determiner
And a start level is calculated by performing an exclusive logic operation on the start point and the end point of the ray. The method of claim 1, wherein the level selector
And if the node crossing occurs, perform node crossing to visit neighboring nodes and selectively increase the level. Determining a start level based on a ray defined by a start point and an end point;
Checking node crossing for the determined start level; And
If the node crossing occurs, performing the level crossing after performing the node crossing. The method of claim 7, wherein the ray tracing method
And if no node crossing occurs, performing multi-level down by a predefined multi-level based on the current level. The method of claim 8, wherein the ray tracing method is
And performing a cross test after the multi level down. The method of claim 9, wherein the ray tracing method is
And performing an altitude read operation and a ray movement operation when the ray reaches the lowest level in the crossover test and no intersection with the ray occurs. 8. The method of claim 7, wherein determining a start level based on a ray defined by a start point and an end point
And calculating a start level by performing an exclusive logic operation on the start point and the end point of the ray. 8. The method of claim 7, wherein the step of performing a selective level up if node crossing occurs.
Performing node crossing to visit neighboring nodes and selectively increasing the level when the node crossing occurs. XOR bit operation is performed on the X coordinate of the start point and the end point of the ray based on the ray defined as the start point and the end point, and at the same time, the XOR bit operation is performed on the Y coordinate of the start point and the Y coordinate of the end point of the ray. A start level determiner which selects a larger number of respective result values and calculates a start level based on the selected result value; And
Check Node Crossing for the determined start level, and if node crossing occurs, perform node crossing to visit neighboring nodes and selectively increase the level, and if node crossing does not occur, increase one or more levels. A ray tracing core comprising a level selector for down.

Images