Lecture 5. Scan Conversion Textures hw2

Transcription

1 Lecture 5 Scan Conversion Textures hw2

2 Announcements Homework deadlines Mini Project Proposals due June 26th

3 E.T. 06 Vector Graphics Algebraic equations describe shapes. Can render type and large areas of color with relatively small file sizes Can be reduced/enlarged with no loss of quality Cannot show continuous tone images (photographs, color blends) Examples: Plotters, Oscilloscopes, Illustrator, Flash

4 Raster Graphics aka. Pixel-based / Bit-mapped graphics Grid of pixels Size of grid = resolution of image Best for large photographic images Poor scalability : zoom in - loss of quality (jagged look) Modification at pixel level - texture mapping, blending, alpha channels, antialiasing, etc. Examples : E.T. 06 CRT, LCD, Dot-matrix printers Adobe Photoshop

5 Rasterization = Scan Conversion = Converting a continuous object such as a line or a circle into discrete pixels. E.T. 06

6 Scan Conversion of Lines

7 Scan conversion of lines Given two points with integer coordinates p 1 =[x 1, y 1 ], and p 2 =[x 2, y 2 ] the algorithm has to find a sequence of pixels approximating the line. Slope: (y 2 -y 1 )/(x 2 - x 1 ) We can always reorder p 1 and p 2 so that x 2 - x 1 is nonnegative. It is convenient to look at only nonnegative slopes; if the slope is negative, change the sign of y. Bunch of simplifications!

8 E.T. 06 Result: slope positive, and x2>x1 +y positive slope, x 2<x1 (x1,y1) (x2,y2) +x (x2,y2) negative slope (x1,y1) (x1,y1) y=0 (x2,y2) y=0 (x1,-y1) (x2,-y2)

9 Slope Slope reduction: it is convenient to have the slope of the line between 0 and 1; then we are able to step along x axis. slope > 1, cannot step along x To handle slope > 1, swap x and y slope < 1, can step along x

10 DDA Assume that the slope is between 0 and 1 Simplest algorithm (pompously called differential digital analyzer): Step along x, increment y by slope at each step. Round y to nearest pixel. float y = y1; float slope = (y2-y1)/(float)(x2-x1); int x; for(x = x1; x <= x2; x++) { } drawpixel(x, floor(y)); y += slope;

11 Bresenham Algorithm What is wrong with DDA? It requires floating-point operations. These operations are expensive to implement in hardware. They are not really necessary if the endpoints are integers. Idea: instead of incrementing y and rounding it at each step, decide if we just go to the right, or to the right and up using only integer quantities.

12 Increment decision pixel corners on different sides of the line increment both x and y pixel corners on the same side of the line increment only x Need: fast way to determine on which side of a line a point is.

13 Half-plane test Implicit equation can be used to perform the test. ( n (q # p)) " 0 (n (q # p)) $ 0 t he point on the same side with the normal n the point on the other side n q p p q

14 Implicit line equation The implicit equation of the line through p 1 =[x 1, y 1 ], and p 2 =[x 2, y 2 ] is (n,q-p 1 ) = 0, with n = [y 2, -y 1, x 1, -x 2 ] We need to test on which side of the line is the point q+d 1 = [x,y] +[1/2,1/2] d 1 [x,y] To do this, we need to determine the sign of F = (n,2q+2d 1-2p 1 ) Note that multiplication by two makes everything integer again! Key idea: compute this quantity incrementally.

15 Incremental computation At each step q = [x,y] changes either to [x+1,y] (step to the right) or to [x+1,y+1] (step to the right and up ); in vector form, the new value of q is either q +D 1 or q+d 2, with D 1 =[1,0] and D 2 =[1,1] Fnext = (n,2q+2d + 2d 1-2p 1 ) = (n,2q+ 2d 1-2p 1 ) + 2(n,D) = F + 2(n,D), where D is D 1 or D 2 At each step, to get new F we have to increment old F either by (n,d 1 ) or (n,d 2 ) (n,d 1 ) = y 2 -y 1 (n,d 2 ) = (y 2 -y 1 ) - (x 2 -x 1 )

16 Bresenham algorithm Assume the slope to be between 0 and 1. int y = y1; int dy = y2-y1; int dxdy = y2-y1+x1-x2; int F = y2-y1+x1-x2; int x; for( x = x1; x < =x2; x++ ) { drawpixel(x,y); if( F < 0 ) { F += dy; } else { y++; F+= dxdy; } }

17 Bresenham algorithm In For your an implementation you need to handle all slopes! First, reorder endpoints so that x 1, <= x 2 Then consider 4 cases: y 2, -y 1 >= 0, x 2, -x 1 >= y 2, -y 1 positive slope <= 1 y 2, -y 1 >= 0, x 2, -x 1 < y 2, -y 1 positive slope > 1 y 2, -y 1 < 0, x 2, -x 1 >= y 1, -y 2 negative slope >= -1 y 2, -y 1 < 0, x 2, -x 1 < y 1, -y 2 negative slope < -1 In each case, make appropriate substitutions in the algorithm.

18 Scan converting polygons

19 Polygons convex with self-intersections non-convex with holes We focus on the convex case

20 Scan Conversion of Convex Polygons General idea: decompose polygon into tiles scan convert each tile, moving along one edge

21 E.T. 06 Scan Convert a Convex Polygon void ScanY(Vertex2D v[],int num_vert,int bot_ind) Array of vertices ccw order V Index of vertex with smallest y Algorithm 1. Find left edge of tile Go around clockwise, starting from v[bot_ind], until an edge that is not contained in a scanline is found... vn-3 v1 vn-2 v0 vn-1

22 E.T. 06 Scan Convert a Convex Polygon void ScanY(Vertex2D v[],int num_vert,int bot_ind) Array of vertices ccw order Algorithm 1. Find left edge of tile 2. Find right edge of tile V Index of vertex with smallest y This time go ccw! vn-3... v1 vn-2 v0 vn-1

23 E.T. 06 Scan Convert a Convex Polygon void ScanY(Vertex2D v[],int num_vert,int bot_ind) Array of vertices ccw order V Index of vertex with smallest y Algorithm 1. Find left edge of tile 2. Find right edge of tile 3. Scan convert all scan lines going from left to right... vn-3 v1 vn-2 v0 vn-1

24 E.T. 06 Scan Convert a Convex Polygon void ScanY(Vertex2D v[],int num_vert,int bot_ind) Array of vertices ccw order V Index of vertex with smallest y Algorithm 1. Find left edge of tile 2. Find right edge of tile 3. Scan convert all scan lines going from left to right... vn-3 v1 vn-2 v0 vn-1

25 E.T. 06 Scan Convert a Convex Polygon void ScanY(Vertex2D v[],int num_vert,int bot_ind) Array of vertices ccw order V Index of vertex with smallest y Algorithm 1. Find left edge of tile 2. Find right edge of tile 3. Scan convert all scan lines going from left to right Repeat until... either endpt is v1 vn-3 reached vn-2 v0 vn-1

26 Convex Polygons void ScanY( Vertex2D v[], int num_vertices, int bottom_index) { } Initialize variables remaining_vertices = num_vertices; while(remaining_vertices > 0) { } Find the left top row candidate Determine the slope and starting x location for the left tile edge Find the right top row candidate Determine the slope and starting x location for the right tile edge for(row = bottom_row; row < left_top_row && row < right_top_row; row++) { ScanX(ceil(left_pos),ceil(right_pos),row); left_pos += left_step; right_pos += right_step; }... bottom_row = row; vn-3 v1 vn-2 v0 E.T. 06 vn-1

27 Convex Polygons void ScanY( Vertex2D v[], int num_vertices, int bottom_index) { } Initialize variables remaining_vertices = num_vertices; while(remaining_vertices > 0) { } Find the left top row candidate Determine the slope and starting x location for the left tile edge Find the right top row candidate Determine the slope and starting x location for the right tile edge for(row = bottom_row; row < left_top_row && row < right_top_row; row++) { ScanX(ceil(left_pos),ceil(right_pos),row); left_pos += left_step; right_pos += right_step; }... bottom_row = row; vn-3 v1 vn-2 v0 E.T. 06 vn-1

28 Convex Polygons void ScanY( Vertex2D v[], int num_vertices, int bottom_index) { } Initialize variables remaining_vertices = num_vertices; while(remaining_vertices > 0) { } Find the left top row candidate Determine the slope and starting x location for the left tile edge Find the right top row candidate Determine the slope and starting x location for the right tile edge for(row = bottom_row; row < left_top_row && row < right_top_row; row++) { ScanX(ceil(left_pos),ceil(right_pos),row); left_pos += left_step; right_pos += right_step; }... bottom_row = row; vn-3 v1 vn-2 v0 E.T. 06 vn-1

29 Convex Polygons void ScanY( Vertex2D v[], int num_vertices, int bottom_index) { } Initialize variables remaining_vertices = num_vertices; while(remaining_vertices > 0) { } Find the left top row candidate Determine the slope and starting x location for the left tile edge Find the right top row candidate Determine the slope and starting x location for the right tile edge for(row = bottom_row; row < left_top_row && row < right_top_row; row++) { ScanX(ceil(left_pos),ceil(right_pos),row); left_pos += left_step; right_pos += right_step; }... bottom_row = row; vn-3 v1 vn-2 v0 E.T. 06 vn-1

30 Convex Polygons void ScanY( Vertex2D v[], int num_vertices, int bottom_index) { } Initialize variables remaining_vertices = num_vertices; while(remaining_vertices > 0) { } Find the left top row candidate Determine the slope and starting x location for the left tile edge Find the right top row candidate Determine the slope and starting x location for the right tile edge for(row = bottom_row; row < left_top_row && row < right_top_row; row++) { ScanX(ceil(left_pos),ceil(right_pos),row); left_pos += left_step; right_pos += right_step; }... bottom_row = row; vn-3 v1 vn-2 v0 E.T. 06 vn-1

31 Convex Polygons void ScanY( Vertex2D v[], int num_vertices, int bottom_index) { } Initialize variables remaining_vertices = num_vertices; while(remaining_vertices > 0) { } Find the left top row candidate Determine the slope and starting x location for the left tile edge Find the right top row candidate Determine the slope and starting x location for the right tile edge for(row = bottom_row; row < left_top_row && row < right_top_row; row++) { ScanX(ceil(left_pos),ceil(right_pos),row); left_pos += left_step; right_pos += right_step; }... bottom_row = row; vn-3 v1 vn-2 v0 E.T. 06 vn-1

32 Find a tile Find the left top row candidate while( left_top_row <= bottom_row && remaining_vertices > 0) { Move to next edge: edge_start = left_edge_end; Be careful with C % operator, (N-1) % M will give -1 for N = 0, need to use (N+M-1) % M to get (N-1) mod M = N-1 left_edge_end = (left_edge_end+num_vertices-1)%num_vertices; left_top_row = ceil(v[left_edge_end].y); remaining_vertices--; } We found the first edge that sticks out over bottom_row determine the slope and starting x location for the left tile edge. if(left_top_row > bottom_row ) { left_step = (v[left_edge_end].x - v[edge_start].x)/ (v[left_edge_end].y - v[edge_start].y); left_pos = v[edge_start].x + (bottom_row-v[edge_start].y)*left_step; } vn-3 v1 vn-2 v0

33 Find a tile Find the right top row candidate; determine the slope and starting x location for the right tile edge. Exactly as for the left edge. Scan convert a single row: void ScanX(int left_col, int right_col, int row, int R, int G, int B) { } if( left_col < right_col) { } for( int x = left_col; x < right_col; x++) { } draw_pixel(x,y);

34 E.T. 06 Antialiasing

35 Texture Mapping

36 Why Textures? Brick wall - each brick separate polygon There may be bumps, rough spots... But a wall is just one large flat rectangle Then model it as one and glue a picture of a brick wall on the polygon. E.T. 06

37 Texture mapping Texture slides are based on E. Angel s slides y z x geometry screen image

38 Texture Example The texture (below) is a 256 x 256 image that has been mapped to a rectangular polygon which is viewed in perspective

39 E.T. 06 Steps in Texture Mapping 1. Specify a texture and create a texture object 2. Indicate how the texture is to be applied to each pixel 3. Enable texture mapping 4. Draw the scene supplying both texture and geometric coordinates

40 Specifying the Texture Define a texture image from an array of pixels in memory E.T. 06 void glteximage2d( GLenum target, GLint level, GLint internalformat, GLsizei width, GLsizei height GLint border, GLenum format, GLenum type, const GLVoid * texels ) GL_TEXTURE_2D if multiple resolutions, 0 otherwise components and resolutions GL_RGBA 2 M +2b GL_RGBA GL_UNSIGNED_BYTE

41 Converting A Texture Image If dimensions of image are not power of 2 gluscaleimage( format, w_in, h_in, type_in, *data_in, w_out, h_out, type_out, *data_out );! *_in is for source image! *_out is for destination image Image interpolated and filtered during scaling

42 Specifying a Texture: Other Methods Specifying a Texture: Other Methods Use frame buffer as source of texture image! uses current buffer as source image glcopyteximage2d(...) glcopyteximage1d(...) Modify part of a defined texture gltexsubimage2d(...) gltexsubimage1d(...) no size restriction Do both with glcopytexsubimage2d(...), etc.

43 E.T. 06 Texture Objects Stores texture data and makes it readily available. Have many textures and go back and forth using them Much faster than reloading an image using glteximage Sometimes a high-performance subset.

44 E.T. 06 Texture Objects 1. Generate texture names glgentextures(n, *texnames ); Return n unused names in array texnames creates, uses and stops using texture objects 2. Initially bind texture objects to texture data glbindtextures(target, id); Following calls to glteximage(), gltexsubimage(), gltexparameteri() store data in this texture object 3. If high-performance available - priorities 4. Bind and rebind to make textures available for models 5. Cleaning up gldeletetextures(n, *texnames);

45 E.T. 06 Steps in Texture Mapping 1. Specify a texture and create a texture object 2. Indicate how the texture is to be applied to each pixel 3. Enable texture mapping 4. Draw the scene supplying both texture and geometric coordinates

46 Texture Application Methods Filter Modes! minification or magnification! special mipmap minification filters Wrap Modes! clamping or repeating Texture Functions! how to mix primitive s color with texture s color! blend, modulate or replace texels

47 Filter Modes Example: gltexparameteri( target, type, mode ); Texture Magnification Polygon Texture Minification Polygon target: GL_TEXTURE_*D type : GL_TEXTURE_MAG_FILTER, GL_TEXTURE_MIN_FILTER mode: GL_NEAREST, GL_LINEAR, GL_*_MIPMAP_*(for min only) E.T. 06

48 E.T. 06 GL_NEAREST texel with coords nearest the center of the pixel is used. may cause aliasing artifacts. less computation GL_LINEAR weighted avg of 2x2 array texels that lie nearest to the center of the pixel. smoother results

49 Mipmaps multum in parvo -- many things in a small place A texture LOD technique Prespecify a series of prefiltered texture maps of decreasing resolutions Requires more texture storage Eliminates shimmering and flashing as objects move

50 MIPMAPS Arrange different versions into one block of memory

51 MIPMAPS With versus without MIPMAP

52 Mipmapped Textures Mipmap allows for prefiltered texture maps of decreasing resolutions Lessens interpolation errors for smaller textured objects Declare mipmap level during texture definition glteximage*d( GL_TEXTURE_*D, level, ) GLU mipmap builder routines glubuild*dmipmaps( ) (also scales orig. image) OpenGL 1.2 introduces advanced LOD controls

53 MipMaps OpenGL 1.2 introduces advanced LOD controls before: no on the fly changes (e.g. additional levels) could not pick lowest resolution popping effect GL_TEXTURE_BASE_LEVEL GL_TEXTURE_MAX_LEVEL GL_TEXTURE_MIN_LOD GL_TEXTURE_MAX_LOD glubuild*dmipmaplevels(...) IF YOU FAIL TO LOAD A (NEC.) MIPMAP LEVEL BUT CALL FOR IT, TEXTURES GET DISABLED

54 Wrapping Mode Example: gltexparameteri( GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP ) gltexparameteri( GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT ) t texture s GL_REPEAT wrapping GL_CLAMP wrapping

55 Texture Functions Controls how texture is applied gltexenv{fi}[v]( GL_TEXTURE_ENV, prop, param ) GL_TEXTURE_ENV_MODE modes! GL_MODULATE! GL_BLEND! GL_REPLACE -use texture with lighting - different blending values per each color -overwrite current color with texture. Set blend color with GL_TEXTURE_ENV_COLOR [ GL_DECAL ] - if texture or model has translucency,it will be preserved

56 Perspective Correction Hint Texture coordinate and color interpolation! either linearly in screen space! or using depth/perspective values (slower) Noticeable for polygons on edge glhint( GL_PERSPECTIVE_CORRECTION_HINT, hint ) where hint is one of! GL_DONT_CARE! GL_NICEST! GL_FASTEST

57 E.T. 06 Steps in Texture Mapping 1. Specify a texture and create a texture object 2. Indicate how the texture is to be applied to each pixel 3. Enable texture mapping 4. Draw the scene supplying both texture and geometric coordinates

58 E.T. 06 glpixelstorei(gl_unpack_alignment, 1);... glenable() / gldisable() GL_TEXTURE_1D,GL_TEXTURE_2D,...

59 E.T. 06 Steps in Texture Mapping 1. Specify a texture and create a texture object 2. Indicate how the texture is to be applied to each pixel 3. Enable texture mapping 4. Draw the scene supplying both texture and geometric coordinates

60 Mapping a Texture Based on parametric texture coordinates gltexcoord*() specified at each vertex " &$%# Texture Space 1 #$%# Object Space '!$%"(%)%'&*+$%&*,(. &$%& 2 3 #$%&! '&*-$%&*+( / 0 '&*,$%&*-(

61 E.T. 06 Automatic Generation of Texture Coordinates gltexgen{ifd}[v] (coord, pname, param); glenable(gl_texture_gen_s/t/r) coord: GL_S, GL_T, GL_R PNAME GL_TEXTURE_GEN_MODE GL_OBJECT_PLANE GL_EYE_PLANE PARAM GL_OBJECT_LINEAR, GL_EYE_LINEAR GL_SPHERE_MAP [v] array of values

62 E.T. 06 Automatic Generation of Texture Coordinates Specify a plane: Ax+By+Cz+D = 0 Texture coordinates are defined as distances from this plane Example: GLfloat plane[4] = {A, B, C, D}; gltexgeni(gl_s, GL_TEXTURE_GEN_MODE, GL_EYE_LINEAR); gltexgenfv(gl_s, GL_EYE_PLANE, plane); glenable(gl_texture_gen_s);

63 Eye vs. Object Plane

64 Applying Textures II! specify textures in texture objects! set texture filter! set texture function! set texture wrap mode! set optional perspective correction hint! bind texture object! enable texturing! supply texture coordinates for vertex! coordinates can also be generated

65 Bump Mapping Bump Mapping = + Geometry Bump map Bump mapped geometry

66 Displacement Mapping Bump mapped normals are inconsistent with surface geometry Shadow problems Displacement mapping affects surface geometry wikipedia.org

67 TGA image Pixel Hw2 Help i Has 4 entries : BGRA image=>pixels() one big array 4*(i)

68 Sphere Think of a unit circle first - parametric equation: (cos u, sin u) gives you a point on circle. Function of one variable u [0, π] Sphere has 2 variables: the angles u,v f(u,v) = (cos(v)cos(u), cos(v)sin(u), sin(v)) u [0, π], v [ π/2, π/2] m meridians, n parallels note the poles

69 Alpha Channels