Upload
oswin-gibbs
View
213
Download
0
Tags:
Embed Size (px)
Citation preview
02/26/02 (c) 2002 University of Wisconsin, CS 559
Last Time
• Canonical view pipeline
• Orthographic projection– There was an error in the matrix for taking a simple orthographic
volume and transforming it into the canonical view space
– The slides now online are correct
• In Shirley’s chapter on Transformation, note notation errors in the discussion of homogeneous coordinates
Local Coordinate
Space
World Coordinate
Space
View Space
3D Screen Space
Display Space
Projection
02/26/02 (c) 2002 University of Wisconsin, CS 559
Today
• Perspective viewing– Simple case
– Completely general case
02/26/02 (c) 2002 University of Wisconsin, CS 559
Perspective Projection
• Abstract camera model - box with a small hole in it
• Pinhole cameras work in practice - camera obscura, etc
02/26/02 (c) 2002 University of Wisconsin, CS 559
Parallel lines meetcommon to draw film planein front of the focal point
02/26/02 (c) 2002 University of Wisconsin, CS 559
Vanishing points
• Each set of parallel lines (=direction) meets at a different point: The vanishing point for this direction– Classic artistic perspective is 3-
point persepctive
• Sets of parallel lines on the same plane lead to collinear vanishing points: the horizon for that plane
• Good way to spot faked images
02/26/02 (c) 2002 University of Wisconsin, CS 559
Basic Perspective Projection
• Assume you have transformed to view space, with x to the right, y up, and z back toward the viewer
• Assume the origin of view space is at the center of projection
• Define a focal distance, d, and put the image plane there (note d is negative)– This doesn’t quite fit into our viewing model, but we’ll come back
to that
02/26/02 (c) 2002 University of Wisconsin, CS 559
Basic Perspective Projection
• If you know P(xv,yv,zv) and d, what is P(xs,ys)?– Where does a point in view space end up on the screen?
xv
yv
-zvd
P(xv,yv,zv)P(xs,ys)
02/26/02 (c) 2002 University of Wisconsin, CS 559
Basic Case
• Similar triangles gives:
v
vs
z
x
d
x
v
vs
z
y
d
y
yv
-zv
P(xv,yv,zv)P(xs,ys)
View Plane
d
02/26/02 (c) 2002 University of Wisconsin, CS 559
Simple Perspective Transformation
• Using homogeneous coordinates we can write:
dzz
y
x
d
y
x
v
v
v
v
s
s
vs
d
PP
0100
0100
0010
0001
02/26/02 (c) 2002 University of Wisconsin, CS 559
Perspective View Volume
• Recall the orthographic view volume, defined by a near, far, left, right, top and bottom plane
• The perspective view volume is also defined by near, far, left, right, top and bottom planes – the clip planes– Near and far planes are parallel to the image plane: zv=n, zv=f
– Other planes all pass through the center of projection (the origin of view space)
– The left and right planes intersect the image plane in vertical lines
– The top and bottom planes intersect in horizontal lines
02/26/02 (c) 2002 University of Wisconsin, CS 559
Clipping Planes
xv
-zv
Near Clip Plane
Far Clip PlaneView
Volume
Left ClipPlane
Right ClipPlane
fn l
r
02/26/02 (c) 2002 University of Wisconsin, CS 559
Where is the Image Plane?
• Notice that it doesn’t really matter where the image plane is located, once you define the view volume– You can move it forward and backward along the z axis and still get
the same image, only scaled
• But we need to know where it is to define the clipping planes– Assume the left/right/top/bottom planes are defined according to
where they cut the near clip plane
• Or, define the left/right and top/bottom clip planes by the field of view
02/26/02 (c) 2002 University of Wisconsin, CS 559
Clipping Planes
xv
-zv
Near Clip Plane
Far Clip PlaneView
Volume
Left ClipPlane
Right ClipPlane
fFOV
02/26/02 (c) 2002 University of Wisconsin, CS 559
OpenGL
• gluPerspective(…)– Field of view in the y direction (vertical field-of-view)
– Aspect ratio (should match window aspect ratio)
– Near and far clipping planes
– Defines a symmetric view volume
• glFrustum(…)– Give the near and far clip plane, and places where the other clip
planes cross the near plane
– Defines the general case
– Used for stereo viewing, mostly
02/26/02 (c) 2002 University of Wisconsin, CS 559
Perspective Projection Matrices
• We want a matrix that will take points in our perspective view volume and transform them into the orthographic view volume– This matrix will go in our pipeline just before the orthographic
projection matrix
(l,b,n)(r,t,n) (l,b,n)
(r,t,n)
02/26/02 (c) 2002 University of Wisconsin, CS 559
Mapping Lines
• We want to map all the lines through the center of projection to parallel lines– Points on lines through the center of projection map to the same
point on the image– Points on parallel lines map orthographically to the same point on
the image– If we convert the perspective case to the orthographic case, we can
use all our existing methods
• The intersection points of lines with the near clip plane should not change
• The matrix that does this, not surprisingly, looks like the matrix for our simple perspective case
02/26/02 (c) 2002 University of Wisconsin, CS 559
General Perspective
• This matrix leaves points with z=n unchanged
• It is just like the simple projection matrix, but it does some extra things to z to map the depth properly
• We can multiply a homogenous matrix by any number without changing the final point, so the two matrices above have the same effect
0100
00
000
000
0100
00
0010
0001
nffn
n
n
n
fnfnPM
02/26/02 (c) 2002 University of Wisconsin, CS 559
Complete Perspective Projection
• After applying the perspective matrix, we still have to map the orthographic view volume to the canonical view volume:
0100
00
000
000
1000
200
02
0
002
nffn
n
n
fn
fn
fn
bt
bt
bt
lr
lr
lr
POscreenview MMM
02/26/02 (c) 2002 University of Wisconsin, CS 559
OpenGL Perspective Projection
• For OpenGL you give the distance to the near and far clipping planes
• The total perspective projection matrix resulting from a glFrustum call is:
0100
200
02
0
002
fn
nf
fn
fnbt
bt
bt
nlr
lr
lr
n
OpenGLM
02/26/02 (c) 2002 University of Wisconsin, CS 559
Near/Far and Depth Resolution
• It may seem sensible to specify a very near clipping plane and a very far clipping plane– Sure to contain entire scene
• But, a bad idea:– OpenGL only has a finite number of bits to store screen depth– Too large a range reduces resolution in depth - wrong thing may be
considered “in front”– See Shirley for a more complete explanation
• Always place the near plane as far from the viewer as possible, and the far plane as close as possible
02/26/02 (c) 2002 University of Wisconsin, CS 559
Clipping
• Parts of the geometry to be rendered may lie outside the view volume– View volume maps to memory addresses– Out-of-view geometry generates invalid addresses– Geometry outside the view volume also behaves very strangely
under perspective projection• Triangles can be split into two pieces, for instance
• Clipping removes parts of the geometry that are outside the view
• Best done in screen space before perspective divide– Before dividing out the homogeneous coordinate
02/26/02 (c) 2002 University of Wisconsin, CS 559
Clipping
• Points are trivial to clip - just check which side of the clip planes they are on
• Many algorithms for clipping lines exist– Next lecture
• Two main algorithms for clipping polygons exist– Sutherland-Hodgman (today)
– Weiler (next lecture)
02/26/02 (c) 2002 University of Wisconsin, CS 559
Clipping Points
• A point is inside the view volume if it is on the “inside” of all the clipping planes– The normals to the clip planes are considered to point inward, toward the
visible stuff
• Now we see why clipping is done in canonical view space
• For instance, to check against the left plane:– X coordinate in 3D must be > -1
– In homogeneous screen space, same as: xscreen> -wscreen
• In general, a point, p, is “inside” a plane if:– You represent the plane as nxx+nyy+nzz+d=0, with (nx,ny,nz) pointing
inward
– And nxpx+nypy+nzpz+d>0
02/26/02 (c) 2002 University of Wisconsin, CS 559
Polygon-Rectangle Clipping (2D)
• Task: Clip a polygon to a rectangle
• Easy cases:
• Hard Cases:
02/26/02 (c) 2002 University of Wisconsin, CS 559
Sutherland-Hodgman Clip (1)
• Clip the polygon against each edge of the clip region in turn– Clip polygon each time to line containing edge
– Only works for convex clip regions (Why?)
02/26/02 (c) 2002 University of Wisconsin, CS 559
Sutherland-Hodgman Clip (2)
• To clip a polygon to a line:– Consider the polygon as a list of vertices
– One side of the line is inside the clip region, the other outside
– Think of the process as rewriting the polygon, one vertex at a time
– Check start vertex: if “inside”, emit it, otherwise ignore it
– Process vertex list as follows…
02/26/02 (c) 2002 University of Wisconsin, CS 559
Sutherland-Hodgman (3)
• Look at the next vertex in the list:– polygon edge crosses clip edge going from out to in: emit crossing
point, next vertex
– polygon edge crosses clip edge going from in to out: emit crossing
– polygon edge goes from out to out: emit nothing
– polygon edge goes from in to in: emit next vertex
02/26/02 (c) 2002 University of Wisconsin, CS 559
Sutherland-Hodgman (4)
Inside Outside
s
p
Output p
Inside Outside
sp
Output i
Inside Outside
s
p
No output
Inside Outside
sp
Output i,p
i
i
02/26/02 (c) 2002 University of Wisconsin, CS 559
Inside-Outside Testing
• Edges store a vector pointing toward the outside of the clip region
• Dot products give inside/outside information
Outside Insiden
s
f
i
x
0
0
0
x)(fn
x)(in
x)(sn
02/26/02 (c) 2002 University of Wisconsin, CS 559
Sutherland-Hodgman (5)
• In 3D, clip against planes instead of lines– Six planes to clip against
– Inside/Outside test still works
• Suitable for hardware implementation– Only need the clip edge, the endpoints of the current edge, and the
last output point
– Polygon edges are output as they are found, and passed right on to the next clip region edge
02/26/02 (c) 2002 University of Wisconsin, CS 559
Inside/Outside in Screen Space
• In screen space, clip planes are xs=±1, ys=±1, zs=0, zs=1
• Inside/Outside reduces to comparisons before perspective divide
ss
sss
sss
wz
wyw
wxw
0
02/26/02 (c) 2002 University of Wisconsin, CS 559
Additional Clipping Planes
• Useful for doing things like cut-away views– Use a clip plane to cut off part of the object
– Only works if piece to be left behind if convex
• OpenGL allows you to do it
• Also one way to use OpenGL to identify objects in a region of space (uses the selection mechanism)
02/26/02 (c) 2002 University of Wisconsin, CS 559
Other Ways to Reject
• If a polygonal object is closed, then no back-facing face is visible– Front-facing faces must occlude all back-facing ones
– Reject back-facing polygons in view space• Transform face normal and check
• Bounding volumes enclosing many polygons can be checked against the view volume– Done in software in world or view space
• Visibility can reject whole chunks of geometry without even looking at them