551/651:
Advanced CS Computer Graphics
Accelerating Ray Tracing Continued
Antialiasing
David Luebke
1
Recap: Accelerating Ray Tracing
q
Problem:
s
s
s
q
Lots of pixels = lots of rays
Computing shadows generates even more rays
Supersampling generates even more rays
Solutions:
s
s
s
Trace rays more quickly
Shoot fewer rays
Reorder ray computation
David Luebke
2
Antialiasing
Aliasing: signal processing term with very
specific meaning
q Aliasing: computer graphics term for any
unwanted visual artifact
q Antialiasing: computer graphics term for
avoiding unwanted artifacts
q Well tackle these in order
q
David Luebke
3
Signal Processing
Raster display: regular sampling of a
continuous function (Really?)
q Think about sampling a 1-D function:
q
David Luebke
4
Signal Processing
q
Sampling a 1-D function:
David Luebke
5
Signal Processing
q
Sampling a 1-D function:
David Luebke
6
Signal Processing
q
Sampling a 1-D function:
s
What do you notice?
David Luebke
7
Signal Processing
q
Sampling a 1-D function: what do you notice?
s
Jagged, not smooth
David Luebke
8
Signal Processing
q
Sampling a 1-D function: what do you notice?
s
s
Jagged, not smooth
Loses information!
David Luebke
9
Signal Processing
q
Sampling a 1-D function: what do you notice?
s
s
q
Jagged, not smooth
Loses information!
What we can do about these?
s
s
s
Use higher-order reconstruction
Use more samples
How many more samples?
David Luebke
10
The Sampling Theorem
Obviously, the more samples we take the
better those samples approximate the original
function
q The sampling theorem:
q
A continuous bandlimited function can be completely
represented by a set of equally spaced samples, if the
samples occur at more than twice the frequency of the
highest frequency component of the function
David Luebke
11
The Sampling Theorem
In other words, to adequately capture a
function with maximum frequency F, we need
to sample it at frequency N = 2F.
q N is called the Nyquist limit.
q
David Luebke
12
The Sampling Theorem
q
An example: sinusoids
David Luebke
13
The Sampling Theorem
q
An example: sinusoids
David Luebke
14
The Sampling Theorem
q
Show Figure 4.2 in the book (p. 113)
David Luebke
15
Fourier Theory
All our examples have been sinusoids
q Does this help with real world signals? Why?
q Fourier theory lets us decompose any signal
into the sum of (a possibly infinite number of)
sine waves
q
David Luebke
16
Fourier Theory:
The Ugly Stuff
Next comes the math
q If you are not comfortable with signal theory,
you might want to read Chapter 4 in W&W or
find a good tutorial on the web
q
David Luebke
17
The End
David Luebke
18