Aliasing and Anti-Aliasing

Question 1

How is aliasing caused?

Answer 1

Aliasing is caused by taking samples at too low a frequency?

Question 2

What rate do we sample above to reconstruct a signal correctly?

Answer 2

The Nyquist rate, if the frequency of the signal is f, samples must be taken at least 2f

Question 3

What happens to an image if it’s sampled more?

Answer 3

It becomes more accurate.

Question 4

What is unweighted sampling?

Answer 4

We measure the area of coverage of the line of the pixel and set that percentage to the colour of the foreground over the background

Question 5

What is the benefit of unweighted sampling?

Answer 5

Smooth transition from on to off

Question 6

What is the disadvantage of unweighted sampling?

Answer 6

Hard to compute.

Question 7

What is super-sampling?

Answer 7

More sampling per pixels. Samples are subdivided into sub-pixels and compared

Question 8

What does super-sampling do for ray-tracing?

Answer 8

Trace more rays per pixel

Question 9

What is the disadvantage of super-sampling?

Answer 9

More computation needed.

Question 10

What is the advantage of super-sampling?

Answer 10

Better image quality

Question 11

What is adaptive aliasing?

Answer 11

Find the pixels on edges and do more samples per pixel there instead of non-edges.

Question 12

What is the process of adaptive super-sampling?

Answer 12

• Sample scene at each pixel corner
• If the 4 samples vary by more then the current threshold
  -Take 5 more samples to create 4 subpixels
• Recursively repeat this process on each of the subpixels until the samples are less than the current threshold or a limit has been reached
• Average all the samples for the anti-aliased result