antialiasing - Department of Computer Science and ...olano/class/435-08-2/antialiasing.pdfNo...

1

Antialiasing

2

  Visual artifacts -  Jagged lines and edges - High frequencies appearing as low - Small objects missed - Texture distortions - Strobing and popping - Backward movement

3

4

5

6

7

8

  Two basic stages - Sampling - Reconstruction

  Assuming discrete sampling

9

10

11

12

13

14

15

16

17

  Shannon's sampling theory (1D): - A band limited signal f(t) with cut off frequency wF

may be perfectly reconstructed from its samples f(nT0) if 2π/T0 >= 2wF

- wF == Nyquist limit   Alternatively:

- A signal can be reconstructed exactly from samples only if the highest frequency is less than half the sampling rate

18

  Two samples of the same sine wave

19

  Regular supersampling   Jittered supersampling   Adaptive supersampling   Stochastic sampling

20

21

  Reconstruction: recreate a continuous signal from a set of samples

  Tasks of reconstruction filter - Remove extraneous replicas of signal spectrum - Pass the original signal base unchanged

22 Box Filter Tent Filter

23

B-Spline Filter

Mitchell-Netravali Filter

Gaussian Filter

Catmull-Rom Filter

24

25

26

27

28

29

30

31