Sampling a signal

1
Sampling a signal
Reconstruction
Display
Spatial domain
Result on screen
2
Sampling a signal
Reconstruction
Display
Fourier domain
Result on screen
3
Filtering
Sampling
Reconstruction
Spatial domain
Result on screen
4
Filtering
Sampling
Reconstruction
Frequency domain
Result on screen
5
Filtering solution
Frequency domain
Spatial domain
6
Subsampling
Number of subsamples determines spacing of
copies Weighting of sub- samples
determines shape of kernel
7
Bad recon- struction
D/A converter convolves with a box function We
would like this to be a sinc function
8
Bad recon- struction
Frequency domain
9

• Filtering is one solution to aliasing problem
• Random sampling is another
• Jittering
• Trades aliasing for noise
• Noise is less objectionable to the human eye