![Page 1: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/1.jpg)
6.003: Signals and Systems
Sampling and Quantization
December 1, 2009
![Page 2: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/2.jpg)
Last Time: Sampling and Reconstruction
Uniform sampling (sampling interval T):
tn
x[n] = x(nT )
Impulse reconstruction:
tn
xp(t) =∞∑n=−∞
x[n]δ(t− nT )
Relation:
xp(t) = x(t)p(t) where p(t) =∞∑n=−∞
δ(t− nT )
![Page 3: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/3.jpg)
Fourier Representation
Sampling followed by impulse reconstruction
≡ multiplication by an impulse train in time
≡ convolving by an impulse train in frequency
ω
X(jω)
−W W
1
ω
P (jω)
−ωs ωs
2πT
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs ωs= 2πT
1T
![Page 4: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/4.jpg)
Fourier Representation
Bandlimited reconstruction: lowpass filter, remove high frequencies.
ω
X(jω)
−W W
1
ω
P (jω)
−ωs ωs
2πT
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
1T
−ωs2ωs2
T
![Page 5: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/5.jpg)
The Sampling Theorem
If signal is bandlimited → sample without loosing information.
If x(t) is bandlimited so that
X(jω) = 0 for |ω| > ωm
then x(t) is uniquely determined by its samples x(nT ) if
ωs = 2πT> 2ωm.
The minimum sampling frequency, 2ωm, is called the “Nyquist rate.”
![Page 6: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/6.jpg)
Aliasing
Aliasing results if there are frequency components with ω > ωs2 .
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
X(jω)1
![Page 7: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/7.jpg)
Aliasing
Aliasing results if there are frequency components with ω > ωs2 .
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
X(jω)1
![Page 8: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/8.jpg)
Aliasing
Aliasing results if there are frequency components with ω > ωs2 .
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
X(jω)1
![Page 9: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/9.jpg)
Aliasing
Aliasing results if there are frequency components with ω > ωs2 .
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
X(jω)1
![Page 10: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/10.jpg)
Anti-Aliasing Filter
To avoid aliasing, remove frequency components that alias before
sampling.
−ωs2ωs2ω
1×
−ωs2ωs2ω
Tx(t)
p(t)
xr(t)xp(t)
ReconstructionFilter
Anti-aliasingFilter
![Page 11: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/11.jpg)
Aliasing
Aliasing increases as the sampling rate decreases.
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
X(jω)1
![Page 12: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/12.jpg)
Aliasing
Aliasing increases as the sampling rate decreases.
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
Anti-aliased X(jω)
![Page 13: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/13.jpg)
Aliasing
Aliasing increases as the sampling rate decreases.
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
Anti-aliased X(jω)
![Page 14: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/14.jpg)
Aliasing
Aliasing increases as the sampling rate decreases.
ω
Xp(jω) = 12π (X(j · ) ∗ P (j · ))(ω)
−ωs2ωs2
1T
ω
P (jω)
−ωs ωs
2πT
ω
Anti-aliased X(jω)
![Page 15: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/15.jpg)
Anti-Aliasing Demonstration
Sampling Music
ωs = 2πT
= 2πfs
• fs = 11 kHz without anti-aliasing
• fs = 11 kHz with anti-aliasing
• fs = 5.5 kHz without anti-aliasing
• fs = 5.5 kHz with anti-aliasing
• fs = 2.8 kHz without anti-aliasing
• fs = 2.8 kHz with anti-aliasing
J.S. Bach, Sonata No. 1 in G minor Mvmt. IV. Presto
Nathan Milstein, violin
![Page 16: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/16.jpg)
Quantization
Analog to digital conversion requires not only sampling in time but
also quantization in amplitude.
-1-1
0
0
1
1Input voltage
Out
putv
olta
ge 2 bits 3 bits 4 bits
00
01
10
0 0.5 1-1
0
1
Time (second)0 0.5 1
Time (second)0 0.5 1
Time (second)
-1 0 1Input voltage
-1 0 1Input voltage
Bit rate = (# bits/sample)×(# samples/sec)
e.g., CD quality audio: bit rate =
(16 bits/sample)×(44,100 samples/sec)≈ 0.7 Mbits/sec
![Page 17: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/17.jpg)
Check Yourself
The original Milstein violin piece is taken from a CD. The sam-
pling rate was 44.1 kHz and each sample was quantized to 16
bits.
If we re-quantize it to fewer bits, how few bits can we use
without causing significant problems for lecture demos?
![Page 18: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/18.jpg)
Quantization Demonstration
Quantizing Music
• 16 bits/sample
• 8 bits/sample
• 4 bits/sample
• 3 bits/sample
• 2 bits/sample
• 1 bit/sample
J.S. Bach, Sonata No. 1 in G minor Mvmt. IV. Presto
Nathan Milstein, violin
![Page 19: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/19.jpg)
Quantizing Images
Converting an image from a continuous representation to a discrete
representation involves the same sort of issues.
This image has 189× 189 pixels, with brightness quantized to 8 bits.
![Page 20: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/20.jpg)
Quantizing Images
8 bit image 7 bit image
![Page 21: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/21.jpg)
Quantizing Images
8 bit image 6 bit image
![Page 22: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/22.jpg)
Quantizing Images
8 bit image 5 bit image
![Page 23: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/23.jpg)
Quantizing Images
8 bit image 4 bit image
![Page 24: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/24.jpg)
Quantizing Images
8 bit image 3 bit image
![Page 25: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/25.jpg)
Quantizing Images
8 bit image 2 bit image
![Page 26: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/26.jpg)
Quantizing Images
8 bit image 1 bit image
![Page 27: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/27.jpg)
Check Yourself
What is the most objectionable artifact of coarse quantization?
8 bit image 3 bit image
![Page 28: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/28.jpg)
Dithering
One very annoying artifact is banding caused by clustering of pixels
that quantize to the same level.
Banding can be reduced by dithering.
Dithering: adding a small amount (±1 quantum) of random noise to
the image before quantizing.
Since the noise is different for each pixel in the band, the noise
causes some of the pixels to quantize to a higher value and some to
a lower. But the average value of the brightness is preserved.
![Page 29: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/29.jpg)
Quantizing Images with Dither
7 bit image 7 bits with dither
![Page 30: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/30.jpg)
Quantizing Images with Dither
6 bit image 6 bits with dither
![Page 31: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/31.jpg)
Quantizing Images with Dither
5 bit image 5 bits with dither
![Page 32: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/32.jpg)
Quantizing Images with Dither
4 bit image 4 bits with dither
![Page 33: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/33.jpg)
Quantizing Images with Dither
3 bit image 3 bits with dither
![Page 34: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/34.jpg)
Quantizing Images with Dither
2 bit image 2 bits with dither
![Page 35: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/35.jpg)
Quantizing Images with Dither
1 bit image 1 bit with dither
![Page 36: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/36.jpg)
Check Yourself
What is the most objectionable artifact of dithering?
3 bit image 3 bit dithered image
![Page 37: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/37.jpg)
Robert’s Technique
One annoying feature of dithering is that it adds noise.
The noise can be reduced using Robert’s technique.
Robert’s technique: add a small amount (±1 quantum) of random
noise before quantizing, then subtract that same amount of random
noise.
![Page 38: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/38.jpg)
Quantizing Images with Robert’s Method
7 bits with dither 7 bits with Robert’s method
![Page 39: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/39.jpg)
Quantizing Images with Robert’s Method
6 bits with dither 6 bits with Robert’s method
![Page 40: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/40.jpg)
Quantizing Images with Robert’s Method
5 bits with dither 5 bits with Robert’s method
![Page 41: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/41.jpg)
Quantizing Images with Robert’s Method
4 bits with dither 4 bits with Robert’s method
![Page 42: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/42.jpg)
Quantizing Images with Robert’s Method
3 bits with dither 3 bits with Robert’s method
![Page 43: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/43.jpg)
Quantizing Images with Robert’s Method
2 bits with dither 2 bits with Robert’s method
![Page 44: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/44.jpg)
Quantizing Images with Robert’s Method
1 bits with dither 1 bit with Robert’s method
![Page 45: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/45.jpg)
Quantizing Images: 3 bits
8 bits 3 bits
dither Robert’s
![Page 46: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/46.jpg)
Quantizing Images: 2 bits
8 bits 2 bits
dither Robert’s
![Page 47: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/47.jpg)
Quantizing Images: 1 bit
8 bits 1 bit
dither Robert’s
![Page 48: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/48.jpg)
Sampling and Quantization
Combining sampling and quantization.
• Resampling (discrete-time sampling) and progressive refinement
• JPEG
![Page 49: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/49.jpg)
Discrete-time Sampling
DT sampling is much like CT sampling.
×x[n] xp[n]
p[n] =∑k δ[n− kN ]
n
xp[n]
0
n
x[n]
0
n
p[n]
0
![Page 50: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/50.jpg)
Discrete-time Sampling
As in CT, sampling introduces additional copies of X(ejΩ).
×x[n] xp[n]
p[n] =∑k δ[n− kN ]
0 2π3
4π3
2π−2π3
−4π3
−2π
13
Ω
Xp(ejΩ)
0 2π−2π
1
Ω
X(ejΩ)
0 2π3
4π3
2π−2π3
−4π3
−2π
2π3
Ω
P (ejΩ)
![Page 51: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/51.jpg)
Discrete-time Sampling
Sampling a finite sequence gives rise to a shorter sequence.
n
xb[n]
0
n
xp[n]
0
n
x[n]
0
Xb(ejΩ) =∑n
xb[n]e−jΩn =∑n
xp[3n]e−jΩn =∑k
xp[k]e−jΩk/3 = Xp(ejΩ/3)
![Page 52: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/52.jpg)
Discrete-time Sampling
But the shorter sequence has a wider frequency representation.
0 2π3
4π3
2π−2π3
−4π3
−2π
13
Ω
Xb(ejΩ) = Xp(ejΩ/3)
0
0 2π−2π
13
Ω
Xp(ejΩ)
2π−2π
1
Ω
X(ejΩ)
0
![Page 53: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/53.jpg)
Discrete-time Sampling
![Page 54: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/54.jpg)
Discrete-time Sampling
![Page 55: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/55.jpg)
Discrete-time Sampling
![Page 56: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/56.jpg)
Discrete-time Sampling
![Page 57: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/57.jpg)
Discrete-time Sampling
![Page 58: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/58.jpg)
Discrete-time Sampling
![Page 59: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/59.jpg)
Discrete-time Sampling
![Page 60: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/60.jpg)
Discrete-time Sampling
Insert zeros between samples to upsample the images.
n
xb[n]
0
n
xp[n]
0
n
x[n]
0
Xb(ejΩ) =∑n
xb[n]e−jΩn =∑n
xp[3n]e−jΩn =∑k
xp[k]e−jΩk/3 = Xp(ejΩ/3)
![Page 61: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/61.jpg)
Discrete-time Sampling
Then filter out the additional copies in frequency.
0 2π3
4π3
2π−2π3
−4π3
−2π
13
Ω
Xb(ejΩ) = Xp(ejΩ/3)
0
0 2π−2π
13
Ω
Xp(ejΩ)
2π−2π
1
Ω
X(ejΩ)
0
![Page 62: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/62.jpg)
Discrete-time Sampling: Progressive Refinement
![Page 63: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/63.jpg)
Discrete-time Sampling: Progressive Refinement
![Page 64: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/64.jpg)
Discrete-time Sampling: Progressive Refinement
![Page 65: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/65.jpg)
Discrete-time Sampling: Progressive Refinement
![Page 66: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/66.jpg)
Discrete-time Sampling: Progressive Refinement
![Page 67: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/67.jpg)
Sampling and Quantization
Combining sampling and quantization.
• Resampling (discrete-time sampling) and progressive refinement
• JPEG
![Page 68: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/68.jpg)
JPEG
Discrete-time Fourier representations are useful for coding digital
images.
Example: JPEG (“Joint Photographic Experts Group”) encodes im-
ages by a sequence of transformations:
• color encoding
• DCT (discrete cosine transform): a kind of Fourier series
• quantization to achieve perceptual compression (lossy)
• Huffman encoding: lossless information theoretic coding
We will focus on the DCT and quantization of its components.
• the image is broken into 8× 8 pixel blocks
• each block is represented by its 8× 8 DCT coefficients
• each DCT coefficient is quantized, using higher resolutions for
coefficients with greater perceptual importance
![Page 69: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/69.jpg)
JPEG
Discrete cosine transform (DCT) is similar to a Fourier series, but
high-frequency artifacts are typically smaller.
Example: imagine coding the following 8× 8 block.
For a two-dimensional transform, take the transforms of all of the
rows, assemble those results into an image and then take the trans-
forms of all of the columns of that image.
![Page 70: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/70.jpg)
JPEG
Periodically extend a row and represent it with a Fourier series.
x[n] = x[n+ 8]
n0 8
There are 8 distinct Fourier series coefficients.
ak = 18∑n=<8>
x[n]e−jkΩ0n ; Ω0 = 2π8
![Page 71: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/71.jpg)
JPEG
DCT is based on a different periodic representation, shown below.
y[n] = y[n+ 16]
n0 16
![Page 72: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/72.jpg)
Check Yourself
Which signal has greater high frequency content?
x[n] = x[n+ 8]
n0 8
y[n] = y[n+ 16]
n0 16
![Page 73: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/73.jpg)
Check Yourself
The first signal, x[n], has discontinuous amplitude. The second sig-
nal, y[n] is not discontinuous, but has discontinuous slope.
x[n] = x[n+ 8]
n0 8
y[n] = y[n+ 16]
n0 16
The magnitude of its Fourier series coefficients decreases faster with
k for the second than for the first.
![Page 74: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/74.jpg)
JPEG
Periodic extension of an 8×8 pixel block can lead to a discontinuous
function even when the “block” was taken from a smooth image.
original row
n0
8 pixel ”block”
n0
x[n] = x[n+ 8]
n0 8
![Page 75: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/75.jpg)
JPEG
Periodic extension of the type done for JPEG generates a continuous
function from a smoothly varying image.
original row
n0
8 pixel ”block”
n0
y[n] = y[n+ 16]
n0 16
![Page 76: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/76.jpg)
JPEG
Although periodic in N = 16, y[n] can be represented by just 8 distinct
DCT coefficients.
y[n] = y[n+ 16]
n0 16
bk =7∑n=0y[n] cos
(πk
N
(n+ 1
2
))
This results because y[n] is symmetric about n = −12 , and this sym-
metry introduces redundancy in the Fourier series representation.
Notice also that the DCT of a real-valued signal is real-valued.
![Page 77: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/77.jpg)
JPEG
The magnitudes of the higher order DCT coefficients are smaller
than those of the Fourier series.
−3−2−1 0 1 2 3 4 m−3−2−101234n Fourier series
m
|ak|
0 1 2 3 4 5 6 7 m01234567n DCT
m
|bk|
![Page 78: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/78.jpg)
JPEG
Humans are less sensitive to small deviations in high frequency com-
ponents of an image than they are to small deviations at low frequen-
cies. Therefore, the DCT coefficients are quantized more coarsely
at high frequencies.
Divide coefficient b[m,n] by q[m,n] and round to nearest integer.
q[m,n] m →
16 11 10 16 24 40 51 61
12 12 14 19 26 58 60 55
14 13 16 24 40 57 69 56
n 14 17 22 29 51 87 80 62
↓ 18 22 37 56 68 109 103 77
24 35 55 64 81 104 113 92
49 64 78 87 103 121 120 101
72 92 95 98 112 100 103 99
![Page 79: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/79.jpg)
Check Yourself
Which of the following tables of q[m,n] (top or bottom) will result
in higher “quality” images?
q[m,n] m →16 11 10 16 24 40 51 6112 12 14 19 26 58 60 5514 13 16 24 40 57 69 56
n 14 17 22 29 51 87 80 62↓ 18 22 37 56 68 109 103 77
24 35 55 64 81 104 113 9249 64 78 87 103 121 120 10172 92 95 98 112 100 103 99
q[m,n] m →32 22 20 32 48 80 102 12224 24 28 38 52 116 120 11028 26 32 48 80 114 139 112
n 28 34 44 58 102 174 160 124↓ 36 44 74 112 136 218 206 154
48 70 110 128 162 208 226 19498 128 156 174 206 256 240 202
144 184 190 196 224 200 206 198
![Page 80: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/80.jpg)
JPEG
Finally, encode the DCT coefficients for each block using “run-
length” encoding followed by an information theoretic (lossless)
“Huffman” scheme, in which frequently occuring patterns are
represented by short codes.
The “quality” of the image can be adjusted by changing the values
of q[m,n]. Large values of q[m,n] result in large “runs” of zeros,
which compress well.
![Page 81: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/81.jpg)
JPEG: Results
1%: 1432 bytes 10%: 2457 bytes 20%: 3533 bytes
40%: 4795 bytes 80%: 8345 bytes 100%: 26k bytes
![Page 82: 6.003: Signals and Systems - Massachusetts Institute of ... · The Sampling Theorem If signal is bandlimited →sample without loosing information. If x(t) is bandlimited so that](https://reader030.vdocuments.pub/reader030/viewer/2022040312/5e4b466a13a20613a5666fe4/html5/thumbnails/82.jpg)
Summary
Today we looked at a number of applications of sampling.
• Sampling and Quantizing Music
• Sampling and Quantizing Images
• Upsampling and Downsampling Images
• JPEG Image Coding