ECE 33
Signals, Spectra, and
Signal Processing
Sampling Theorem and Aliasing
Lesson 3
Start
Engr. Santos James Emmanuel C. Malunda
Faculty, ECE Dept
College of Engineering and Architecture
Lesson 3: Quantization of
Continuous-valued Signals
• At the end of this lesson, the student shall be
able to convert continuous-valued signals into
discrete-valued signals.
NEXT
Analog to Digital Converter
A/D Converter
xa(t) x[n] xq[n] 1011000...
Sampler Quantizer Coder
Analog Discrete-time Quantized Digital
Next
Signal Signal (discrete-valued) Signal
Signal
Analog to Digital Conversion
1. Sampling
– converting continuous-time to discrete-time signal
– taking the samples of continuous-time signal into
discrete-time signal
Previous
2. Quantization
Next
– each sample is truncated or rounded off to the nearest
value from a finite set of values
3. Coding
– each discrete-value is represented by a binary number
Quantizing
• converting a discrete-time continuous-valued
signal into digital signal by expressing each
sample by a finite set of values
Previous
Next
• produces an error called Quantization Error or
Quantization Noise
(Proakis, 1996)
Quantizer Operation
xq[n] = Q{x[n]}
eq[n] = xq[n] - x[n]
Previous
Next
where
• xq[n] = quantized signal
• Q = quantizer funtion
• eq[n] = quantization error
(Proakis, 1996)
Previous
Next
Quantizing (example)
(Proakis, 1996)
Previous
Next
Δ = quantization step size
Δ = resolution
dynamic range = xmax - xmin
Previous
Next
L = no. of quantization levels
x[n] xq[n] xq[n] eq[n]
n Discrete-time signal Truncation Rounding Rounding
0 1.0 1.0 1.0 0.0
1 0.9 0.9 0.9 0.0
2 0.81 0.8 0.8 -0.1
3 0.729 0.7 0.7 -0.029
Previous
4 0.6561 0.6 0.7 0.0439
Next
5 0.59049 0.5 0.6 0.00951
6 0.531441 0.5 0.5 -0.031441
7 0.4782969 0.4 0.5 0.0217031
8 0.43046721 0.4 0.4 -0.03046721
9 0.387420489 0.3 0.4 0.012579511
(Proakis, 1996)
Quantization error
• Rounding:
Previous
Next
• Truncation:
4
1
Previous
Next
-1
-2
-3
-4
4 (0100)
3 (0011)
2 (0010)
1 (0001)
(0000)
Previous
0
(1000)
Next
-1 (1001)
-2 (1010)
-3 (1011)
-4 (1100)
0011 0100 0010 1010 1100 1011 1000
Try this (from lesson 2):
(Proakis, 1996)
Given the analog signal:
xa(t) = 3 cos 2000πt + 5 sin 6000πt + 10 cos 12000πt
Previous
Next
The signal is sampled at a rate of 5000
samples/s and is quantized into 1024 distinct
voltage levels. What is the resolution Δ?
Summary
• quantization error • Rounding:
eq[n] = xq[n] - x[n]
Previous
• resolution • Truncation:
END
