Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
2 views

DSP L03 Signal Sampling and Quantization II

Uploaded by

Shereen Elmetwally
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
2 views

DSP L03 Signal Sampling and Quantization II

Uploaded by

Shereen Elmetwally
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 27

1

SBES152
Biomedical Signal
Processing
Lecture 03
Signal Sampling and Quantization: Practical
Considerations
Shereen El-Metwally sh.elmetwally@eng1.cu.edu.eg

Credit: Adapted from slides by Dr. Muhammed


Rushdi
2

Recap: DSP General Scheme


3

Recap: The Nyquist-


Shannon Sampling
Theorem
4

Signal Reconstruction
5

Signal Reconstruction: fs
=• The
2 freconstruction
max filter (anti-image filter)
is an ideal low-pass filter
• Impractical
6

Signal Reconstruction: fs
> •There
2 fmaxis a separation between the highest-
frequency edge of the baseband spectrum and
the lower edge of the first replica.
•A practical low-pass reconstruction filter
can be designed.
7

Signal Reconstruction: fs
 There is a spectral overlapping between the
<original
2 fmaxbaseband spectrum and the spectrum of
the first replica and so on.
 Even when we apply an ideal lowpass filter to
remove these images, in the baseband there are still
some foldover frequency components from the
adjacent replica.
 If an analog signal with a frequency f is
undersampled (i.e., f > fs/2), its aliasing
frequency component falias in the baseband is : falias
= fs - f

fmax =
B
8

Examples on Sampling (1)

Rewrite the cosine signal using Euler’s identity:


9

Examples on Sampling (1)

All images of the original spectrum must be removed, since


they convey no additional information.
1
0

Examples on Sampling (2)

Rewrite the cosine signals using Euler’s


identity:
1
1

Examples on
Sampling (2)
The sampling theorem condition is satisfied; hence, we can
recover the original spectrum using a reconstruction lowpass
filter
1
2

Examples on Sampling (3)


1
3

Examples on
Sampling (3)
The sampling theorem condition is violated.

f = 5 kHz is undersampled ( f > fs/2), its aliasing


frequency component falias in the baseband is:

Aliasing noise: falias = fs – f = 8 kHz – 5kHz = 3 kHz.


fc = fs/2 = 4 kHz
Butterworth Analog Low-Pass
Filters
Filter magnitude frequency response is given by:

2k

where k is the filter order, and fc is the cutoff (3 dB) frequency.


Half-power point
Cutoff (3dB) frequency is the frequency at
which the filter magnitude response falls
to 1/√2 its maximum value.
In Decibels:
At fc : 20log*|H(f)| = 20 log (1/√2) = -3 dB
1
4

Practical Sampling: Anti-


Aliasing Filtering

Analog signal Analog signal


spectrum spectrum after sampled signal
filtering spectrum
1
5

Practical Sampling: Anti-


Aliasing Filtering
• To satisfy the sampling theorem condition, we apply
an anti-aliasing filter to limit the input analog signal,
so that all the frequency components are less than the
folding frequency (half of the sampling rate).
i.e., All frequencies satisfy: f < fs/2

• Due to nonzero attenuation of the magnitude


frequency response of the anti-aliasing lowpass filter,
the aliasing noise from the adjacent replica still
appears in the baseband.

• We can control the aliasing noise by either using a


higher-order lowpass filter or increasing the
1
6

Practical Sampling: Example Anti-


Aliasing Filter
• Example filter: A second-order Butterworth
lowpass filter with unit gain.
• Practically speaking, aliasing will occur
anyway with some degree.

Filter Magnitude
Frequency Response

fc is the filter
cutoff frequency
Filter order N =
1
7

Practical Sampling:
Aliasing Level
• The aliasing level percentage using the
symmetry of the Butterworth magnitude
function and its first replica:
1
8

Practical Sampling: Aliasing


Level
1
9

Practical Sampling: Effect of


Increased Sampling Rate on
Aliasing Level
2
0

Practical Sampling: Choosing the


Anti- Aliasing filter order
3
0

Frequently Asked Questions


(FAQ)
• Q: The Fourier transform of a sampled signal
shows frequency-shifted copies of the Fourier
transform of the original signal. Does this
mean that “the original signal contains
higher frequencies than it already has”?

• A: No, these frequency images or replicas are


because of the sampling effect. All images of the
original spectrum must be removed during
reconstruction, since they convey no additional
information.
3
1

Frequently Asked Questions


(FAQ)
What does it mean to have a negative frequency in
the frequency spectrum of a signal? Img
• “The concept of negative and positive
frequency can be as simple as a wheel
rotating one way or the other way.
A signed value of frequency can
indicate both the rate and direction Re
of rotation.”
• A complex exponential is expressed as:
ejωt = cos ωt + j sin ωt = 1 ωt
• The vector (cos t, sin t) rotates counter-clockwise at ω
=1 radian/sec, and completes a circle every 2π seconds.
• The vector (cos - t, sin - t) rotates in the other direction
with frequency ω = -1 radian/sec, (not shown).
3
2

Frequently Asked Questions


(FAQ)
What does it mean to have a negative frequency in
the frequency spectrum of a signal?
• For real signals, there are always two equal complex
exponentials rotating in opposite directions, so their real
parts combine and imaginary parts cancel out, resulting
in a real sinusoid.

sin 2πf0 t =
3
3

Frequently Asked Questions


(FAQ)
What does it mean to have a negative amplitude in
the frequency spectrum of a signal?

Every real signal contains equal amounts of


positive and negative frequencies.

You might also like