p01 - pengantar digital signal processing
TRANSCRIPT
-
7/25/2019 P01 - Pengantar Digital Signal Processing
1/47
Digital Signal Processing:An Introduction and Some Examples of its
Everyday Use
Ardik Wijayanto
-
7/25/2019 P01 - Pengantar Digital Signal Processing
2/47
Target
Sampling
FFT
Filter
Slide 2
-
7/25/2019 P01 - Pengantar Digital Signal Processing
3/47
Materi Dalam 1 Semester
Pendahuluan (1 TM) Dasar DSP : Sinyal dan Sistem (2 TM) + 1 Tugas
ADC dan Segala Aspeknya (1 TM)
Persamaan Beda Koef Linier Konstan (2 TM)
Konvolusi (1 TM) + 1 Tugas
Transformasi Laplace (1 TM)
Transformasi Z (1 TM) Filter Digital dan Segala Aspeknya (2 TM) + 1 Tugas
Transformasi Fourier (1 TM) + 1 Tugas
TM = Tatap Muka
-
7/25/2019 P01 - Pengantar Digital Signal Processing
4/47
DSP ???
-
7/25/2019 P01 - Pengantar Digital Signal Processing
5/47
Slide 5
Contents
What is DSP?
What is DSP used for?
Speech & Audio processing
Image & Video processing Adaptive filtering
DSP Devices and Architectures
Summary & Conclusions
-
7/25/2019 P01 - Pengantar Digital Signal Processing
6/47
Slide 6
What is DSP?
Digital Signal Processing the processingor manipulation of signals using digital
techniques
ADC DAC
Digital
Signal
ProcessorAnalogueto Digital
Converter
Digital to
Analogue
Converter
Input
SignalOutput
Signal
-
7/25/2019 P01 - Pengantar Digital Signal Processing
7/47
PSD - Ardik / Bima
-
7/25/2019 P01 - Pengantar Digital Signal Processing
8/47
PSD - Ardik / Bima
-
7/25/2019 P01 - Pengantar Digital Signal Processing
9/47
PSD - Ardik / Bima
-
7/25/2019 P01 - Pengantar Digital Signal Processing
10/47
Slide 10
What is DSP Used For?
And much more!
-
7/25/2019 P01 - Pengantar Digital Signal Processing
11/47
Slide 11
-
7/25/2019 P01 - Pengantar Digital Signal Processing
12/47
Slide 12
The world is filled withsignals
1-D signals:
Seismic vibrations
EEG and EKG
Speech
Sonar
Audio & music
Dow-Jones averages
2-D signals: Photographs
Medical images
Radar
Satellite data Fax
Fingerprints
And of course there are 3-d signals (video, volumetric data sets)
and beyond
-
7/25/2019 P01 - Pengantar Digital Signal Processing
13/47
Slide 13
And we want toprocessthem
Recognizewhats in a signal Target detection
Speech recognition
Image analysis
Predicta future value of thesignal
Stock market prediction Interpolatemissing values
of a signal Conceal lost video packets
Restorea signal that hasbeen degraded
Noise removal Echo cancellation
Enhancea signal Image contrast enhancement
Compressa signal Faster transmission
Less storage space
Synthesizea realistic example ofa signal Speech synthesis
Image texture generation
Choosespecific input signals tocontrol a process Chemical process control
Aerodynamic control
-
7/25/2019 P01 - Pengantar Digital Signal Processing
14/47
PSD - Ardik / Bima
-
7/25/2019 P01 - Pengantar Digital Signal Processing
15/47
Slide 15
-
7/25/2019 P01 - Pengantar Digital Signal Processing
16/47
Slide 16
Speech Processing
Speech coding/compression
Speech synthesis
Speech recognition
-
7/25/2019 P01 - Pengantar Digital Signal Processing
17/47
Slide 17
Some Properties of Speech
The blue--- s---p--o---------t i-s--on--the-- k--ey a---g--ai----n------
oo in blueo in spotee in keye in agains in spotk in key
-
7/25/2019 P01 - Pengantar Digital Signal Processing
18/47
Slide 18
Some Properties of Speech
ee in keyo in spotoo in blue e in again
Vowels
s in spot k in key
Consonants
Quasi-periodic
Relatively high signal power
Non-periodic (random)
Relatively low signal power
-
7/25/2019 P01 - Pengantar Digital Signal Processing
19/47
Slide 19
Speech Coding
BTS
BSC
MSC
TRAU
64 kbits/s
13 kbits/s
22.8 kbits/s
-
7/25/2019 P01 - Pengantar Digital Signal Processing
20/47
Slide 20
Speech Coding Linear Prediction
A(z)
s(n)+
d(n)
se(n)
d(n)
A(z)
++
se(n)
sr(n)
Try topredictthe current sample value; Transmit theprediction error.
-
7/25/2019 P01 - Pengantar Digital Signal Processing
21/47
Slide 21
Speech Coding Vocoder
G
Pulse Train
Random Noise
Vocal Tract
Model
V/U
Synthesized Speech
Decoder
Original Speech
Analysis:
Voiced/Unvoiced decision
Pitch Period (voiced only)
Signal power (Gain)
Signal Power
Pitch
Period
Encoder
LPC-10:
-
7/25/2019 P01 - Pengantar Digital Signal Processing
22/47
Slide 22
To be or
not to bethat is thequestion
Textnormalization
expands
abbreviationsdates, times,money..etc
Parsing Pronunciation
Prosodyrules
Tu bee awrnawt tu beedhat iz dhekwestchun
semantic &
syntactic partsof speechanalysis of text
phonetic description
of each word, dictionarywith letter-to-soundrules as a back up
Waveformgeneration
Synthesizedspeech
Apply word
stress, durationand pitch
Phonetic-to-
acoustictransformation
phonetic form
Inputtext
Text-to-Speech Synthesis
Text-to-speech synthesis sounds very natural these days.
-
7/25/2019 P01 - Pengantar Digital Signal Processing
23/47
Slide 23
Speech Synthesis Applications
Speaking clocks
Spoken (variable) announcements
Talking emails + talking heads for mobile
Synthesis of location-based information(e.g. traffic information)
Interactive systems (e.g. catalogue ordering,
Yellow Pages, ...)
-
7/25/2019 P01 - Pengantar Digital Signal Processing
24/47
Slide 24
Speech/Speaker Recognition
Speech Recognition What has been spoken?
Speaker dependent Recognition system trained
for a particular persons voice.
Speaker independent Recognition system
expected to deal with a wide variety of speakers.
Speaker Recognition Who has spoken?
Not easy
Sometimestherearenogapsbetweenwords.
Sometim esthereareg aps inthe mid dleofwords.
Accents, dialects and Stress eggsist.
-
7/25/2019 P01 - Pengantar Digital Signal Processing
25/47
Slide 25
Speech Recognition System
Featureextractionspeech Phonemerecognition
Phonememodels
SentencerecognitionWordrecognition
Wordpronunciation
Semanticknowledge
decision
Syntacticknowledge
Dialogueknowledge
-
7/25/2019 P01 - Pengantar Digital Signal Processing
26/47
Slide 26
Digital Audio
Standard music CD:
Sampling Rate: 44.1 kHz
16-bit samples 2-channel stereo
Data transfer rate = 21644,100 = 1.4 Mbits/s
1 hour of music = 1.43,600 = 635 MB
-
7/25/2019 P01 - Pengantar Digital Signal Processing
27/47
Slide 27
Audio Coding (Contd)
Key standards: MPEG: Layers I, II, and III (MP3); AAC.
used in DAB, DVD
Dolby AC3, Dolby Digital, Dolby Surround.
Typical bit rates for 2-channel stereo:
64kbits/s to 384 kbits/s.
Subband- or transform-based, making use
of perceptual masking properties.
-
7/25/2019 P01 - Pengantar Digital Signal Processing
28/47
Slide 28
Audio Coding (Contd)
5.1 channels (3/2) with LFE channel:
Left, Right, Centre,
Left Surround, Right Surround,
Low Frequency Effects (LFE) (Reduced Bandwidth). LFE loudspeaker can, in general, be placed anywhere in the
listening room.
Typical 3/2 multichannel stereo configuration:
RightSurround
Right
Surround
Left
Centre
Left
-
7/25/2019 P01 - Pengantar Digital Signal Processing
29/47
Slide 29
Audio Coding Masking
Auditory Masking: Spectral: Strong frequency components mask weaker
neighbouring frequency components.
Temporal: Strong temporal events mask recent and
future events.
SPL/dB
10ms 160ms
Temporal Masking
timefreq/kHz1
SPL/dB
Spectral Masking
-
7/25/2019 P01 - Pengantar Digital Signal Processing
30/47
Slide 30
200 300 400 500 600 700 800
10
20
30
40
50
60
Hz
dB
Masking Example
-
7/25/2019 P01 - Pengantar Digital Signal Processing
31/47
Slide 31
Image/Video
Still Image Coding: JPEG (Joint Photographic Experts Group):
Discrete Cosine Transform (DCT) based
JPEG2000: Wavelet Transform based
Video Coding:
MPEG (Moving Pictures Experts Group):
DCT-based,
Interframe and intraframe prediction,
Motion estimation.
Applications: Digital TV, DVD, etc.
-
7/25/2019 P01 - Pengantar Digital Signal Processing
32/47
Slide 32
JPEG Example
Original
JPEG (100:1)JPEG (4:1)
Video compression: Example of a
-
7/25/2019 P01 - Pengantar Digital Signal Processing
33/47
Slide 33
Video compression: Example of a
packet loss & need to interpolate:
Packet loss means
a horizontal slice of
data is missing
The decoder holds
over the slice fromthe previous frame
to conceal the loss.
Because of camera
pan, slice from
previous framedoesnt line up well
Interpolating the missing data in the current frame using the data from
above & below would likely provide better visual concealment of the loss
-
7/25/2019 P01 - Pengantar Digital Signal Processing
34/47
Slide 34
Example: contrast enhancement
Original magnetic resonance
brain scanContrast enhancement using
histogram equalization
-
7/25/2019 P01 - Pengantar Digital Signal Processing
35/47
Slide 35
ECE161C: DSP II
Image processing and computer vision
Topics include:
image formation: cameras, radiometry,
and color
2D DSP, Discrete Cosine Transform Filtering, Edge detection
Multiscale representations, texture
Least squares model fitting, motion
Statistics
Principal components and face detection
Video compression
Quantization of 24-bit true color
-
7/25/2019 P01 - Pengantar Digital Signal Processing
36/47
Slide 36
Quantization of 24-bit true color
down to 8-bit color:
24-bit color original 8-bit color quantized version
A little random dithering would help mask the quantization banding on the right
ECE 172A Introduction to
-
7/25/2019 P01 - Pengantar Digital Signal Processing
37/47
Slide 37
ECE 172A Introduction to
Intelligent Systems
Main topics covered in the courseinclude:
1. Introduction to Intelligent Systems andSensor-based Robots
2. Model-Based approach in perception
3. Image segmentation
4. Edge Detection5. Region growing
6. Texture analysis
7. Object recognition and imageunderstanding
8. Extraction of 3-dimensional cues: passiveand active approaches
Project (about 5 weeks long):1. Vehicle Detection & Re-identification
2. Person Detection and Tracking
3. Robust Image Classification
Input Image
Image Classified as a
Day Image
-
7/25/2019 P01 - Pengantar Digital Signal Processing
38/47
Slide 38
Adaptive Filtering
Self-learning: Filter coefficients adapt in responseto training signal.
W(z) +x(n)y(n)
e(n)
d(n)
)()(2)()1( nnenn x
Filter update: Least Mean Squares (LMS)algorithm
d i il i li i
-
7/25/2019 P01 - Pengantar Digital Signal Processing
39/47
Slide 39
Adaptive Filtering Applications
Echo cancellation (telephone lines)
Used in modems (making Internet access possible!!)
Acoustic echo cancellation
Hands-free telephony
Adaptive equalization
Active noise control
Medical signal processing
e.g. foetal heart beat monitoring
S O h A li i A
-
7/25/2019 P01 - Pengantar Digital Signal Processing
40/47
Slide 40
Some Other Application Areas
Image analysis, e.g: Face recognition, Optical Character Recognition (OCR);
Restoration of old image, video, and audio signals;
Analysis of RADAR data;
Analysis of SONAR data;
Data transmission (modems, radio, echocancellation, channel equalization, etc.);
Storage and archiving;
Control of electric motors.
DSP D i & A hi
-
7/25/2019 P01 - Pengantar Digital Signal Processing
41/47
Slide 41
DSP Devices & Architectures
Selecting a DSP several choices: Fixed-point;
Floating point;
Application-specific devices
(e.g. FFT processors, speech recognizers,etc.).
Main DSP Manufacturers:
Texas Instruments (http://www.ti.com)
Motorola (http://www.motorola.com)
Analog Devices (http://www.analog.com)
T i l DSP O ti
-
7/25/2019 P01 - Pengantar Digital Signal Processing
42/47
Slide 42
Typical DSP Operations
Filtering Energy of Signal
Frequency transforms
1
0
)()(L
ii inxany
for (n=0; n
-
7/25/2019 P01 - Pengantar Digital Signal Processing
43/47
Slide 43
Traditional DSP Architecture
X RAM Y RAM
Multiply/Accumulate
Accumulator
x(n-i) ai
y(n)
N.B. Most modern DSPs have more advanced features.
DSP t EPSON
-
7/25/2019 P01 - Pengantar Digital Signal Processing
44/47
Slide 44
DSP at EPSON
EPSON Scotland Design Centre develops a
broad range of technologies to minimize
power consumption and maximize cost
effectiveness in mobile DSP applications.
Energy-saving Firmware
SDC C Skill
-
7/25/2019 P01 - Pengantar Digital Signal Processing
45/47
Slide 45
SDC Core Skills
System Integration
System modelling
DSP
Speech Testing
Speech enhancement
Speech synthesis
Speech Recognition
Speech
PerformanceAssessment
MP3
Audio
Channel coding
Basebandprocessing
Mobile
Computer&
Networking
CAD Tools
Administration
Services
System on Chip (SoC)
Firmware design
AMR Coding
Speech compression
CPU (Oak, ARM)
H/w & S/wCo-design
Other digital audio
SDC Firmware Development
-
7/25/2019 P01 - Pengantar Digital Signal Processing
46/47
Slide 46
SDC Firmware Development
Algorithm
Definition
Floating-pointand
Fixed-pointCo-Simulation
Co-Design
Implementation Co-Verification
COSSAPMatlab ...
Product Development
Behavioural,RTL, Logic ...
With Barcelona and TokyoDesign Centres
MCU, DSP ...
Summary & Conclusions
-
7/25/2019 P01 - Pengantar Digital Signal Processing
47/47
Slide 47
Summary & Conclusions
DSP used in a wide range of everyday applications Looked at:
Speech coding; Speech synthesis & recognition;
Image/Video;
Adaptive filtering.
Other areas include:
Image analysis (e.g. face recognition, OCR, etc.);
RADAR/SONAR;
Data transmission and reception;
And many more..!!