Fundamentals of Digital Signal Processing (DSP)
DSP is Everywhere!
Audiological equipment Hearing aids Otoacoustic systems Audiometers Aural rehabilitation programs ABRs Telecommunications Cellular phones Voice over Internet Audio CD, DVD, DAT players MP3 players Biomedical monitoring equipment Digital Television
Analog vs Digital
Analog
infinitely detailed continuous in time continuous in amplitude
Digital
discrete numbers discrete in time discrete in amplitude
Challenge: a digital representation that adequately maps the analog values.
Analog Systems
Acoustic domain Electric domain Acoustic domain
Programmable Analog Systems
Acoustic domain Electric domain Acoustic domain
Digital domain
Digital Systems
Acoustic domain Electrical domain Digital domain
Electrical domain
Acoustic domain
Why digital?
Flexibility Programmability Reproducibility and Precision (??) Advanced Signal Processing
Multichannel compression Precise frequency shaping Feedback cancellation Noise reduction Directional processing
DSP Applications
Digital signal analysis/display Digital recording, processing, and reproduction Digital signal synthesis
DSP Buzzwords
A to D Converter D to A Converter Sampling Quantization Aliasing Anti-aliasing filter Algorithm Oversampling
Sampling Process
Analog Signal
Sampling Interval (Ts)
Sampled Numbers
Analog-to-Digital Conversion
5 0
Analog
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-5 5
Discrete
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-5 5
Digital
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-5
Undersampling or Aliasing
Undersampling results in a form of distortion termed as aliasing.
http://www.dsptutor.freeuk.com/aliasing/AliasingDemo.html
How to prevent aliasing?
Nyquist Criterion: When a signal is sampled, the sampling rate must be greater than twice the highest frequency of the input signal. What if the highest frequency is unknown?
Use a low pass filter to remove unwanted frequencies. Set the sampling rate greater than twice the bandwidth of the low pass filter. The low pass filter is called an anti-aliasing filter.
Quantization
The sampled values are converted into bit representation The process is called Quantization The performance of a quantizer is dependent on the number of bits, also called bit resolution.
Binary Logical Circuits Flip Flops
#1 ON #2 OFF #3 ON #4 OFF
Each of these values is called a bit. The string of values 1010 is a binary representation. A string of 8 bits is called a byte.
An example of binary counting
20 15 10 5 0 -5 -10 -15 -20 0 0.2 0.4 0.6 0.8 1
Voltage Flip Flop #1 Flip Flop #2 5 < V < 10 on (1) on (1) 0<V<5 on (1) off (0) -5 < V < 0 off (0) on (1) -10 < V < -5 off (0) off (0)
Binary 11 10 01 00
Dec 3 2 1 0
Example (contd.)
20 15
Analog fine grain signal
Voltage
10 5 0 -5 -10 -15 -20 0 0.2 0.4 0.6 0.8 1
Decimal number
5 4 3 2 1 0 -1 -2 0
Digital quantized signal
0.2
0.4
0.6
0.8
Example (contd.)
10
Analog fine grain signal
-5
-10 0 0.2 0.4 0.6 0.8 1
5 4 3 2 1 0 -1 -2 0
Digital quantized signal
0.2
0.4
0.6
0.8
3-bit Analog-to-Digital Converter
Voltage
7.5 < V <= 10 5 < V <= 7.5 2.5 < V <= 5 0 < V <= 2.5 -2.5 < V <= 0 -5 < V <= -2.5 -7.5 < V <= -5 -10 < V <= -7.5
#1
#2
#3
on(1) off(0) on(1) off(0) on(1) off(0) on(1) off(0)
Binary Decimal
111 110 101 100 011 3 010 001 000 7 6 5 4 2 1 0
on(1) on (1) on(1) on (1) on(1) off (0) on(1) off (0) off(0) on(1) off(0) on(1) off(0) off(0) off(0) off(0)
3-bit Quantization
10 5
-5
-10 0 0.2 0.4 0.6 0.8 1
-2 0
0.2
0.4
0.6
0.8
Characteristics of an A/D Converter
Input range
The voltage range that the A/D converter can handle. Can be unipolar (either +ve or ve voltages), or bipolar (both +ve and ve)
Resolution
Represented by the number of bits. The number of possible states is given by 2N, where N is the number of bits.
Sampling Rate
Rate at which samples are measured.
Digital Signal Processing
A series of trigonometric and arithmetic operations. Series of steps called algorithm Algorithm classes
Spectral analysis Digital filtering Coding and compressing data Noise reduction Etc.
Conclusions
Digital Signal Processing (DSP) is often used in modern audiological equipment. Fundamental concepts in DSP
Sampling discretization of the time axis Quantization discretization of the amplitude axis
Digital Processing a series of instructions to manipulate the digital numbers.
