ECED 4502 Digital Signal Processing Jacek Ilow Dalhousie University Lecture 1 1 Summary of Lecture Introduction Concepts Analog/Digital Systems Digital vs. Analog processing Hot applications and recent developments Basic operations on sequences Basic sequences Lecture 1 2 Introduction Discrete: – Only takes discrete values (as opposed to continuous) Time – The independent variable – Could be other than time (e.g., space), but we will use the word time to mean the independent variable Lecture 1 3 Introduction ... Signal – A function that conveys information – Domain: One or more independent variables We will look at signals with discrete domains This means that the signal is only defined on those discrete points – Amplitude of signal Can be either discrete or continuous Lecture 1 4 Introduction ... Processing – We will look at systems which take in a signal as input and produce another signal as output – Systems can also be classified in the same way as the signals Continuous time systems Discrete time systems Lecture 1 5 Concepts Continuous time signals: f(t) defined for all values of t, where t is a continuous variable Discrete time signal: fn=f(n) defined only for discrete (e.g., integral) values of n Sequences: Representation of a discretetime signal as a sequence of numbers, fn Lecture 1 6 Concepts... Digital signal: both domain (time) and amplitude of signal are discrete Analog signal: continuous in both time and value Lecture 1 7 Analog/digital systems Analog signal Digital signal Discrete signal Analog SP A/D converter QuantizSampler er Fs≥ 2.Fmax Error is Analog signal introduced -voltage -speech -pressure or -tape data from -simulations -digital devices DSP D/A Digital Signal Processor -digital computer -dedicated dig. hw -programmable hw Digital signal Lecture 1 8 Issues Reconstruction accuracy – Conditions for perfect reconstruction Digital signal is not just an approx. representation of an analog signal – Could be generated digitally – The processing being performed may not be realizable in analog The theory of discrete time signal processing is independent of continuous Lecture 1 9 Digital vs. analog processing DSP implementations are flexible, programmable and modular More precise and repeatable Performance and cost effectiveness (riding the microelectronics wave) Direct mapping of mathematical expressions with less approximation possible (enables sophisticated algorithms) Lecture 1 10 Digital vs. analog ... Digital hardware can be multiplexed better than analog. Allows integration of multiple operations and services on a h/w platform Digital storage is more reliable, cheaper and more compact Lecture 1 11 On the other hand Analog SP still offers higher bandwidth Higher dynamic range Can be very low power Lecture 1 12 Hot application in DSP Entertainment: – digital audio – digital TV – multimedia voice mail, video and audio conferencing Medical imaging/ data analysis Information services, databases Meteorology Lecture 1 13 Hot apps... Non-destructive testing of materials Digital process control for manufacturing Automotive electronics Sensor data transmission, storage, reconstruction Voice and image synthesis Environmental monitoring Lecture 1 14 Developments in DSP DVD – CD-ROM size – Order of magnitude higher data storage – Same disk can store movies, audio, or data Very low bit-rate audio coding – Internet telephones Scalable from video coding a few Kbps to several Mbps Lecture 1 15 Basic operations on sequences Multiply two× sequences z=x.y is defined as z[n] = x[n].y[n] Add two sequences z =x.y is defined as z[n] = x[n]+y[n] Multiply a sequence by a constant z=αx is defined as z[n] = αx[n] Shifting by n0 z[n] = x[n−n0] Lecture 1 16 Basic sequences Unit sample sequence δ[n]= 0, 1, { n≠0 n=0 1 0 Lecture 1 17 Basic sequences... Unit step sequence u[n]= 0, 1, { n<0 n≥0 1 0 Lecture 1 18 Basic sequences... Sinusoidal sequence x[n]=Αcos(ω0 n+φ) 0 Lecture 1 19 Basic sequences... Exponential sequence 0<α<1 x[n]=Aαn, A 0 x[n]=Aαn, α>1 A 0 Lecture 1 20 Exponential sequence If α and A are complex α = |α|ejωo and A= |A|ejφ x[n] = Aαn = |A|ejφ|α|nejωon = |A||α|nej(ωon+φ) = |A||α|ncos(ωon+φ) + j|A||α|nsin(ωon+φ) ωo is called the frequency, φ the phase Lecture 1 21 Sinusoids and exponentials A signal with frequency ωo+2πk is identical to one with frequency ωo Exponential signal – Ae j(ωo+2πk) n = Ae jωone j2πkn = Ae jωon Sinusoidal – Acos((ωo+2πk)n+φ) = Acos(ωon+φ) We only need to consider an interval of length 2π (e.g.,−π< ωo≤ π or 0≤ωo<2π) This is not true for continuous time signals Lecture 1 22