Basic Intro to DSP

Basic Intro to DSP

Digital signal processing

Lets introduce ourselves to digital signal processing. It is concerned with the representation of signals by sequence of numbers or symbols and the processing of these sequences

There are three main component of DSP

  1. Signals
  2. Representation of signals
  3. Processing of these signals

A signal can be defined as a function that conveys information about the state or behaviour of a physical system for example while I’m speaking I’m generating audio signals using my vocal cord, When I say out the alphabet a I’m generating an audio signal that conveys the information of alphabet ‘A’ so mathematically the signals are represented as functions of one no more independent variables. If you take the same example of audio signals the amplitude of my voice can be plotted as a function of time i.e time on the ‘X’ axis and on the ‘Y’ axis have the corresponding amplitude at each instance of time. We get this signal here the independent variable list time ‘t’ and the signal which is my voice is plotted as a function of time, if you see this signal is changes continuously in time also it has continuous amplitudes so this is an analog signal. Another form of signals are discrete signals: they are signals defined on discrete units of time to obtain discrete signals or we have to do is sample the analog signal set discrete units of time. For instance let’s sample this analog signal, for that first let us split that time X into discrete units i.e we mark the X axis at equal distances then we obtain the amplitude of signal at this discrete times. Now if we plot only these points we get it discrete signal.

Many people often confuse analog signals with continuous time signals let me make this very clear for you analog signals and continuous time signals are not the same. Another common misconception is that discrete signals and digital signals are the same. But digital signals and discrete signals are not same.

Now let’s learn what is signal processing: signal processing involves analysing modifying and synthesising signals to pull or extract meaning out of it signal processing. It can be broadly classified into analog signal processing and digital signal processing. Analog signal processing deals with transformation of analog signals and the processing is done using electrical networks consisting of active and passive elements i.e opAmp resistors, capacitors and inductors

On the other hand digital signal processing deals with the processing of discrete signals and the processing is done by general purpose computers, or by digital circuits such as ASIC, FPGS, DSP chips etc.

DSP system consist mainly five blocks, they are pre filter , ADC, DSP, DAC, post filter. Low pass filter is filter out unwanted high frequency components from the raw analog input signal for example if you’re processing audio signals we only need frequencies in the range of 20Hz to 20KHz and this pre filter will filter out any frequencies about 20 KHz as they can cause. ADC will convert analog signals to digital signals then these signals are fed to the DSP where it is analysed and processed and the synthesised output is fed to the DAC. The DAC converts digital signals back into analogue signals and finally this post filter is used to filter out any unwanted high frequency components in the generated analog signal. Most often we also provide an amplifier at the beginning and at the end. This is to amplify the incoming and outgoing signals. Now coming to the applications we use digital signal processing in many areas:

  • It is used to process speech and audio: this involves speech recognition analysis, noise filtering, echo cancellation etc
  • It is also used to process image and video this involves compression enhancement, reconstruction and restoration of images and videos.
  • It also involves face detection like in the case of face unlocking phones this piece are also used in military and telecommunication applications for example sonar and navigation radar tracking modulation and demodulation etc
  • Next application is in biomedical and healthcare sector for example analysis of ECG and X ray signals.
  • Another application is consumer electronics most digital equipments in your home like smart phones, televisions, digital cameras etc. have a DSP embedded in it to accellerate its performance

Advantages of DSP: It can perform complicated signal analysis with relative easiness without the use of complex circuits. Also the processing can be easily modified with simple changes in software that is DST systems are flexible. Also with dsp we have access to many error detection and correction features for example parity generation and correction. Also with DSP systems data storage is easier since digital storage devices are becoming cheaper day by day storage of data is much easier in digital forms, it is also easier to transport and recreate digital data with 100% fidelity. It is low cost with advancements in VLSI technology the cost of dsp have come down by many faults than their analog counterparts.

Disadvantages of DSP systems:

They are made out of a lot of transistors and these transistors together consume a lot more power than analogue signal processes. Another problem is the higher learning curve required for the operation of DSP systems each DSP has different hardware architecture and software instructions due to this highly skilled engineers are needed to programme the device and proper training on DSP required for programming for various applications.

To view or add a comment, sign in

More articles by Kishan Rajoria

  • Working of MLOps (Part-3)

    MLOps follows a similar pattern to DevOps the practices that driver’s seamless integration between your development…

    1 Comment
  • Key questions for finding data to build ML models

    Since data is the essential ingredient to power ML algorithms, it always helps to build an understanding of the…

  • Roles and Requirement of MLOPs (Part-2)

    1. Subject matter experts a.

    1 Comment
  • Machine Learning Model Operationalization (ML Ops): Part-1

    During the industrial revolution the rise of the physical machines required organizations to systematize form factories…

    1 Comment
  • Forecasting Error

    When doing forecasting whether our forecasting model is accurate or not because forecasting is an estimation of future…

  • Exponential Smoothing model

    As we know exponential smoothing models are very efficient models of smoothing and these models help us effortless…

  • Forecasting Principles and methods

    Advanced models of time series analysis and these are known as exponential smoothing models. The name explanation…

  • Forecasting Introduction and Methods-2

    For the time series forecasting there are some fundamental requirements. Type of method you are going to use you need…

  • Forecasting Introduction & Methods-1

    Utility operation and maintenance management point of view we have to take many decisions based on forecasting…

Others also viewed

Explore content categories