Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

What to study to learn more about digital signal processing?

Status
Not open for further replies.

Always_Confused_Sudent

Junior Member level 3
Junior Member level 3
Joined
Dec 26, 2003
Messages
28
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
321
What to study..

Hi there...I'm an undergraduate student and I had a digital signal processing course that covered the basic topics like DFT,FFT,Ztransform,.. It was just plain theory , I still don't know the applications of all the theory we studied...I need some suggstions on what to focus on and master, and what useful projects can I do to understand the idea of digital signal processing because it seems interesting.I have some experience using PICmicro's,so is it possible to implement digital signal processing ideas using a PICmicro ? I also know some MATLAB.
 

What to study..

You may implement digital filter using PIC microcontroller.
**broken link removed**

For Matlab:
Practical FIR Filter Design in MATLAB
**broken link removed**

Fingerprint Recognition System
Source code is available.
**broken link removed**

JPEG Compression
An example for jpeg compression and the restoration.
https://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=4328&objectType=file

DTMF decoder
Displays the input waveform of a DTMF and determines the High and Low frequency content of a DTMF.
https://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=3785&objectType=file
 
What to study..

I think the fundamental of DSP theory is the <signal and system>, I suggest that you should focus on understanding the essential of FFT, including the transform between time and frequency, periodicity and discreteness. that will help you to master DFT.

In my opinion, all the dsp do is to filter and for the filter. So if there is anything you must master, that should be the filter design. Fortuenately, now we can design a digital filter easily using the computer tools such as Matlab, so you need only to know the difference between different filter design methology, and no need to touch too deep in the mathematica principle.

Regards!
 

Re: What to study..

u should understanding the fft and dft, and it also involves in alot of mathematic, but it's ok if u know MATLAB.
 

Re: What to study..

hi,

i would recommend u to go thro this one.. which is the best book i have known till now,, i would say! its a Wiley-IEEE press and would contain all the concepts u learnt in ur course but the amount of info u get from this book... it s immense than u learnt from ur course and also it self-readable!



and for matlab.. i would recommend the publications by "orchard" signals and systems using matlab is the book. u could do a search and get it with terms "signals and systems orchard" or more relevant strings.. these 2 combinations could build ur confidence and concepts in this field i beleive!!! go for it!

/Am
 

What to study..

Hi, I would recommend you concentrate on different transforms in DSP . Go further than just DFT and FFT maybe into AFT, Gabor, DCT , DST etc.. You could then work in the image processing field where these transforms are applied. My suggestion is that do a lot of ground work since these are all mathematical concepts and hence require a very strong basic understanding of the fundamentals
Regards :)
AG
 

What to study..

DFT and FFT are hard for students for they include lot of mathematical concepts, but they are importance for understand the way of DSP's work.
 

Re: What to study..

Hi Always_Confused_Sudent,

Many students who first encountered DSP for the first time get a shock. Some still remain daunted even after the course. LoL! So don't be bogged down.

Let me put it in a simple language.

DSP is simply a time-based technique instead of the analog-frequency technique to tackle real-world signals which are all analog, despite your square waves, digitally-modulated signals and so on are "thought to be digital", they are actually analogous. It's the phases, voltage/current levels and polarities and that make analog to be digital.

A good start is to pick a book on speech processing where you try understand how to design digital audio filters to extract speech signals from background noises.

Advanced levels would be to study on video processing, image processing, coding and compression.

I presume you have background on Fourier Transforms, DFTs, FFTs and basics in Matlab, which has many simulink tools to help you simulate a design, or you can use the commands to analyse the signals.

If you have foundations in State-Space Analysis and Linear Algebra, you can perform matriculated calculations of filters in a very novel way where eigenvalues come in, Z-transform in this case appears much clearer to compare designs in Analog and Digital.


PhD (Imperial College London)
Analog Devices Plc (Ireland)
 
Re: What to study..

DSP is an interesting subject. The best way to study it is to understand the BIG PICTURE first.
There are several ways to characterize a linear system which are difference and differential equations,impulse response,step response,complex frequency domain and the transfer function.There are relationship among these characterizations, for example ,you can get the frequency domain representation of the system described by a difference equation by performing the z-transform.
Various characterizations help you understand how the system behaves in a different point of view.For instance the spectrum you obtain by performing DTFT ,The Z plane and the DFT spectrum , each has it's own importance.The DTFT is where we usually start from,DFT is a sampled version of the DTFT, the Z plane gives you information about stablity and causality of the system.

My point here is understand the general concepts first , then study the mathematics of it in more details when you feel that you are ready.

Regards
 
Re: What to study..

Always_Confused_Sudent said:
Hi there...I'm an undergraduate student and I had a digital signal processing course that covered the basic topics like DFT,FFT,Ztransform,.. It was just plain theory , I still don't know the applications of all the theory we studied...I need some suggstions on what to focus on and master, and what useful projects can I do to understand the idea of digital signal processing because it seems interesting.I have some experience using PICmicro's,so is it possible to implement digital signal processing ideas using a PICmicro ? I also know some MATLAB.
Hi,
The concepts in DSP are used whenever you interface a digital system to an analog system, which in short is everywhere almost. So things that we could simply do with anlog components like inductors, capacitors and resistors in analog domain is possible to be duplicated in digital domain on digital signals and this is primarily preferred method due to cost, reproducibility and performance. So you see DSP applications in all spheres of processing except in extremely high speed/frequency applications. The typical examples where DSP are employed being speech processing in telecommunications equipment, ultrasound and MRI data processing in medical sciences and many induatrial and military applications and last and most important in cosumer electronics for multimedia signal processing. they are everywhere in short.
If you are thorough with the basics of the DSPs you studied then pick up an applications book like speech processing or synthesis or spectral estimation or filtering in communications systems. it would be good if you could get to understand the proakis book and the book signal processing with matlab will be very handy for you since you already know some matlab.

hope this helps.
 
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top