Signal Processing Review

Signal Processing area focuses on issues regarding the efficient processing and transmission of real-time data. Some examples of sources of data include sound, images, and sensor output signals. Signal processing algorithms deal with efficiently transforming the signals resulting from these sources into digital data streams. 
The first thing we learned in this course was a difference between normal microprocessors and DSP processor. We worked on board which had  TMS 320f2833  processor designed by Prof. Y.S. Rao. We used CCS v6 and CCS v8. Lab Sessions included a theory session of 15-20 minutes followed by a practical session. We initially used C codes and then converted them to apply them on board on CCS software. For some programs, we used the actual board and for some we used simulation. Since simulation requires version 6 of software,  (Prof. K.T Talele) sir provided us with is own laptop to work on. We simulated Phase Modulation techniques learned in Digital Communication. We also reviewed a patent in the last experiment to study the practical application of DSP.
The best part of the entire course was an efficient way of setting targets and uploading the experiments well in time. This created a discipline in submission. The targets were challenging and required in-depth knowledge of the subject. Thus, this course was a great learning experience.

LED Binary Counter

There are 4 LEDs on board. Thus, we can make a 16 bit binary counter.  The leftmost LED is MSB and Rightmost is LSB. The programs were implemented on board. The results were observed.

Application of Signal Processing

For the application of Signal Processing, a patent was studied. The Application that we choose was an Equalizer. A graphic audio equalizer alters an audio frequency response in selected frequency bands using a plurality of filters of variable level, fixed frequency band, and fixed center-frequency frequency. A graphic audio equalizer for altering an audio frequency response in selected frequency bands using a plurality of filters of variable level, fixed frequency band, and fixed center frequency.

Sensor Interfacing

Sensors and Sensor Interfacing are an important part of Signal Processing. Microphone, LDR are few sensors on an external board which is connected to our DSP board. These sensors give data which is processed on our board and is then analyzed. This practical was indeed a learning experience.

ADC-DAC

All practical signals are analog in nature. To process them digitally we require ADC and to give them back analog signals we require DAC. This experiment gives a brief overview of the ADC DAC interfacing. The results were observed on an oscilloscope. It was an informative experiment.

Linear Filtering using OAM/OSM

OAM is Overlap Add Method and OSM is Overlap Save Method. Here we added the last two values of the first divided signal and the first two values of the next divided signal. Then we got the output where we added the signals accordingly. Thus we conclude that the input signal represents a near infinite length signal which is broken down into smaller signals for proper calculation. The Board was used and results were observed.

Discrete Fourier Transform and Fast Fourier Transform,

DFT and FFT are methods to convert a signal from time to frequency domain. Fast Fourier transform is slightly complicated but makes the work easy as it takes less time to arrive at the results. We had learned this is Signals and Systems hence it was easy to understand the working. The results were tested on board. Snapshots were taken and uploaded.

Simulator

Code composer studio provided an option of simulating the code on the Computer without uploading on Board. This feature was discontinued after version 6. This feature was used to simulate Phase Modulation Techniques. BPSK, DPSK, QPSK and Harmonic Generation was implemented. Respective graphs were plotted to observe visual results.

Correlation

Correlation is a technique to compare two signals. It is classified as autocorrelation and cross-correlation. We had a basic idea of this as this was done in the previous semester theory lectures.
Karl Pearson's Coefficient of correlation was found using FFT magnitude.  The code was implemented on board. Snapshots were taken at regular intervals.

Convolution

We were introduced to Linear and Circular Convolution. In this experiment, we first tried to write a program in C language. Then the program was translated to board using required syntax and most importantly removing all the scanf and printf statements. We got a lot of errors but rectified it. The overall experience was enjoyable yet tiring.