Signal Processing Application

For this experiment, we had to select a signal processing application and collect the information related to it using published papers and patents. We formed a group of three and the application we selected was Seismic signal processing. The group members are as follows: Ankita Dhavale, Kanchan Sawant, Nivedita Petkar.

I selected the the below mentioned paper and my review is as follows:

IEEE paper review:

A geophysical survey collects data in the form of reflection seismograms (also called seismic traces) which are received at the surface of the ground due to sources (such as explosions, vibrations, thumping) activated at the surface of the ground.

Seismic Signals are statistical in nature with non-stationary character. A new technique signal processing called wavelet analysis has been employed to obtain the better resolution for the detection of thin layer and to provide improved data for stratigraphic interpretation. Wavelets are powerful signal processing tools that have found applications in a broad spectrum of geophysical signal processing and imaging. Compactly supported non orthogonal wavelets do not have phase distortion problem and provide a better choice for seismic data processing.

DSP Processor

In this experiment, we studied and performed different operations on the DSP Processor TMS320F28375. The software for implementing the code was Code Composer Studio.
Various arithmetic and logical instructions like ADDB, SUBB, MPYB, AND, NOT are used. The various shift operations used are LSL, LSR, ROR, ROL.

FIR Filter Design Using Frequency Sampling Method

In this experiment, we have designed FIR filter using Frequency Sampling Method and plotted the frequency response of low pass and high pass filter.
Ripples are observed in stop band with decreasing amplitude. It is observed that the phase plot is linear and similar for both LPF and HPF with same order. The output is less distorted.
https://drive.google.com/drive/folders/0B3FKfBjwMZ_EdU9QeldQV25xS0k

FIR filters can be designed using different types of window functions. In this experiment, we have designed the filter using Hanning filter.
A window function is used to truncate the infinite samples of h_d(n). Phase plot observed was linear and so the output of filter was a delayed version of input.
https://drive.google.com/drive/folders/0B3FKfBjwMZ_EQ2drdmF6eEx4MXM

Chebyshev filter

In this experiment also, we plotted the magnitude and pole zero plot of Chebyshev-1. This experiment was also performed using Scilab.
The characteristics show the ripple behavior in pass band and monotonic in stop band. The observed specifications from plot matched the theoretical ones.
https://drive.google.com/drive/folders/0B3FKfBjwMZ_ERlVyazdWTThPdjQ

Butterworth Filter

In this experiment we have plotted the characteristics of Butterworth low pass filter and high pass filter. Also we analysed the output response of the respective filter. The code is implemented using Scilab.
The input specifictions such as Ap, As, Wp, Ws were verified theorectically and graphically.
https://drive.google.com/drive/folders/0B3FKfBjwMZ_EMGRkZXR0ZFFPZUE

OAM and OSM

In this experiment, we have performed linear convolution using Overlap Addition Method (OAM) and Overlap Save Method (OSM).
In this we have decomposed the input signal in parts and then output is computed.These methods are suitable to find linear convolution of long input sequence.
https://drive.google.com/drive/folders/0B3FKfBjwMZ_ERDZWdnp4bWRpOFU

Fast Fourier Transform

In this experiment, we have performed 4-point FFT and 8-point FFT. Then we have compared FFT with DFT.
We found that the number of complex multiplications and additions as well as real multiplication and addition of FFT are less than that of DFT. Hence by using FFT, the computation is reduced as compared to DFT which, in turn, consumes  less time.
https://drive.google.com/drive/folders/0B3FKfBjwMZ_ELWVjRjNpekl1Mk0

Discrete Fourier Transform

In this experiment, we have found the 4-point DFT and 8-point DFT and analysed the spectrum of the utput signal.

We have analysed that as the values are increased in 8-point DFT, the quality of the spectrum is improved.

We have concluded that as the order N increases, the frequency spacing reduces, which in turn reduces the approximation error in the representation of spectrum.

 https://drive.google.com/drive/folders/0B3FKfBjwMZ_EVHd2Rm1qeUVBd1U

Convolution and Correlation

We have started with the basics of DSPP that is convolution and correlation. We have encoded the methods and data in C language and have calculated the precise values which are approximately closer to the theoretical values.

Initially, we entered the length of input x(n) and impulse response h(n). Then we made the function to calculate the output signal y(n).

We concluded that, in linear convolution, if both the input signals are causal, then the resultant output signal is also causal.

Correlation gives an output signal which is both sided and gives us the degree of similarity between the inputs. Correlation of a signal with itself is called auto correlation its output is always a palindrome.

https://drive.google.com/drive/folders/0B3FKfBjwMZ_EOVJyVHBsOUlKcEE