Lab9:-Basic Operations on Digital signal Processor

The aim of this experiment was to perform basic operations on DSPP processor.The kit used was TMS320F28375.Basic arithmetic, logical and shifting instructions were executed on the processor.The values stored in the registers were noted after execution.It was observed that logical shifting operations perform bitwise shifting of data.Also logical operations only represent true or false value(1 or 0).

Lab8:- FIR filter design using FSM

The aim of this experiment was to perform FIR filter design is Frequency Sampling Method. Here the output h[n] is obtained by IDFT of h(w) obtained by the magnitude spectrum of filter required.When the input specifications are kept same as  previous experiment wherein FIR filter design was performed using Windowing method, it is observed that as the order increases , the number of lobes in the stop band increases.The advantage of FSM over windowing is the efficient frequency sampling structure which is obtained when most of the frequency samples are zero. Like window method, the h[n] (transfer function) obtained is symmetric.

Research paper review:Audio Compression using Fourier transform

International Journal of Science and Research (IJSR) ISSN (Online): 2319-7064 Index
Copernicus Value (2015): 78.96 | Impact Factor (2015): 6.391, Volume 6 Issue 2, February 2017
Summary:
Audio compression has been carried out by Discrete Fourier Transform to differentiate one frequency from another. Results obtained on performing DFT of an audio speech signal has been analyzed with the help of parameters like Level Detection, Framing, Windowing, Power spectrum calculation. The approximate signal is drawn by taking the maximum frequency components of the maximum power frame. The approximate signal and the original signal are compared graphically.
Paper link: https://www.ijsr.net/archive/v6i2/ART2017951.pdf
Plagiarism link: https://drive.google.com/open?id=0B06hNCrbOenyVHZtRVl5Zkctd2s

Patent Review:Mixed lossless audio compression

Patent Review
Patent No: US 8,630,861 B2
Publication date: *Jan. 14, 2014
Inventor: Wei-Ge Chen, Sammamish, WA (US); Chao He Redmond, WA (US)          
Summary:
A mixed lossless audio compression has application to a unified lossy and lossless audio compression scheme that combines lossy and lossless audio compression within a same audio signal. The mixed lossless compression configures a transition frame between lossy and lossless coding frames to produce seamless transitions. Some of the obstacles wile unifying lossy and lossless audio compression
techniques include:-
 Transition between lossy and lossles compression can introduce certain audible discontinuities in the decoded audio signal. In simple words, audible noise can be produced when switching between lossy and lossless compression.
 In lossy compression schemes, audio signal samples on an overlapped window basis, whereas lossless compression techniques do not. Redundantly coding the overlapped portion with both lossy and lossless compression may reduce the achieved compression ratio.

To tackle these obstacles, the inventor has provided a solution to divide the signal into frames
(separately for lossy and lossless) and mixed lossless frames that serve as transition between the two other types. These mixed lossless frames are compressed by performing a lapped transform on an overlapping window in the lossy compression case followed by inverse lapped transform to produce a
single audio signal frame, which is losslessly compressed.
Patent link:  https://google.com/patents/US8630861B2

Lab7:-FIR filter design using Windowing method

We can design a linear phase FIR filter using 2 methods:-Windowing and FSM(Frequency Sampling method). The aim of this experiment is to design a linear phase FIR filter using windowing method.In this method the input  time domain infinite Digital Time signal  is multiplied with a windowing function which is  a finite DT signal, to obtain an FIR filter transfer function in the form of an array of values. There are different types of windowing functions. The choice of windowing  function is decided by the value of stop band attenuation and Cn.
A Scilab code was used to design an FIR filter. Hanning window was the windowing function used .The stopband attenuation was decided accordingly. A low pass filter was designed and the observed and decided parameter values were compared.

LAB6:-Design of Chebyshev Filter

The aim of this experiment was to design a Chebyshev filter using Scilab software.Passband attenuation and frequency (Ap,Fp)  and Stopband attenuation and frequency (As,Fs) were given as input parameters to the filter design.The sampling frequency was set greater than 5 times max(Fs,Fp) .The magnitude response of the filter was plotted with frequency in Hz and Magnitude in dB.Ripples were observed in the pass band and the number of extremes in the ripples corresponded to the order of the filter.

Lab5:-Design of Butterworth Filter

The aim of this experiment was to design a Butterworth filter using Scilab software.As,Ap,Fs,Fp and sampling frequency were given as input parameters to the filter design and BLT method was used for filter design.The sampling frequency given as input was kept higher than 5 times the max(Fs,Fp). Magnitude response of the filter was plotted wherein frequency was kept in Hz and Magnitude in dB.It was observed that Magnitude response​ was monotonic (without ripple).

LAB4:-Filtering of data sequence

The aim of this experiment was to perform OAM(Overlap Add Method) and OSM(Overlap Save Method) in FIR-filter(Finite Impulse Response) to filter a sequential data input. Independent functions to perform OAM  and OSM separately were executed in C language which took length of the signal(in my case:13)as one of its arguments. It was observed that both techniques of filtering yielded the same output. Also we essentially observed that the function decomposed the input signal into blocks and computed them separately. This meant that both OAM and OSM are block processing techniques.

LAB3:-Fast Fourier Transform

The aim of this experiment was to perform Fast Fourier transform of a DT signal. A function in C was executed to perform Radix-2 Cooley and Tuckey's DITFFT algorithm which took the length of the signal as one of its arguments(length taken must be in exponential powers of 2).A counter was also implemented to calculate total number of real and complex multiplications and additions. The results obtained were verified mathematically using formulae for the same. Bit reversal was observed and the fact that FFT is a computationally faster algorithm than DFT was understood.  

LAB2:-Discrete Fourier Transform

The aim of this experiment was to perform Discrete Fourier Transform(DFT) of a DT signal. We executed a function in C to perform DFT of an N point signal by providing length of the input signal as one of its arguments. Multiplications and additions of complex terms were done separately by considering only their coefficients. The results obtained were plotted in the form of a magnitude spectrum(Magnitude vs frequency).It was observed that DFT results are periodic,inverse-DFT converges and DFT of an expanded signal results on compression of magnitude spectrum.     

LAB1:-Convolution and Correlation

The aim of the experiment was to perform linear convolution, circular convolution and linear convolution using circular convolution. We executed a C code for LC,CC and LC by CC on Terminal in Linux OS by using simple commands such as ('gcc','./a.out' etc.).The results obtained by running this code were verified by mathematical formulation. Key observations that were made were that Circular Convolution gives aliased output and Autocorrelation of input signal did not change even when it was delayed.