DSPP BLOGS

Patent name: Multi-rate Digital filter for audio sample-rate conversion US Patent No: 6,487,573 B1 This patent emphasises on method for providing sample rate conversion filter on an input stream of sampled data provided at a first rate, to produce an output stream of data at a second rate different from the first rate. The input stream of sampled data is operated on with a first low order interpolation  filter routine to produce first intermediate data  The stream of intermediate data is operated on routine, having a substantially small number of operations to calculate the to produce a second stream of intermediate data. The Second Stream of intermediate data is Operated on  routine to produce the output stream of data.

This blog presents normal discussion about the architecture of dual mode audio reconstruction filter which can be used for AC-3 and MPEG.It utilises the real FFT (Fast Fourier Transform), the efficient MPEG windowing. The digital audio which is an integral part of any video or multimedia application substantiates the need for an efficient compression algorithm due to its considerable consumption of the overall bandwidth. The MPEG audio coding algorithm is the first international standard for the compression of audio signals. Both MPEG and AC-3 make no assumptions about the source of the input signal, and they can be applied to both speech and high fidelity audio. Link:  http://ieeexplore.ieee.org/document/628679/

The aim of this experiment was to perform operations using DSP Processor.  We learned programming using DSP Hardware. In this experiment, we took two, 4 point signal values performed addition, subtraction, multiplication and division using Processor. We also performed bitwise logical operations and shifting of signal values. After performing each of these operations we verified the register values and compared them with the values of these registers before the execution of these instructions. We observed the practical implementation of the processor and its instructions. 

The aim was FIR Filter design using Frequency Sampling Method. We designed the digital filter using frequency sampling method. The input specifications were given as For LPF / HPF filter Design : (1) Pass band Attenuation (Ap) (2) Stop band Attenuation (As ) (3) Pass band Frequency (Fp) in Hz (4) Stop band Frequency (Fs) in Hz (5) Sampling Frequency in Hz  For BPF / BSF filter Design : (1) Pass band Attenuation (Ap) (2) Stop band Attenuation (As ) (3) Pass band Frequency (Fp1, Fp2) in Hz (4) Stop band Frequency (Fs) in Hz (5) Sampling Frequency in Hz        In this experiment we observed that phase response will be same for low pass and high pass filter if the orders are kept same. We also verified the values of Ap and As.

The aim was Linear Phase FIR Filter design using window function. In this experiment we designed a digital filter using windowing technique and studied the spectrum of the filter. The input specifications were given as For LPF / HPF filter Design : (1) Pass band Attenuation (Ap) (2) Stop band Attenuation (As ) (3) Pass band Frequency (Fp) in Hz (4) Stop band Frequency (Fs) in Hz (5) Sampling Frequency in Hz  For BPF / BSF filter Design : (1) Pass band Attenuation (Ap) (2) Stop band Attenuation (As ) (3) Pass band Frequency (Fp1, Fp2) in Hz (4) Stop band Frequency (Fs) in Hz (5) Sampling Frequency in Hz    F or the phase spectrum we concluded that it is linear for FIR filter. Also the observed values of As and Ap are close to the input values

The aim of this experiment was designing analog and digital Chebyshev Filter .  In this experiment we designed a digital Chebyshev filter from analog Chebyshev filter using BLT. The input specifications were given as (1) Pass band Attenuation (Ap)  (2) Stop band Attenuation (As ) (3) Pass band Frequency (Fp) in Hz (4) Stop band Frequency (Fs) in Hz (5) Sampling Frequency in Hz In the end we concluded that in both low pass and high pass filters, poles are inside the unit circle and hence they are stable. For low pass filter there is a definite zero at z=-1 while for high pass filter there is a definite zero at z=1. The values of Ap and As as input are approximately same.

The aim of this experiment was designing analog and digital Butterworth filter. We designed a digital filter from analog filter and studied the aliasing effect due to sampling in Impulse Invariant Method and the frequency warping effect in BLT Method. The Input Specifications were given as (1) Pass band Attenuation  (2) Stop band Attenuation  (3) Pass band Frequency (4) Stop band Frequency  (5) Sampling Frequency        We arrived at a conclusion that for both low pass and high pass filter, poles lie inside the unit circle. Hence both the filters are stable. But we also observed in the result that the values of Ap and As are not approximately same. Hence for better stability the order of the filter needs to be increased.

The aim   of this experiment was to perform filtering of Long Data Sequence using Overlap Add Method and Overlap Save Method. We implemented the filtering of Long Input Sequence using Overlap Add / Overlap Save Algorithm. The Input Specifications were given as length of long data sequence and signal values and length of impulse response M and Signal values. We concluded that Overlap Add Method(OAM) and Overlap Save Method(OSM) are efficient methods to calculate the convolution between long length signal and finite impulse signal.

The aim of this experiment was to perform Fast Fourier Transform. In this experiment, we developed a program to perform FFT of N point signal. For this we gave input specifications as length of signal N and signal values.  After performing this experiment we concluded that from perspective of arithmetic computations, the number of arithmetic calculations in FFT are less than DFT.  Therefore, FFT is preferred over DFT.

The aim of this experiment was to perform Discrete Fourier Transform. Initially, we studied the theory of DFT and solved a sum for the same. Using these formulae and results, we proceeded with the experiment. We developed a function to perform DFT of N point signal and concluded on the effect of zero padding on magnitude spectrum.  We found that as N increases, frequency spacing reduces, approximation of error in representation of spectrum decreases and resolution of spectrum increases. Therefore, the visual appearance of the spectrum increases.

The aim of this experiment was to study mathematical operation such as Linear convolution, Circular convolution, Linear convolution using circular convolution.We developed a function to find Linear Convolution and Circular Convolution We then concluded on aliasing effect in Circular convolution. The aim of the second part of the experiment was to study mathematical operation of correlation and measure degree of similarity between two signals. We wrote a function to find correlation operation.                                                                               We calculated correlation of a DT signals and verify the results using mathematical formulae.