Basic operations using DSPP Processor

Unlike previous experiments which were performed in Scilab or  through C language , this experiment was a hardware based demo experiment. Instructions for arithmetic, Logical and Shift operations were performed on the processor kit.The changes in the register values before and after execution were noted. The processor is capable of performing many such complex operations. Real world scenarios require the use of proper hardware for implementation. A wide range of processors are available for signal processing and a selection of board should be made taking into consideration all the parameters of the system to be implemented.

FIR Filter Design: Frequency Sampling Method

This is the second method of FIR Filter design. The steps of finding the frequency response of the input is same as that of the windowing method. Then, DFT of the frequency response is calculated. The values are adjusted depending on the type of filter i.e. low pass filter or high pass filter. Inverse of the frequency sampled signal gives the desired filter design. The designing and plotting of magnitude response was done on Scilab. 

Research Paper Review

Paper Title:Speech Recognition using Frequency Transformations

Authors: Jorge Salomon Fuentes, Dr. Chit-Sang Tsang

Every person has different speech utterances based on age, sex, tone etc. An improvement in speech recognition using frequency transformation and pattern recognition is proposed with increased accuracy and simple hardware. The performance of the system is shown using a database of 124 words. Different models are used for the different aspects of speech recognition in such a way that their combination improves the efficiency of the system. The speech patterns of a single user was used to recognize the 124 words stored in the database.
The algorithm developed was speaker and language independent. A study to implement it in different languages is in progress.The accuracy of the algorithm was measured using a training kit   

Patent Review

ADAPTATION OF A SPEECH RECOGNITION SYSTEM ACROSS MULTIPLE REMOTE SESSIONS WITH A SPEAKER

Patent No.: US 6,766,295 B1

Inventors: Hy Murveit, Portola Valley, CA (U S); Ashvin Kannan, Redwood City, CA (Us)

Publication Date:Jul. 20, 2004

Filing Date : May 10’ 1999

A speaker can remotely access a speech recognition system. The medium of communication can be a telephone. An acoustic model is used to initialize an independent speaker in a remote session for the first time.  Upon the termination of the session, the speech data of the speaker is stored in memory with an independent identification key . During a subsequent remote session, the speaker is identified on the basis of his speech and the corresponding acoustic model is altered. New samples are taken each time the user speaks. The stored acoustic model is modified each time and the result made more accurate.

Even if the speaker uses the remote session for a relatively small amount of time, modifications are made in the acoustic model and the accuracy of recognition is improved. Thus different users can use multiple remote sessions and their speech data can be stored and modified separately. 

FIR Filter Design: Windowing Method

There are two methods that we have learnt for Finite Impulse Response filters. Windowing Method and Frequency Sampling Method. In this post I am going to talk about windowing method. In this method the input  time domain infinite signal  is multiplied with a windowing function which is finite, to obtain an FIR filter. There are different types of windowing functions. The choice of windowing  function is decided by the value of stop band attenuation.
A Scilab code was written to design an FIR filter. Hanning window was the windowing function used.The stopband attenuation was chosen accordingly. A low pass filter was designed and the observed and input values compared.   

Chebyshev Filter Design

This is an IIR filter design technique where the magnitude  response of the filter has ripples. We designed Chebyshev-I filter which has ripples in the pass band. The number of peak and valley points of the ripple gives the order of the filter.
A Low pass filter and a high pass filter were designed with the input values of stop band attenuation, stop band attenuation ,pass band frequency, stop band frequency and sampling frequency. BLT Method of digital filter design was used to find the transfer function from the analog filter response. The analog filter was designed with the help of formulae. The observed values in the magnitude response were found to be close to the input values.

Butterworth Filter Design

This is a filter design technique to design an IIR filter. The magnitude response of the filter is ripple free. A code was written in Scilab ,which is an open source software to plot the frequency response of an IIR filter with the given input specifications. An analog filter was designed with the input values and a digital filter was designed from this analog filter using BLT Method.
A low pass filter as well as a high pass filter were designed. The observed values were compared with the input specifications to check for accuracy.