This exercise enables the user to design a filter and then process an existing speech or audio file.
THE SET OF SPEECH AND AUDIO FILES HAVE TO BE DOWNLOADED AND SAVED IN THE SAME FOLDER TO WHICH YOU DOWNLOAD THE VARIOUS SPEECH PROCESSING EXERCISES IN ORDER FOR THIS PROGRAM TO WORK SEAMLESSLY. THE SPEECH AND AUDIO FILES CAN BE DOWNLOADED USING THE FOLLOWING LINK:
WE STRONGLY URGE YOU TO DOWNLOAD THE SPEECH AND AUDIO FILES BEFORE DOWNLOADING ANY OF THE SPEECH PROCESSING EXERCISES.
Speech processing designates a team consisting of Prof. Lawrence Rabiner (Rutgers University and University of California, Santa Barbara), Prof. Ronald Schafer (Stanford University), Kirty Vedula and Siva Yedithi (Rutgers University). This exercise is one of a set of speech processing exercises that are intended to supplement the teaching material in the textbook “Theory and Applications of Digital Speech Processing” by L R Rabiner and R W Schafer.
A basic signal processing operation is filtering of an existing signal using a user-designed filter. The goal of the filtering operation is to remove extraneous (unwanted) signal components at either or both the low frequency or the high frequency end of the spectrum. Such undesired signal components include dc offset, low frequency hum, and low and high frequency noise
which are often created during the speech recording process. The choice of filter type includes lowpass, highpass, and bandpass filter. This exercise allows the user to provide a set of ideal filter specifications and then design a filter which approximates the ideal filter using the MATLAB function, firpm.m, an FIR filter design algorithm. The user can examine the resulting filter in the time and frequency domains, and then filters the designated speech signal using the designed filter. The filter design and signal processing operations can be iterated as often as desired, until a desired filtering result is obtained.
For optimal results use the speech files at this link: