Before starting this module, you may want to look at the optional foundation material on Signals.
This module introduces a simple model of speech production will be helpful throughout the rest of the course. We’ll be using it as a theoretical concept to help us understand how speech signals are made, and therefore how we can work with them in various ways. We’ll also be seeing that is can be actually implemented and used for generating synthetic speech.
Both the theoretical concept and the practical use of this model will become very important at several places later in the course.
Download the slides for the module 1 and 2 videos
Download the additional slides for the class on 2019-10-03 : Foundations – Signals
Download the additional slides for the class on 2019-10-03 : Module 2
Total video to watch in this module: 15 minutes
Here is a spreadsheet that implements a very simple source-filter model. The source is an impulse train, and you can change its fundamental period. The filter has three coefficients, which you can also change. Explore these parameters and see what effect they have on the output from the filter.
- Apple Numbers format which is preferred because it provides sliders to change the values
- Excel format in which you need to change the values by typing new values the cells. Try values for the fundamental period between 2 and 50, and values for the filter co-efficients between -1 and 1.
Once you understand how it works, test your understanding by extending the filter to have 4 co-efficients.
Links
- There is some great material from Rice University on how the auditory system works.
- A Java applet to generate any signal by summing together sine waves
After finishing this module, you may want to look at the optional foundation material on Phonetics.
