Activity 3.3 - Sound signal analysis

Interpreting the recorded speech waveforms
Determining formants of the recorded speech waveforms by using sound signal analysis tools (e.g. Fast Fourier Transform or Liner Prediction)
Understanding a spectrogram

A sound sensor, interface and software that displays sound waveforms and allows to perform sound analysis (e.g. CMA Coach 6 or Raven Lite program which also produce Sound spectrogram)

In this activity students use a sound sensor to record the human voice sounds. Software for recording and analyzing sound is required. It can be for example:

Coach 6 software together with a sound sensor and interface, or
Raven Lite program, an application originally developed for use by ornithologists studying bird sounds, can be downloaded at: http://store.birds.cornell.edu/Raven_Lite_p/ravenlite.htm and used for free, or
Java simulation: Model of the Human Speech Production, which includes as well sound analysis as sound synthesis, available at: http://www2.spsc.tugraz.at/add_material/courses/scl/vocoder/simulation.html

By using Signal analysis tools like Fourier Transform or Linear Prediction students can display the frequency spectrum of the recorded sounds and find the frequency components.
Analyzing a simple sine wave signal produced by a tuning fork is especially illuminating, since frequencies can be calculated in different ways and compared.
Further students investigate sounds of different vowels and determine the formants of these vowels.

How could we recognize a vowel from its formants?
Which frequencies determine the voice differences between a male and a female ‘aa’-vowel?
Which frequencies determine the voice differences between a ‘aa’-vowel and ‘oo’-vowel said in the same pitch.