Speech Production

The production of speech is a complex physiological process that relies on the interplay of several organs of the vocal tract. Sounds are produced by varying the place and manner of articulation.

Phonemes can be classified based on two main aspects of articulation:

• Place of articulation: Where in the mouth is the airflow obstructed?
• Manner of articulation: How is the airflow obstructed?

Additionally, phonation or voicing distinguishes between sounds based on whether the vocal cords vibrate (voiced, e.g., /z/) or not (unvoiced, e.g., /s/).

Spectrograms are tools used to analyse and represent the speech signal. A Fourier transform is involved in creating a spectrogram, which represents waveforms (amplitude over time) as amplitudes of different frequencies over time. In a spectrogram, the x-axis represents time, and the y-axis represents frequency, with darkness indicating higher amplitudes for specific frequencies. Formants, which are the distinguishing frequency components of human speech, can be identified on spectrograms. For instance, vowels can be distinguished quantitatively by their formant frequencies, and the change of formants across time provides information about words.

To conclude, Speech production is a complex process involving various vocal organs and the precise control of articulation (place and manner) and voicing. Spectrograms are valuable tools for analyzing the acoustic properties of speech and identifying key features like formants.