|Talking Heads: Speech Production|
Measuring and Modeling Speech Production
Figure 3: The source-filter model for selected vowels.
vocal tract shapes are shown corresponding, from top to bottom, to the vowels
"ah" (/a/), "ee" (/i/), and "oo" (/u/). In this
example, we have used schematized vocal tract shapes from the
Haskins Laboratories articulatory synthesizer. Plotted in the
same graph for each tract shape is the smoothed transfer function that is
computationally derived by the synthesizer, and the hypothetical energy spectrum
that would result from using these functions to filter a glottal source spectrum
with a fundamental frequency of 100 Hz. Note that although all three vowels have
the same fundamental frequency, their spectra differ according to the filter
characteristics of the different vocal tract shapes.
Tract Model | Gestural Modeling | State of the Art