Prominentie: Akoestische, lexicale en syntactische correlaten
|Title||Prominentie: Akoestische, lexicale en syntactische correlaten|
|Year of Publication||2002|
|Conference Name||Afgeronde Projecten|
|Publisher||Nederlandse Vereniging voor Fonetische Wetenschappen|
|Conference Location||Amsterdam, The Netherlands|
In my talk both acoustical and lexical/syntactic correlates of prominence are discussed. Prominence is defined at the word level and is based on listener judgments. A selection of useful acoustic input features is tested for classification of prominent words, with the help of Feed Forward Nets. Spoken sentences from many different speakers, taken from the Dutch Polyphone corpus of telephone speech, are used. For an independent test set of 1,000 sentences about 79% of the words are correctly classified whether or not as prominent. At the text input level we also developed an algorithm, using linguistic/syntactical features derived from text only, to predict prominence. The prediction agrees with the perceived prominence in 81% of the cases for the independent test set. It can be concluded that, naive listeners are able to mark prominence. The results of this thesis show that acoustical and linguistic correlates of prominence can be determined automatically, and that these acoustical correlates can be used to accurately predict prominence on the basis of only 12 appropriate features. Prominence assignment of naive listeners is valuable because the determined acoustical and linguistic correlates are able to predict prominence. Agreement measures show that prominence prediction is undistinguishable from the prominence assignment of naive listeners.