Abstract

In this thesis a data-driven and linguistically interpretable intonation model for the automatic analysis and synthesis of fundamental frequency (F0) contours was developed. The model can be characterised as parametric, contour-based, and superpositional. Its intonation representation consists of a superposition of global and local contour classes and can be derived in a purely data-driven manner, which guarantees consistency and easy adaptability to new data. The model's linguistic interpretability was examined by automatic linguistic corpus analyses resulting in hypotheses about possible relations between contours and linguistic concepts. These hypotheses were subsequently tested by perception experiments. By these means a systematic linguistic anchoring of the model was achieved in form of a decision tree to predict the linguistically appropriate contour class. The adequacy of its predictions was assured by a further perception test. Due to its simultaneous signal proximity and linguistic anchoring, the model covers the entire chain from text to signal and therefore can be used for intonation analysis and generation on a linguistic as well as on a phonetic-acoustic level. It is qualified for employment in speech technology applications as well as in phonetic fundamental research to automatically analyse raw speech data.