Classification of Prosodic Events using Quantized Contour Modeling

We present Quantized Contour Modeling (QCM), a Bayesian approach to the classification of acoustic contours. We evaluate the performance of this technique in the classification of prosodic events. We find that, on BURNC, this technique can successfully classify pitch accents with 63.99% accuracy (.4481 CER), and phrase ending tones with 72.91% accuracy.

[1]  Andrew Rosenberg,et al.  Automatic detection and classification of prosodic events , 2009 .

[2]  Julia Hirschberg,et al.  Evaluation of prosodic transcription labeling reliability in the tobi framework , 1994, ICSLP.

[3]  Paul Taylor,et al.  The rise/fall/connection model of intonation , 1994, Speech Communication.

[4]  Julia Hirschberg,et al.  Detecting Pitch Accents at the Word, Syllable and Vowel Level , 2009, NAACL.

[5]  Xuejing Sun,et al.  Pitch accent prediction using ensemble machine learning , 2002, INTERSPEECH.

[6]  Shrikanth S. Narayanan,et al.  Fine-grained pitch accent and boundary tone labeling with parametric F0 features , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[7]  Rudi C. Villing,et al.  Automatic Blind Syllable Segmentation for Continuous Speech , 2004 .

[8]  Mari Ostendorf,et al.  TOBI: a standard for labeling English prosody , 1992, ICSLP.

[9]  Mari Ostendorf,et al.  Prediction of abstract prosodic labels for speech synthesis , 1996, Comput. Speech Lang..

[10]  Gina-Anne Levow,et al.  Context in multi-lingual tone and pitch accent recognition , 2005, INTERSPEECH.

[11]  Gina-Anne Levow,et al.  Unsupervised and Semi-supervised Learning of Tone and Pitch Accent , 2006, NAACL.

[12]  Julia Hirschberg,et al.  Discourse Structure in Spoken Language: Studies on Speech Corpora , 1995 .

[13]  P Taylor,et al.  Analysis and synthesis of intonation using the Tilt model. , 2000, The Journal of the Acoustical Society of America.