Paper Info
Title
Automatic Prosody Boundary Labeling of Mandarin Using Both Text and Acoustic Information
Abstract
Prosody is an important factor for a high quality text-to- speech (TTS) system. Prosody is often described with a hierarchical structure. So the generation of the hierarchical prosody structure is very important both in the corpus building and the real-time text analysis, but the prosody labeling procedure is laborious and time consuming. In this paper, an automatic prosody boundary label system is presented, in which the classification and regression tree (CART) framework is used. In this system, we build a prosody model using acoustic information and the text information based on large speech corpus with prosodic structure label (ASCCD). Experiments show this model can achieve prosody boundary detection 90.86% accuracy.
Year
DOI
Venue
2008
10.1109/CHINSL.2008.ECP.100
ISCSLP
Keywords
Field
DocType
real-time text analysis,chinese information processing,speech corpus,acoustic information,high quality text-to-speech system,cart,acoustic- prosodic feature,trees (mathematics),regression analysis,speech synthesis,prosodic structure label,text information,index terms— prosody boundary,classification and regression tree framework,mandarin,prosody prediction,natural language processing,automatic prosody boundary labeling,real time,speech,speech recognition,hidden markov models,acoustics,text to speech,classification algorithms,text analysis,entropy,information processing,indexing terms,labeling
Speech corpus,Decision tree,Prosody,Computer science,Boundary detection,Artificial intelligence,Natural language processing,Speech synthesis,Pattern recognition,Speech recognition,Statistical classification,Hidden Markov model,Mandarin Chinese
Conference
ISBN
Citations 
PageRank 
978-1-4244-2943-1
0
0.34
References 
Authors
6
3
Name
Order
Citations
PageRank
Chong-Jia Ni1204.84
Wenju Liu221439.32
Bo Xu324136.59