Automatic labeling of prosodic patterns

Wightman, Colin W.; Ostendorf, Mari

doi:10.1109/89.326607

Cited by 226 publications

(143 citation statements)

References 28 publications

Supporting

Mentioning

138

Contrasting

Order By: Relevance

“…Some of these approaches detect boundaries by analyzing the acoustic features of the input utterance, such as its energy contour, the speaking rate, and the fundamental frequency F 0 (Swerts 1997, Wightman 1994. It is true that some of the approaches take into account the linguistic content of the input utterance (Batliner 1996, Stolcke 1996 to some degree.…”

Section: Introductionmentioning

confidence: 99%

Chinese Utterance Segmentation in Spoken Language Translation

Zong

Ren

2003

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper presents an approach to segmenting Chinese utterances for a spoken language translation (SLT) system in which Chinese speech is the source input. We propose this approach as a supplement to the function of sentence boundary detection in speech recognition, in order to identify the boundaries of simple sentences and fixed expressions within the speech recognition results of a Chinese input utterance. In this approach, the plausible boundaries of split units are determined using several methods, including keyword detection, pattern matching, and syntactic analysis. Preliminary experimental results have shown that this approach is helpful in improving the performance of SLT systems.

show abstract

Section: Introductionmentioning

confidence: 99%

Chinese Utterance Segmentation in Spoken Language Translation

Zong

Ren

2003

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…S.Narayanan [6] applied acoustic and syntactic features in maximum entropy framework to label prosody boundary and accent automatically. M.Ostendorf [7] used decision trees and a Markov sequence model to predict the prosody boundary of the text. Ma [8] applied decision tree to label prosody boundary using both text and acoustic information.…”

Section: Introductionmentioning

confidence: 99%

“…So automatic prosody labeling attracts more and more attention now. Some research work had been done in this field [6][7][8][9][10]. S.Narayanan [6] applied acoustic and syntactic features in maximum entropy framework to label prosody boundary and accent automatically.…”

Section: Introductionmentioning

confidence: 99%

Automatic Prosody Boundary Labeling of Mandarin Using Both Text and Acoustic Information

Liu

2008

2008 6th International Symposium on Chinese Spoken Language Processing

View full text Add to dashboard Cite

Prosody is an important factor for a high quality text-tospeech (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.

show abstract

“…Price et al [15] proposed a labeling system consisting of seven labels, called prosodic break indices, which express the degree of perceived decoupling or separation between every pair of words in an utterance. Based on this labeling system, Wightman and Ostendorf [21] developed an algorithm for automatically generating the prosodic labels.…”

Section: Using Prosodic Informationmentioning

confidence: 99%

“…The raw energy contours in dB were normalized to a z-score scale within speaker and tone across stress categories to neutralize intrinsic dierences in energy due to tonal category [20]. Following Wightman and Ostendorf [21], the mean energy was analyzed in the target syllable itself without normalization to the mean energy in the following syllable or the mean energy in the utterance.…”

Section: Training and Testing Datamentioning

confidence: 99%

Using stress to disambiguate spoken Thai sentences containing syntactic ambiguity

Potisuk

Harper²,

Gandour³

Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96

View full text Add to dashboard Cite

We h a v e developed a Bayesian classier to determine whether syllables in connected Thai speech are weakly or strongly stressed by using ve acoustic parameters: syllable rhyme duration, mean F0, F 0 standard deviation, mean energy, and the standard deviation of the energy. With speaker-dependent data normalization, we a c hieved a classication accuracy of 99%. The classication accuracy drops to 96% when we used speaker-independent normalization. We h a v e also developed prosodic constraints that can use this stress information to syntactically disambiguate a class of ambiguous sentences that arise from the use of compounding in spoken Thai.

show abstract

Automatic labeling of prosodic patterns

Cited by 226 publications

References 28 publications

Chinese Utterance Segmentation in Spoken Language Translation

Chinese Utterance Segmentation in Spoken Language Translation

Automatic Prosody Boundary Labeling of Mandarin Using Both Text and Acoustic Information

Using stress to disambiguate spoken Thai sentences containing syntactic ambiguity

Contact Info

Product

Resources

About