Pronuncation modeling applied to automatic segmentation of spontaneous speech

Kipp, Andreas; Wesenick, Maria-Barbara; Schiel, Florian

doi:10.21437/eurospeech.1997-358

Cited by 26 publications

(2 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The figures express the number and the percentages of symbols that were substituted, deleted and inserted in the AGT when we compare it to the Tref. These data indicate that the quality of this initial AGT is already reasonable if we consider that data on agreement between human transcribers vary between 93.1% and 94.4% for care ful speech [10] and between 78.8% and 86.2% for less careful speech [6]. As this AGT is a simple concatenation of canoni cal forms in which no cross-word processes are represented, the first thing to do is to find out how many of the processes presented in Table 2 take place at word boundaries.…”

Section: Quantitative Resultsmentioning

confidence: 84%

Phonetic transcriptions in the spoken dutch corpus: how to combine efficiency and good transcription quality

Cucchiarini¹,

Binnenpoorte²,

Goddijn³

2001

7th European Conference on Speech Communication and Technology (Eurospeech 2001)

View full text Add to dashboard Cite

This paper reports on an experiment aimed at establishing how phonetic transcriptions for the large CGN corpus can be obtained most efficiently. This experiment explores the po tential of an automatically generated transcription (AGT) by comparing an AGT with a reference transcription (Tref) of the same material, to determine whether and how the AGT can be improved to make it more similar to Tref. The results indicate that the AGT can be optimized through pronunciation variation modelling so as to make human corrections more efficient or even superfluous, at least for some speech styles.

show abstract

Section: Quantitative Resultsmentioning

confidence: 84%

Phonetic transcriptions in the spoken dutch corpus: how to combine efficiency and good transcription quality

Cucchiarini¹,

Binnenpoorte²,

Goddijn³

2001

7th European Conference on Speech Communication and Technology (Eurospeech 2001)

View full text Add to dashboard Cite

show abstract

“…To interpret these data we have to compare them to those concerning transcriptions by human transcribres. Data on agreement between human transcribers appear to vary between 93.1% and 94.4% (which correspond to deviation percentages of 6.9% and 5.6%, respectively) for careful speech [8], and between 78.8% and 86.2% (which correspond to deviation percentages of 21.2% and 13.8%, respectively) for less careful speech [9]. The data in Table 2 indicate that the quality of this initial Taut is already reasonable, especially if we consider that the agreement data reported in [8 and 9] 2 it can be inferred that Taut could be improved if some of the processes causing discrepancies between Taut and Tref were modeled.…”

Section: Results: Validation Of Tautmentioning

confidence: 96%

Validation and improvement of automatic phonetic transcriptions

Binnenpoorte

Cucchiarini

2002

7th International Conference on Spoken Language Processing (ICSLP 2002)

View full text Add to dashboard Cite

The ultimate aim of our research is to show that good-quality phonetic transcriptions of large speech corpora can be obtained by employing automatic techniques initially developed for ASR. The experiment presented in this paper has two aims. The first is to show how the quality of an automatic transcription that is easily obtained through lexicon lookup can be measured in a way that is methodologically sound. The second is to show how, while measuring the quality of an automatic transcription, it is possible to obtain information that can subsequently be used to improve the automatic transcription where necessary. As a result, correction by human transcribers should become more efficient or even superfluous.

show abstract