Prosodic boundary detection using syntactic and acoustic information

Kocharov, Daniil; Kachkovskaia, Tatiana; Skrelin, Pavel A.

doi:10.1016/j.csl.2018.07.001

Cited by 7 publications

(6 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…(KLIMKOV et. al., 2017); (1) Classificador Random Forest, usado para detectar os limites prosódicos reais usando um pequeno conjunto de recursos acústicos (KOCHAROV et. al., 2019); (1) SSML -Speech Synthesis Markup Language, que atribui uma variedade de tags de prosódia SSML com base na estrutura de temática de cada frase (DOMINGUEZ et.…”

Section: Discussionunclassified

Características prosódicas associadas aos sinais de pontuação

Galdino

Silva

Oliveira

2021

CadLin

View full text Add to dashboard Cite

O objetivo deste artigo é apresentar uma revisão de escopo sobre as características prosódicas associadas aos sinais de pontuação. Foi realizado um levantamento bibliográfico a partir da pesquisa de descritores em inglês e português, organizados de acordo com a seguinte sintaxe: prosódia AND acústica AND discurso AND estrutura AND ("sinais de pontuação" OR "pontuação gráfica" OR "sinal de pontuação"), sem incluir citações e patentes nas bases de dados: OvidMedlin, Public Medicine Library (PubMed), Scopus (Elsevier), Ebscohost (Academic Search Premier), Gale Academic Online e Google Scholar. Observamos que existe uma diversidade de métodos empregados para analisar a correlação entre os sinais de pontuação e as características prosódicas. Os estudos desta revisão confirmaram nossa pergunta de pesquisa, evidenciando a relação entre os sinais de pontuação e os aspectos prosódicos. A maioria dos trabalhos relacionados à tecnologia desenvolveu diferentes redes neurais para transformar texto em fala e/ou para converter fala em texto e mostrou que as pausas são apontadas como indicadores mais fortes dos sinais de pontuação.

show abstract

Section: Discussionunclassified

Características prosódicas associadas aos sinais de pontuação

Galdino

Silva

Oliveira

2021

CadLin

View full text Add to dashboard Cite

show abstract

“…Many years of experience in collecting, processing and analysing speech material have enabled us to create speech corpora of all kinds that can serve as the basis for a wide range of fundamental and applied research. The fully annotated large corpus of read speech CORPRES laid the foundation for a lot of research projects including automatic prosodic boundary detection (Kocharov et al, 2019a), research on vowel reduction (Kocharov et al, 2019b) and phrase-final lengthening (Kachkovskaia et al, 2013), melodic declination (Kocharov et al, 2015), the melody of post-nucleus and others.…”

Section: Speech Corpora In Phonetic Researchmentioning

confidence: 99%

Principles of the St. Petersburg Phonological School in Speech Corpora Design

Skrelin

Kachkovskaia

Kocharov

et al. 2023

Bakhtiniana, Rev. Estud. Discurso

Self Cite

View full text Add to dashboard Cite

The paper discusses the main principles in designing and annotating speech corpora within the framework of the Saint Petersburg phonological school, and provides examples of using corpus data in phonetic research. One of the major principles that we follow is to analyse the speech material at all levels: from segmental to intonational, including speech disfluencies. During segmental phonetic annotation, we suggest listening to each speech sound in isolation (without knowing its context) and relying on spectrographic data. At the syllabic tier, it is crucial to reflect resyllabification. During prosodic annotation, we suggest to rely on listener’s perception of the intonation pattern first, then analyse the actual melodic curves. A speech corpus with multi-level annotation that follows these principles is a valuable source of phonetic data — as segmental and prosodic factors are in constant interaction with each other, and one cannot analyse units of one annotation tier without reference to other tiers.

show abstract

“…Systems for audio-based prosodic boundary detection have traditionally utilized combinations of different features such as the duration of pauses and syllables, F 0 range and resets, intensity, or pitch movement [4,9,11,14,15]. Rather than use such handcrafted combinations of features, however, we chose to employ learned representations from raw audio data.…”

Section: Model For Audio-based Detectionmentioning

confidence: 99%

“…There are two different scenarios for the automatic detection of prosodic boundaries: detection solely from text, most often for the purposes of speech synthesis [7,13,18,20,24], or detection from spoken utterances as a form of audio annotation. In the latter case, some approaches have been based solely on acoustic information (though sometimes with word or syllable boundaries derived from text transcripts) [11,[14][15][16], while others have combined both lexical and acoustic information [4,8,9].…”

Section: Introductionmentioning

confidence: 99%

Detection of Prosodic Boundaries in Speech Using Wav2Vec 2.0

Kunešová

Řezáčková

2022

Text, Speech, and Dialogue

View full text Add to dashboard Cite

Prosodic boundaries in speech are of great relevance to both speech synthesis and audio annotation. In this paper, we apply the wav2vec 2.0 framework to the task of detecting these boundaries in speech signal, using only acoustic information. We test the approach on a set of recordings of Czech broadcast news, labeled by phonetic experts, and compare it to an existing text-based predictor, which uses the transcripts of the same data. Despite using a relatively small amount of labeled data, the wav2vec2 model achieves an accuracy of 94% and F1 measure of 83% on within-sentence prosodic boundaries (or 95% and 89% on all prosodic boundaries), outperforming the text-based approach. However, by combining the outputs of the two different models we can improve the results even further.

show abstract

Prosodic boundary detection using syntactic and acoustic information

Cited by 7 publications

References 15 publications

Características prosódicas associadas aos sinais de pontuação

Características prosódicas associadas aos sinais de pontuação

Principles of the St. Petersburg Phonological School in Speech Corpora Design

Detection of Prosodic Boundaries in Speech Using Wav2Vec 2.0

Contact Info

Product

Resources

About