The present work aims to determine the most appropriated automatic analysis of phrasal intonation patterns in Brazilian Portuguese, understood as the variation of F0. In order to evaluate the automatic analysis done by the software ExProsodia a perceptual analysis was carried out and its results were compared to the ones given by the software. The corpus consisted of one sentence produced three times by 11 female subjects. In the perceptual analysis an intuitive evaluation was done considering the syllabic nuclei as intonation units (control group). The data from the automatic analysis were analyzed in three different ways: i) the cumulative mean of all points of intonation curve; ii) the cumulative mean of points higher than 0Hz; iii) and the cumulative mean of the points selected by ExProsodia. The acoustic parameters considered for the automatic analysis were: F0- from 50Hz to 700Hz; duration- from 20ms to 600ms; intensity- higher than 10% of RMS mean of intonation curve. A Dunnett's test compared the control group with other groups from the automatic analysis (α = 95%, n = 33). In the comparison between the control group and i & ii, it was obtained p < 0.05. It was verified that the units established by the interval given by ExProsodia (iii) were the only ones that showed no significant differences when compared to the control group. The results indicate a similarity between the automatic and perceptual analyses. The automatic analysis done by ExProsodia seems thus trustworthy.

Este ensaio analisa a expressão da emoção na entoação do português brasileiro (PB) por diferentes gêneros (masculino e feminino). Os trechos do discurso emocional foram coletados na Internet e divididos em colérico, neutro, triste e SNI (simulacros de entoação neutra) para ambos os gêneros. Os parâmetros entoacionais foram analisados automaticamente pelo software ExProsodia, que se baseou em medidas de F0 (Hz) e duração (ms). Os resultados mostraram que o discurso emocional no PB pode ser caracterizado pelos graus de tensão na produção de fala. Desse ponto de vista, cada gênero e emoção relacionada pode ser localizado ao longo do eixo de tensão entoacional proposto: (i) triste>neutro>simulacro de entoação neutra>colérica, para locutores masculinos; (ii) simulacro de entoação neutra/neutra/triste>colérica, para locutores femininos. Foi possível verificar que diferença de gênero se relaciona às características fisiológicas expressas por F0 analisado automaticamente e que a tensão entoacional no discurso emocional em ambos os gêneros também pode ser uma abordagem confiável para a análise automática da expressão de emoção no PB.

This study compares fundamental frequency (fo) and fundamental frequency standard deviation (foSD) of COVID-19 patients with the same parameters in the speech of subjects without COVID-19, and verifies whether there is an effect of age and sex in the patient group. Both groups, subjects with and without COVID-19, are formed by Brazilian Portuguese speakers. Speech samples were obtained from 100 patients with mild to severe symptoms of COVID-19, and 100 healthy subjects. A single 31-syllable Portuguese sentence was used as the elicitation material for all subjects. The recordings were divided into four age groups. The acoustic measures were semi-automatically extracted and analyzed by a series of analyses of variance. Patients with COVID-19 present vocal differences in fo-related parameters when compared to healthy subjects, that is, patient voices presented higher fo and foSD with respect to control voices. In addition, for patient voices, there was an age and sex effect on fo SD values. Vocal parameters of women and elderly subjects showed more marked differences in fo-related parameters, indicating that patient voices are higher-pitched and have a higher variation of fo SD. Consequently, fo-related parameters may be tested as vocal biomarkers in the screening of respiratory insufficiency by voice analysis, in patients with severe symptoms of COVID-19.

This work aims to verify the appropriation of the semitones scale for data analysis in studies involving intonational aspects of Brazilian Portuguese (henceforth BP), regarding mainly speakers perception and intuition at the time of production. The semitone scale was selected because it operates with frequencies ranges and with a central frequency to which the neighboring frequencies, in a given range, would be attracted covering the perception issue. The test was performed with 13 BP native speakers (seven women and six men) and consisted primarily in an exposure for a target-utterance. In next step, they were required to reproduce what they had heard: imitating just the intonation despite of the voice qualities and segments realization. The data (f0 values) obtained were plotted in mels, barks, semitones, and ERB-rates and all compared by (i) correlation (r2), (ii) variation from target-utterance values of F0, and (iii) graphics comparison. The results so far obtained and analyzed indicates that the semitones scales can be used with no loses for frequencies under 600 Hz if compared with Hz scale and better fits for heights comparison. In relation to other scales, any difference was noted in absolute values at this moment. [Work funded by CNPq, Process no. 152293/2010-8.]

O objetivo deste artigo é apresentar considerações sobre um modelamento lógico-matemático desenvolvido para abordar o fenômeno da percepção da entoação para o Português Brasileiro. O modelamento foi feito com base no modelo de análise automática da entoação, desenvolvido por Ferreira Netto (2006, 2008, 2010), e utiliza os princípios desenvolvidos por Hart, Collier e Cohen (1990), no que tange ao fenômeno da conjugação entre a percepção e produção das curvas entoacionais. Além disso, aplicamos, em nosso modelo, os limiares de diferenciação tonal, estipulados por Consoni (2011), valores os quais nos asseguram a relativização em um estado neutro e estados relevantes para a percepção, por meio da entoação para palavras isoladas, bem como para frases de contexto. Nosso modelamento resume-se à criação de um sistema conjugado em que os valores de F0 são processados de acordo com uma componente denominada tom médio (FERREIRA NETTO, 2008). Os limiares de diferenciação tonal operam como limites sistêmicos de lateralidade gerados a partir do tom médio e têm o papel de restringir as variações e explicar possíveis variações modalizadas. O modelamento demanda mais estudos de base para uma melhor funcionalidade, além de uma implementação computacional funcional a fim de se verificar a sua aplicabilidade.