Descrição comparativa de aspectos fonetico-acusticos selecionados da fala e do canto em portugues brasileiro

AUTOR(ES)
DATA DE PUBLICAÇÃO

2002

RESUMO

In this dissertation three acoustic phonetic aspects were selected in order to compare speech to singing. The selected aspects were temporal structure, formant pattem and coproduction between consonants and vowels; and for each aspect a study has been elaborated, thus consisting the thesis the sum of this three studies. First we discuss the bad diction myth, widespread among musicians, which propagates that singers and especially female singers sing in an unintelligible way. Our purpose in this investigation is to verify similarities and differences between what we call the two speech types, speech stricto sensu and sung speech, following Sundberg s finding (1977) in order to raise hypotheses from the production s point of view. Sundberg s finding, which focused treble notes produced by a soprano in his experiment, is that the higher the fundamental frequency of the musical note, the more the mandible is lowered, and as a result the first formant of the sung vowel matches the first harmonic of phonation frequency. To elaborate the experiment we focused on the Brazilian classical song and one of its characteristics that is to be composed, in general, in middle frequency regions, or even low frequency. A song by Francisco Mignone, Cantiga de Ninar, for soprano was chosen, from which we extracted a phrase to be the carrier-phrase in a low mean frequency region. The nonsense word /la CV/ was introduced in the carrier phrase, without altering its rhythmic and melodic structure. Once the spoken and sung phrases produced by five Brazilian sopranos were recorded and digitized, we were able to perform the acoustic measurements which were the nonsense word segment duration and the first three formants ofthe vowel target in the CV syllable. In this syllable the consonant was [p,t,k] and vowels were [a, ?épsilon?,e, i, ômicron , o, u], oral, in stressed position. To raise our explaining hypotheses conceming the differences between speech and singing we took into account the acoustic results and interpreted the articulatory singing manoeuvres based on the Acoustic Theory of Speech Production of Fant (1960). Our general conc1usion was that these manoeuvres are subject to at least to two musical constraints: tuning and volume. However, as the note is produced on a vowel it means there must be a trade-offbetween speech and singing to avoid the loss of vowel colour, as well as to avoid unintelligibility. Our finding has to do with how this trade-offhappens: consonants are shortened in singing, the first three formants match the harmonics and instead of coarticulation there is coproduction in singing. We propose that this trade-off implying musical and linguistic cognitive abilities, can be examined later, on the basis of the FAAR Acoustic Articulatory Phonology (Albano, 2001)

ASSUNTO(S)

fala fonetica acustica canto

ACESSO AO ARTIGO

Documentos Relacionados