6.646, Sum: Acoustic data in sociolinguistics

[The Linguist List]

----------------------------------------------------------------------
LINGUIST List:  Vol-6-646. Fri 05 May 1995. ISSN: 1068-4875. Lines: 85
 
Subject: 6.646, Sum: Acoustic data in sociolinguistics
 
Moderators: Anthony Rodrigues Aristar: Texas A&M U. <aristar at tam2000.tamu.edu>
            Helen Dry: Eastern Michigan U. <hdry at emunix.emich.edu>
 
Assoc. Editor: Ljuba Veselinova <lveselin at emunix.emich.edu>
Asst. Editors: Ron Reck <rreck at emunix.emich.edu>
               Ann Dizdar <dizdar at tam2000.tamu.edu>
               Annemarie Valdez <avaldez at emunix.emich.edu>
 
-------------------------Directory-------------------------------------
 
1)
Date: Wed, 3 May 1995 19:41:59 +0200 (MET DST)
From: Mario Cal Varela (iamario at usc.es)
Subject: Sum: Acoustic data in sociolinguistics
 
-------------------------Messages--------------------------------------
1)
Date: Wed, 3 May 1995 19:41:59 +0200 (MET DST)
From: Mario Cal Varela (iamario at usc.es)
Subject: Sum: Acoustic data in sociolinguistics
 
Content-Length: 3199
 
Dear linguists:
Last March I posted a query to the list regarding the handling of
acoustic data in sociolinguistic studies. I asked about how
to normalize values from different speakers and how best to plot formants
onto charts. My excuses for such a delay in posting a summary of the
responses I got.
 
Gunnar Fant's (1973) Speech Sounds and Features was mentioned by several
people with regard to both normalization and mel-scales. However, I have
not had access to this specific work yet so I do not know how he deals
with the problem.
As alternatives to the Mel scale, either the Bark scale or the ERB-rate
scale were also mentioned by some respondents. Sylvia Moosmueller listed
a number of references on Bark-scales:
 
Antti Iivonen (1989): Regional German Vowel Studies. Helksingin
Yliopiston Fonetiikan Laitoksen Monisteita; Mimeographed series of the
Department of Phonetics, University of Helsinki, nr. 15.
 
Bertram Scharf (1959): "Critical Bands and the Loudness of Complex Sounds
Near Threshold". JASA 31/3: 365-370.
 
Zwicker, Flotter and Feldtkellner (1967): Das Ohr als
Nachrichtenempfaenger, Hirzel Verlag: Stuttgart.
 
However, according to, for example, Helmer Strik, Ingolf Franke or David
Deterding, simply shifting values on a log scale may prove just as good as
any other more complex transformations.
 
As for normalization of formant values, one of the respondents suggested
using the schwa or neutral vowel as a reference vowel.
Other procedures are commented on in the following references suggested
by Murray Munro:
 
Hindle, D. (1978) "Approaches to vowel normalization in the study of
natural speech". In D. Sankoff (ed.) Linguistic Variation: Models and
Methods (p. 161-171). [This one compares different procedures to
normalize sociolinguistic data]
 
Disner, S. (1986) "On describing vowel quality". In J. Ohala and J.
Jaeger (eds.) Experimental Phonology (pp. 69 and ff.)
 
Miller, J. D. (1989) "Auditory-perceptual interpretation of the vowel".
JASA 85: 2114-2134.
 
In addition, D. Deterding deals in his PhD thesis with the problem of
speaker normalization:
D. Deterding (1990): Speaker Normalization for Automatic Speech
Recognition, Cambridge University.
 
Once I have made the relevant decisions and obtained any results I will
post another message to the list. Meanwhile, thanks very much again to
all those who responded (Murray Munro, Helmer Strik, Ingolf Franke, David
Deterding, John Kingston, and Sylvia Moosmueller)
 
Mario Cal
--------------------------------------------------------------------------
LINGUIST List: Vol-6-646.