The distribution of speaker information in Dutch fricatives /s/ and /x/ from telephone dialogues

Abstract

Although previous work has shown that some speech sounds contain more speaker-dependent information than others, not much is known about the speaker information of the same segment in different linguistic contexts. The present study therefore investigated whether Dutch fricatives /s/ and /x/ from telephone dialogues contain differential speaker information as a function of syllabic position and labial co-articulation. These linguistic effects, established in earlier work on read broadband speech, were firstly investigated. Using a corpus of Dutch telephone speech, results showed that the telephone bandwidth captures the expected effects of perseverative and anticipatory labialization for back fricative /x/, for which spectral peaks fall within the telephone band, but not for front fricative /s/, for which the spectral peak falls outside the telephone band.Multinomial logistic regression shows that /s/ contains slightly more speaker information than /x/ in telephone speech and that speaker information is distributed across the speech signal in a systematic way; even though differences in classification accuracy were small, for both /s/ and /x/, codas and tokens with labial neighbours are more speaker-specific than onsets and tokens with non-labial neighbours. These findings indicate that speaker information contained by the same speech sound is not the same across contexts.

The effect of word class on speaker-dependent information in the Standard Dutch vowel /aː/

“…A comparison of the three Dutch corner vowels showed that /aː/, which is also used in the present study, contained most speaker-specific information in both the durational and spectral domains, relative to /i/ and /u/ ( Van den Heuvel, 1996). An explanation for these differences is mainly given by articulatory differences between speech sounds, also in relation to their neighboring sounds (Smorenburg and Heeren, 2020), together with the anatomical/physiological differences between individual speakers.…”

“…This yields differences in, for example, closure (or linguo-palatal) contact duration during articulation, and such articulatory differences may in turn alter speech sound acoustics. Recently, Smorenburg and Heeren (2020) showed that speaker classification of fricatives /s/ and /x/ was better with tokens sampled from coda rather than onset positions. Moreover, that study demonstrated that the amounts of between-and within-speaker variation depended on syllabic position (see also He and Dellwo, PREPRINT VERSION of article published in JASA 14-10-2020, https://doi.org/10.1121/10.0002173 8 2017).…”

“…Speech perception research has provided much evidence that speaker information interacts with the interpretation and memory of a spoken message (e.g., Palmeri et al, 1993;Van Berkum et al, 2008). For voice perception, both within-and between-speaker acoustic variation are important (Lavan et al, 2018), whereas the speech production literature shows that speech acoustics, and its variation, depend on linguistic context (e.g., Smorenburg & Heeren, 2020). Taken together, this suggests that also speaker-dependent voice characteristics may be conditioned by linguistic context.…”

“…There is evidence, however, that speaker-dependent information in an utterance is affected by speech style (e.g., Moos, 2010;Dellwo et al, 2015) or speech sound (Van den Heuvel, 1996;Andics, 2013;Kavanagh, 2014). Moreover, Smorenburg and Heeren (2020) recently found that the speaker information contained by the Dutch fricatives /s/ and /x/ to some extent depends on whether the fricative was produced in onset versus coda position. This finding was explained as articulatorily less-demanding positions, such as codas, allowing for more between-speaker variation (see He and Dellwo, 2017).…”

“…In the search for optimal features for acoustic-phonetic FSC, earlier work has compared speaker information carried by different segments (e.g., Van den Heuvel, 1996;Andics, 2013) and different speech styles (Moos, 2010;Dellwo et al, 2015). What is largely lacking from the existing literature, with the exception of Smorenburg and Heeren (2020), is a systematic investigation of how the speaker information carried by a segment may be affected by its position in the utterance within the same speech style. A speech sound's acoustics are altered by linguistic structure, such as whether it is realized in a lexically-stressed or a focused position.…”