Acoustic-phonetic characteristics of older children’s spontaneous speech in interactions in conversational and clear speaking styles

Abstract

This study investigated (a) the acoustic-phonetic characteristics of spontaneous speech produced by talkers aged 9–14 years in an interactive (diapix) task with an interlocutor of the same age and gender (NB condition) and (b) the adaptations these talkers made to clarify their speech when speech intelligibility was artificially degraded for their interlocutor (VOC condition). Recordings were made for 96 child talkers (50 F, 46 M); the adult reference values came from the LUCID corpus recorded under the same conditions [Baker and Hazan, J. Acoustic. Soc. Am. 130, 2139–2152 (2011)]. Articulation rate, pause frequency, fundamental frequency, vowel area, and mean intensity (1–3 kHz range) were analyzed to establish whether they had reached adult-like values and whether young talkers showed similar clear speech strategies as adults in difficult communicative situations. In the NB condition, children (including the 13–14 year group) differed from adults in terms of their articulation rate, vowel area, median F0, and intensity. Child talkers made adaptations to their speech in the VOC condition, but adults and children differed in their use of F0 range, vowel hyperarticulation, and pause frequency as clear speech strategies. This suggests that further developments in speech production take place during later adolescence. [Work supported by ESRC.]

The Acoustic Features and Didactic Function of Foreigner Directed Speech: A literature Review.

“…They observed higher word intensity for repeated words in FDS than NDS. Kendi and Khattab (2019), in their study on Arabic FDS, also reported significantly higher vowel intensity in FDS than NDS (in line with Hazan et al, 2015;cf. Knoll et al, 2015).…”

“…It is important to note that although some studies underscored the absence of heightened mean pitch in FDS, other studies have disclosed a different pattern of results. In recent work, Kendi and Khattab (2019) demonstrated that pitch average midpoints were higher in FDS than NDS (Hazan et al, 2015). However, this significant increase in mean pitch in FDS might be due to specific aspects of their study design.…”

“…We were able to identify 12 studies published to date, that have investigated the presence of vowel hyperarticulation in FDS: 8 studies out of a total of 12 reported vowel hyperarticulation in FDS. Most studies focused on English FDS and identified vowel hyperarticulation as the main feature differentiating FDS from NDS 2009a;Uther et al, 2007;Scarborough et al, 2007;Hazan et al, 2015). At the best of our knowledge, only Kendi and Khattab (2019) have reported vowel hyperarticulation in a language other FDS, A REVIEW.…”

“…While vowel hyperarticulation across languages awaits further research, speech rate adjustments in FDS have been explored directly in cross-linguistic studies (e.g., English vs. French). Warren-Leubecker & Bohannon (1982) and Hazan et al (2015) reported that English FDS has a lower word rate per minute than NDS. Kühnert and Antolík (2017) provided evidence from French using a tandem paradigm: participants with two different native languages (L1s) practiced language exchange to help each other learn an L2; each participant was a native speaker of the L2 the other participant wished to learn.…”

Combining conversational speech with read speech to improve prosody in Text-to-Speech synthesis

“…Spontaneous conversational speech exhibits prosodic characteristics that distinguish it from read speech [3], e.g. syllabic reduction, decreased prosodic range [4] and differing stress placement [5]. Further, conversational turns are highly context-dependent, exhibiting changes in emotion and intent, topic/focus structure, and subtle prosodic changes due to phenomena such as (dis)entrainment.…”