High speed ultrasound/acoustic database of lingual articulation in typically developing children between three and thirteen years old

Zharkova, Natalia (2019). High speed ultrasound/acoustic database of lingual articulation in typically developing children between three and thirteen years old. [Data Collection]. Colchester, Essex: UK Data Archive. 10.5255/UKDA-SN-852289

Typically developing children generally learn to produce most vowels and consonants correctly by the age of three years old. However details concerned with complex variations of speech sound production in words and sentences (or coarticulation) continue to develop during childhood. An example of coarticulation is the difference between two realisations of the consonant /s/, in the words "sea" and "saw", which arises from the influence of the following vowel. Understanding of coarticulatory processes and the constraints on them (for example the ability to differentiate between parts of the tongue) is fundamental to our knowledge of the general processes of speech motor control. The project focusses on subtle developmental changes during childhood and preadolescent years, by collecting and analysing ultrasound imaging data on tongue position and shape in speech. The general aim is to provide articulatory and acoustic data on the maturation of lingual coarticulatory patterns and tongue differentiation from the age of 3 years old to early adolescence. The project contributes to theories of speech production, focussing on the role of tongue constraints in coarticulation development. A long-term practical aim is to gain information that will potentially be useful for diagnosing speech motor problems.

Data description (abstract)

The database consists of synchronised recordings of ultrasound tongue imaging and acoustic data for sixty typically developing children aged between three and thirteen years old, all speakers of Scottish Standard English. There are six tightly defined age groups: 3-year-olds, 5-year-olds, 7-year-olds, 9-year-olds, 11-year-olds, and 13-year-olds. The data, consisting of consonant-vowel syllables in a carrier phrase, were collected at Queen Margaret University, Edinburgh. The ultrasound frame rate was 100 Hz. Each participant from the 7-, 9-, 11-, and 13-year-old groups wore a custom-designed headset, which stabilised the ultrasound transducer in relation to the head. Participants from the 3- and 5-year-old groups did not wear the headset, and the transducer was hand-held by the experimenter. Additionally, each 13-year-old participant produced the same stimuli recorded without the headset. All recordings were made in a sound-treated studio, while all equipment that emitted noise was located in an adjacent room.

Downloads

Altmetric