Susanne Schötz

Acoustic Analysis of Adult Speaker Age

Information about the age of the speaker is always present in speech. It is used as perceptual cues to age by human listeners, and can be measured acoustically and used by automatic age estimators. This chapter offers an introduction to the phonetic study of speaker age, with focus on what is known about the acoustic features which vary with age. The age-related acoustic variation in temporal as well as in laryngeally and supralaryngeally conditioned aspects of speech has been well documented. For example, features related to speech rate, sound pressure level (SPL) and fundamental frequency (F 0 ) have been studied extensively, and appear to be important correlates of speaker age. However, the relationships among the correlates appear to be rather complex, and are influenced by several factors. For instance, differences have been reported between correlates of female and male age, between speakers of good and poor physiological condition, between chronological age and perceived age, and also between different speech sample types (e.g. sustained vowels, read or spontaneous speech). More research is thus needed in order to build reliable automatic classifiers of speaker age.

Does the speaker's voice quality influence children's performance on a language comprehension test?

Abstract Purpose. A small number of studies have explored childrens perception of speakers’ voice quality and its possible influence on language comprehension. The aim of this explorative study was to investigate the relationship between the examiners voice quality, the childs performance on a digital version of a language comprehension test, the Test for Reception of Grammar (TROG-2), and two measures of cognitive functioning. Method. The participants were (n = 86) mainstreamed 8-year old children with typical language development. Two groups of children (n = 41/45) were presented with the TROG-2 through recordings of one female speaker: one group was presented with a typical voice and the other with a simulated dysphonic voice. Result. Significant associations were found between executive functioning and language comprehension. The results also showed that children listening to the dysphonic voice achieved significantly lower scores for more difficult sentences (“the man but not the horse jumps”) and used more self-corrections on simpler sentences (“the girl is sitting”). Conclusion. Findings suggest that a dysphonic speakers voice may force the child to allocate capacity to the processing of the voice signal at the expense of comprehension. The findings have implications for clinical and research settings where standardized language tests are used.

Development of speech motor control: Lip movement variability

This study examined variability of lip movements across repetitions of the same utterance as a function of age in Swedish speakers. The specific purpose was to extend earlier findings by examining variability in both phase and amplitude. Subjects were 50 typically developed native Swedish children and adults (28 females, 22 males, aged 5 to 31 yr). Lip movements were recorded during 15 to 20 repetitions of a short Swedish phrase using three-dimensional articulography. After correction for head movements, the kinematic records were expressed in a maxilla-based coordinate system. Movement onset and offset of the utterance were identified using kinematic landmarks. The Euclidean distance between receivers on the upper and lower lips was calculated and subjected to functional data analysis to assess both phase and amplitude variability. Results show a decrease in both indices as a function of age, with a greater reduction of amplitude variability. There was no difference between males and females for either index. The two indices were moderately correlated with each other, suggesting that they capture different aspects of speech production. Utterance duration also decreased with age, but variability was unrelated to duration. The standard deviation of utterance duration also decreased with age. The present results thus suggest that age related changes in speech motor control continue up until 30 years of age.

A Study of Acoustic Correlates of Speaker Age

Speaker age is a speaker characteristic which is always present in speech. Previous studies have found numerous acoustic features which correlate with speaker age. However, few attempts have been made to establish their relative importance. This study automatically extracted 161 acoustic features from six words produced by 527 speakers of both genders, and used normalised means to directly compare the features. Segment duration and sound pressure level (SPL) range were identified as the most important acoustic correlates of speaker age.

A prototype robot speech interface with multimodal feedback

Speech recognition is available on ordinary personal computers and is starting to appear in standard software applications. A known problem with speech interfaces is their integration into current graphical user interfaces. This paper reports on a prototype developed for studying integration of speech into graphical interfaces aimed towards programming of industrial robot arms. The aim of the prototype is to develop a speech system for designing robot trajectories that would fit well with current CAD paradigms.

The Impact of Visual and Auditory Cues in Age Estimation

Several factors determine the ease and accuracy with which we can estimate a speakers age. The question addressed in our study is to what extent visual and auditory cues compete with each other. We investigated this question in a series of five related experiments. In the first four experiments, subjects estimated the age of 14 female speakers, either from still pictures, an audio recording, a video recording without sound, or a video recording with sound. The results from the first four experiments were used in the fifth experiment, to combine the speakers with new voices, so that there was a discrepancy in how old the speaker looked and how old she sounded. The estimated ages of these dubbed videos were not significantly different from those of the original videos, suggesting that voice has little impact on the estimation of age when visual cues are available.