New research reveals that children can be identified by just the sound of the child’s voice starting at five years old.
Researchers from the US recorded audio clips of voices of children and gave them the opportunity to guess their gender.
They found the talker’s gender could be identified from age five – almost a decade before differences in the vocal tracts of males and females start developing during puberty.
Experts say that humans tend to judge a speaker’s gender by their voice and resonance. However, they also consider age and height.
Researchers at the University of Texas at Dallas and University of California Davis report that we tend to judge speaker gender mostly based on their voice pitch and resonance.
Pitch refers to the sound’s high frequency or low frequency, and resonance is the sound quality that sounds deep, full, and reverberating.
‘Resonance is related to speaker height – think violin versus cello – and is a reliable indicator of overall body size,’ said study author Santiago Barreda at the University of California, Davis.
“Apart from these fundamental cues, there’s a variety of subtle cues about behaviour and how someone chooses to communicate, instead of relying solely on anatomical characteristics.
Researchers are especially interested in how children perceive gender. Young boys and young girls have very similar voices before reaching puberty.
However, adult male voices and female voices can often be quite different, which makes gender identification very easy.
Researches wanted to determine what type of changes take place in the voices of children as they age.
Talkers can be identified almost a decade before differences in the vocal tracts of males and females start developing during puberty, such as growth of the larynx (or voice box, shaded in this artistic rendering)
For the study, the team developed a database of speech audio samples of children speaking, aged from five to 18.
The clips were the sounds of spoken syllables (‘heed’, ‘hod’ and ‘who’d’) as well as these syllables spoken in full sentences (e.g. “The teacher said you should listen to her advice.”
Audio clips were played to 40 participants – 31 females and nine males, all of whom were undergraduate students at the University of Texas at Dallas.
One-half of listeners were asked to hear isolated sounds, and the other half were instructed to listen for complete sentences.
Participants identified either the male or female talker after listening to the audio.
It was found that those who listened to the entire sentence were more adept at correctly identifying which gender the children were. gender differences in voices come across better in sentences.
Researchers also made two other important findings – first, listeners can reliably identify the gender of individual children as young as five years.
These are well before any anatomical differences exist between speakers, nor before any reliable differences of pitch and resonance.
‘Based on this, we conclude that when the gender of individual children can be readily identified, it is because of differences in their behaviour, in their manner of speaking, rather than because of their anatomy,’ said Barreda.
“In other terms, the gender of speech information can be heavily based upon performance and not on differences in physical appearance between male speakers and female speakers.”
These results suggest that it is possible to identify children’s gender from their speech. This is especially true when the talker uses a longer speech length (sentences, not syllables). The image below shows the classification rates of individual talkers aged five- to eleven years old. Numbers indicate talker’s age (males in circle). Based on isolated syllables and sentences, voices within the shaded quadrants could be correctly identified
Second, the researchers found that identification of gender is possible when speakers are identified together with their age and probable physical size.
Also, this means you should be able detect a child’s gender by their voice and likely also know their age or size.
Barreda stated that ‘essentially, there’s too much uncertainty within the speech signal for age, gender and size to be treated as independent decisions.
“One way to solve this problem is to think about, for instance, what does 11-yearold boy sound like? rather than what does 11 year-old male sound like or what does 11-yearold girl sound like? as if they were separate questions.
The Journal of the Acoustical Society of America published the study.