Computer model sheds new light on birdsong
Researchers have developed a computer model that can calculate the sound of a birdsong based on measuring the dimensions of the larynx and the properties of the laryngeal vocal folds. The researchers hope that this model can eventually also inform surgeons how a person’s voice will sound after undergoing surgery for laryngeal cancer.
For many years, Associate Professor Coen Elemans from the Department of Biology at the University of Southern Denmark has been trying to develop a computer model that can accurately predict the sound of a birdsong by knowing the physical properties of the bird’s vocal organ.
Several years later, an international collaboration has finally succeeded in achieving this, and the computer model makes exploring how both animals and people produce voiced sound much easier for researchers.
For example, researchers can now study the biology behind the vocalizations of many songbirds that are so small that studying the vibrations of their vocal folds when they produce sound is practically impossible.
The model can hopefully also be used to predict how people will sound after undergoing surgery for laryngeal cancer, often giving them a whole new voice.
“In our experiment, the model was spot on every time. This is also the first time a model of this type has been examined as thoroughly as we have done here,” says Coen Elemans.
The research has been published in the Proceedings of the National Academy of Sciences in the United States of America.
Researchers lacked data to verify the computer model
Over the past 9 years, Coen Elemans has collaborated with several physicists, mathematicians and engineers to develop a computer model that mimics how birds produce voiced sound.
Unfortunately, the 9 years has been frustrating since no one has been successful until now.
Coen Elemans coincidentally heard about a group of researchers from the University of Maine who had developed the exact computer model he had been trying to develop for many years.
However, the researchers lacked data to test and verify their computer model.
“I can do sophisticated experiments, but I lacked a model, so we collaborated and developed experiments in which we could use my data to test their computer model,” explains Coen Elemans.
Using pigeons to verify an advanced computer model
The researchers tested their computer model on data from six pigeons.
Coen Elemans had studied the pigeons thoroughly, measuring the three-dimensional structure of the larynx, the air pressure through the larynx, the movements of the muscles, the stiffness of the vocal folds, the physical properties of all the biological elements and much, much more.
The researchers entered all these data into the model and then investigated whether the model’s output matched the ones measured in terms of fundamental frequencies, the depth of sound and the vibrations of the vocal folds – all of which express voiced sound production.
“The study focused on determining whether the computer model could accurately predict how an animal sounds, and it did this perfectly, which was extremely positive,” says Coen Elemans.
Investigating songbirds more precisely
Coen Elemans says that the computer model opens up multiple perspectives.
Researchers can use it to examine animal vocalizations that are not well suited to physical examination.
For example, songbirds are often so small that researchers cannot examine the vocal folds in their syrinx.
“We can use the model to investigate how various physical and structural parameters influence the unique vocalizations of the songbirds,” explains Coen Elemans.
Predicting a person’s voice after throat surgery
The second perspective relates to people being operated for laryngeal cancer.
This type of models hopefully may enable doctors to examine how surgery will change a person’s voice, depending on the part of the larynx the doctors remove.
The model thus gives doctors more confidence in the outcome of their surgical procedure.
“Previously, surgeons removed the vocal folds completely, but now they can remove tumours much more accurately, so they no longer remove all the vocal folds. This type of model can be used in the future to assess how surgery affects voiced sound production,” says Coen Elemans.
Cannot replicate the vocalizations of dinosaurs
It is tempting to think that this model could provide insight into how extinct animals such as dinosaurs sounded, but according to Coen Elemans, the model is not suitable.
The model shows which parameters are extremely important for predicting an animal’s vocalizations.
One of the most influential parameters is the tissues in the vocal organ.
“These tissues very rarely fossilize, and without that parameter, investigating dinosaur vocalizations would be guesswork and not calculation,” concludes Coen Elemans.
“High-fidelity continuum modeling predicts avian voiced sound production" has been published in Proceedings of the National Academy of Sciences of the United States of America. In 2017, the Novo Nordisk Foundation awarded a grant to co-author Coen Elemans for the project Give Voice to Your Body: Decoding Vocal Motor Control.