Yanny or laurel?
The internet exploded in argument this week as people debated which of those words a robot was saying in this recording, which originated on Reddit and was posted on Twitter by YouTube vlogger Chloe Feldman.
Hardik Kothare, who works in the speech neuroscience lab at UCSF, was quick to weigh in with his assessment of the sound. To his ear, this was definitely a recording of the word “laurel”.
What do you hear?! Yanny or Laurel pic.twitter.com/jvHhCbMc8I
— Cloe Feldman (@CloeCouture) May 15, 2018
If you’re still on team “yanny,” take a listen to the original recording from Vocabulary.com, where it’s the pronunciation key for just one word: laurel. (Apologies, yanny fans.)
The sound was recorded by a professional opera singer who was one of the original cast members in the Broadway musical Cats, according to Wired.
The dictionary site hasn’t revealed the man’s identity, but said he was one of several trained singers enlisted to record hundreds of thousands of pronunciations, based on the rules of the international phonetic alphabet.
Kothare suggested that the recording was likely “cleverly synthesized” to trick our brain’s powers of speech detection. He said there’s a simple, logical reason why some folks who listen to the viral recording hear “yanny” while others pick up “laurel.”
The importance of frequency
“The human brain is trained to perceive and interpret speech on the fly in a remarkable way,” Kothare said Tuesday on Twitter.
Our ears learn at a young age to pick up clues about the vowels people spit out by focusing on frequencies at which certain sounds tend to resonate. The frequencies for each sound are a little different from person to person and language to language.
“Speech perception and production depends heavily on an internal map of speech sounds,” Kothare said. “You learn this map while learning to speak as a toddler, and also while hearing others speak on a day-to-day basis.”
If you mess with the frequencies in a recording, you can change what people hear – it’s similar to the way that our eyes can be tricked by an optical illusion.
The initial vowel is perceived as /i/ with F1 at the lower band and the higher band as F2. Completely ignoring the spectral power in the ‘middle’ frequencies. The dip makes these people perceive the middle vowel as /æ/ and then the rise makes them go back to /i/ (13/n) pic.twitter.com/P63lPf5eQ9
— Hardik Kothare (@hardikkothare) May 16, 2018
The New York Times tried this out today and created its own yanny-versus-laurel audio switching tool.
It turns out that our brains can shift pretty easily between hearing yanny and laurel, just based on how low or high the frequency of the recording gets.
Couple this with all the cultural and linguistic ways we’ve been trained to hear certain vowels, and you’ve got a perfect recipe for a little audio illusion.
‘You have categories of sounds in your head’
John Houde, who runs the speech neuroscience lab at UCSF where Kothare works, said that the either-or prompt of yanny or laurel is a classic example of what’s known as a forced-choice experiment. Your brain is bound to pick out one word or the other given that prompt.
“Expectations can really bias your perception of speech sounds,” Houde told Business Insider.
“The sound seems to have a mix of some cues for the sounds making up laurel, and also some cues for the sounds making up yanny.”
Houde added that making quick-fire, perceptual decisions about the words and sounds we hear all day is how we get through life.
“You have categories of sounds in your head,” he said. “Kind of like how words are made up of letters, spoken words are made up of phoneme categories. Your brain is trained to listen to the signal and say, ‘What phoneme sequence did I hear?'”
Whichever word you hear in the recording, the speech scientists both say you shouldn’t worry: “It’s totally OK to hear what you are hearing,” Kothare said.
“Your way of perceiving speech is almost by definition right, because it’s served you well for all these years, understanding other people’s speech,” he said.
“There is no right and wrong in speech perception. There’s just how you perceive it and how I perceive it.”
This article was originally published by Business Insider.
More from Business Insider: