What's the deal with this "Laurel vs. Yanny" nonsense? We asked an expert to explain

It was the four-second clip heard round the world.

Much like the infamous dress did in 2015 - which was quite clearly white and gold, BTW - the Yanny vs. Laurel audio clip tore the internet apart. Some people very clearly heard "Yanny," while others know for certain that the robotic voice was saying "Laurel."

So what's the explanation for the divide? To find out, we checked in with Dr. Mario Svirsky, Professor of Hearing Science in the Department of Otolaryngology and Professor of Neuroscience and Physiology at NYU Langone Health.

It turns out, both interpretations of the audio clip are valid: it just depends which components of the sound your brain latches onto.

To explain the Yanny/Laurel conundrum, Svirsky points to the visual illusion known as Rubin's vase: "Is it a vase, or is it two faces looking at each other?" he asks over email.

The answer: "actually, it's neither of those things. It's a bunch of dots on a screen," he says.

Rubin's vase is a two-dimensional image and not a real object, so your brain tries to generate "explanations" for what you're seeing. "The image is ambiguous, so both explanations - vase and faces - are possible," Svirksy explains. "The same thing can happen with speech sounds."

Enter: Yanny and Laurel.

Svirksy captured the first speech sound in the divisive clip - the part where listeners hear either a "Y" or an "L" sound. The image below is a "frequency spectrum," showing the energy at different frequencies for that small piece of audio:

Peep those two big peaks at the 250 Hz and 2,800 Hz marks. According to Svirsky, if your brain disregards the stuff in between and only focuses on those two peaks, you might hear a "Y" for Yanny.

"On the other hand, if your brain prefers to assign more weight to other details of the spectrum, like the peak at 800 Hz or the two valleys - known as 'antiresonances' - between 1500 Hz and 2500 Hz, then it makes more sense to interpret this sound as an 'L,'" he explains.

The conclusion?

Just like Rubin's vase, the four-second Yanny/Laurel audio clip is ambiguous. In an effort to assign an "explanation" to what you're hearing, your brain latches on to certain peaks or valleys of its own choosing - and that determines whether you hear Yanny or Laurel.

"Our brains are amazing computers that can make sense of ambiguity using algorithms that are much more powerful than the best automatic speech recognition systems," Svirsky says. "The problem is that with a small chunk of speech, those algorithms may lead you astray, as they are probably optimized for real world communication."

Our brains are amazing computers.

In his opinion, "If we were provided with a whole sentence instead of a single word, the ambiguity might disappear," because we'd be able to make more accurate estimations of what we're hearing.

Unfortunately, all we've got is four seconds. So let's just agree it's Laurel, okay?