Don't Panic Over Laurel And Yanny

I woke up this morning, hopped on Twitter, and was confronted by the Laurel-Yanny debate. I happened to have my headphones on and took a listen. I clearly heard "Laurel." Show's over. Move on.

I was doing something else later, and ran across it again. This time I played the audio through my laptop speakers . . . I clearly heard "Yanny."

What gives?

I did some research and found where the New York Times had tracked it down to a high school student in Georgia (USA). In the course of his studies the other day, he used his computer mic to record the audio of the word "Laurel" being pronounced at vocabulary,com. When he played it back, it sounded like "Yanny." He posted the audio on Reddit with a poll asking what others heard, and it took off from there.

More than just a huge time waster (it's that too) this was in interesting exercise in psyhoacoustics. The mic used, in addition to the ambient background noise at the moment of recording, added certain frequencies to the original content. Those additional frequencies create the "yanny" effect. Depending on what you use to play back the audio, it might either emphasize those frequencies so it sounds like "yanny" or emphasize the lower frequencies that dominated the original recording.

There's also the matter of the age of the listener. As we age our sensitivity to higher frequencies decreases, so younger people would find it easier to hear "yanny" when the listen to the audio.

Out of curiosity, I used my own base mic to make some recordings. First, I recorded my own voice saying "Laurel." A spectrum plot of that looked like this.

My "baseline" recording

Then I used the mic to record the controversial Laurel-Yanny audio at different distances from my speakers. The first one was sort of middle of the road.

Notice the spikes in the 1500-2000 Hz range. My guess is this is what got picked up and causes the "yanny" effect. On the next one I held the mic closer to the speaker.

This one I heard as closer to sounding like "Laurel." The spikes are still there, but the lower frequencies where "Laurel" comes across most strongly are very full.

Finally, there's this one.

It was recorded with the mic at the greatest distance from the speaker, and the lower frequencies are less dominant, allowing the higher transient frequencies to shine through, making it more likely that you'd hear "Yanny."

If you go to the Times you can find where their staff has created a little app that lets you adjust the pitches and frequencies to hear more "Laurel" or more "Yanny." It's a pretty cool little effect.