My wife, Sasha, shared this audio clip with me this morning and I discovered 2018’s version of The Dress. The clip is a high pitched robotic voice repeating “Yanny” over and over… Or wait, it’s a low pitched man’s voice repeating, “Laurel.” I decided to dig a little deeper.

A few others got there first. The Verge has a writeup going into how different frequencies affect our perception of the audio and how susceptibility to prompting can also affect what you perceive. But I’m usually very susceptible to prompting, and I stood there repeating “Laurel, Laurel, Laurel” to myself over and over, but I couldn’t make it happen. It was so clearly “Yanny.”

I tried different devices, different headphones, and speakers, but I couldn’t make my brain hear anything else. Eventually I stumbled upon some pitch shifted versions. This didn’t do it for me at first, but on the very highest pitched one I started to see how someone could hear “Laurel.” There was something like an undertone there, some garbling in the background that hinted at “Laurel.” For her part, Sasha was finally able to hear “Yanny” by listening to the lowest pitched file.

So I did the only thing I could. I recorded the audio and went into Audition to apply some filters to see if I could figure out what was happening.

The end result was this:

In that audio clip, I explain how I modified the file. First I applied a Fast Fourier Transform (FFT) filter to cut off all audio above 2 kHz. This, finally, made it say “Laurel.” Clear as day. Then something strange happened. When I went back to the original, I found, “Laurel, Laurel, Laurel.” Uh oh. Now I’m in an odd middle ground. I can sometimes hear one. Sometimes the other. It seems to matter what I prime myself for by hearing the pitch shifted or filtered versions first. But I still haven’t entirely wrapped my mind around it.

I went on to try various other filters and shifts. Listen to the audio for the full experience.