Laurel or Yanni??

This is apparently an optical illusion for your ears and it's all over the internet and everyone is talking about it.

Click on this and say what word you hear.

I hear "Laurel" repeated over and over, and can't imagine how anyone could hear it as 'Yanni'. Do any of you hear the voice saying "yanni"?

CBS News talked to somebody who says the difference in how this is heard is due to how well a person hears high frequencies. Those who have better high frequency hearing are supposedly more apt to hear the voice saying 'yanni'.

(It seems to me that a lot of it might be due to the frequency response of the speakers you are listening to it through.)

The White House is (unsurprisingly) divided:
I heard yarry
Team Yanni!
No normal person could possibly hear "yanni".

The frequency explanation might be right. The New York Times has a thing where you can slide a slider to increase or decrease the higher frequencies and if you increase them, it does start to sound like 'yanni'.

Psychology Today does something similar with wav files that download and you have to click on to play (the high-frequency-enhanced 'yanni' one here sounds to me like Elte's 'yarry':
For some reason, I could only hear Yanny, until I watched this video that splits the two sounds:
Now I can only hear Laurel, even on both in that same video.
I can only hear Yanny now if I adjust the pitch in Audacity.
I think it's more "Viral". (I don't mean it sounds like that)

It reminds me of something I created some time back in regards to dealing with mob mentality on another forum. I used the GD+ library on the server to switch out images and a caption randomly everytime the image placename was called. It meant an image might have stated "A pack of Hyenas" (with a corresponding image of Hyenas) or a pack of dogs or dingos etc. (I think the captions reflected their true collective nouns.)

What I am suggesting is that if you were to layer two different words over the top of each other and then adjusted their gains from one being predominant to the other, you could then use a selection of the same file but "tweaked" to emphasise which voice sample you hear clearly. This means that you could randomise the file to be served up and the discrepancy wouldn't necessarily be noticed unless you actually paid attention to the overall filesize (difficult when dealing with streaming media) and had the true files name. (After all it might well be a placename/pointer that's dynamically allocated)

I was thinking of doing a proof of concept with a "Bob vs Dave" controversy, but there really is only so far that you need to go to get the idea.

Users browsing this thread: 1 Guest(s)