Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
Music Social Networks The Internet Science

'Yanny vs. Laurel' Reveals Flaws In How We Listen To Audio (theproaudiofiles.com) 233

Unless you've been living under a rock for the past few days, you've probably heard about the controversy over "Yanny" and "Laurel." The internet has been abuzz over an audio clip in which the name being said depends on the listener. Some hear "Laurel" while others hear "Yanny." Ian Vargo, an audio enthusiast who spends most of his working hours of the day listening to and editing audio, helps explain why we hear the name that we do: Human speech is actually composed of many frequencies, in part because we have a resonant chest cavity which creates lower frequencies, and the throat and mouth which creates higher frequencies. The word "laurel" contains a combination of both which are therefore present in the original recording at vocabulary.com, but the clip that you most likely heard has accentuated higher frequencies due to imperfections in the audio that were created by data compression. To make it worse, the playback device that many people first heard the audio clip playing out of was probably a speaker system built into a cellular phone, which is too small to accurately recreate low frequencies.

This helpful interactive tool from The New York Times allows you to use a slider to more clearly hear one or the other. Pitch shifting the audio clip up seems to accentuate "laurel" whereas shifting it down accentuates "yanny." In summary, this perfect storm of the human voice creating both low and high frequencies, the audio clip having been subject to data compression used to create smaller, more convenient files, and our tendency to listen out of devices with subpar playback components lead to an apparent near-even split of the population hearing "laurel" or "yanny."

'Yanny vs. Laurel' Reveals Flaws In How We Listen To Audio

Comments Filter:
  • by 93 Escort Wagon ( 326346 ) on Friday May 18, 2018 @03:10AM (#56631296)

    And I heard “Turn me on, dead man” - but my wife swore she heard “Number nine”.

  • Another example (Score:5, Interesting)

    by Knightman ( 142928 ) on Friday May 18, 2018 @03:11AM (#56631298)

    Brain storm or Green needle:
    https://youtu.be/5pRY3wlKwm8 [youtu.be]

    Anticipate the word you want to hear and you will hear it.

    • Re:Another example (Score:5, Informative)

      by Mister Transistor ( 259842 ) on Friday May 18, 2018 @03:26AM (#56631334) Journal

      I noticed that "hysteresis" effect, too. When you start at one end or the other, your brain locks onto one or the other sound and you keep recognizing that word past the "critical point" on the slider you lost it at in the other direction. Once you become accustomed to hearing one or the other you get biased to keep hearing it despite it trending in the other direction.

      That happened to me accidentally the other day, it was on some TV show I wasn't really paying attention to, and for the first time I heard "Laurel" distinctly, then my brain shifted to hearing "Yanny" and I rewound the DVR and all I could hear on replays was "Yanny".

      This is sort of similar to the optical illusion of the staircase that can be going up or down until you "flip" it by seeing it going the other way, or the 2 faces/vase silhouettes illusion, or the inside-out face, they all make your brain "click" or "flip" from one interpretation of the image to the other.

      • by Mandrel ( 765308 )

        This gestalt flipping is due to the associative networks in the speech recognition section of your temporal cerebral cortex settling into two different low-energy states created by your fluency in English, one a word, and the other a word-like name.

        And the adaptation to each word is probably due to the short-tern strengthening of synapses between active neurons, widening and deepening the energy well for the recently-heard word.

        By moving the slider slowly enough, I can move the switch-over from close t

      • by ls671 ( 1122017 )

        Exactly, but just like for those optical illusions, I am able to condition myself to see or hear either known alternatives. Anybody else?

      • I noticed that "hysteresis" effect, too. When you start at one end or the other, your brain locks onto one or the other sound and you keep recognizing that word past the "critical point" on the slider you lost it at in the other direction. Once you become accustomed to hearing one or the other you get biased to keep hearing it despite it trending in the other direction.

        I was wondering about that was well. I first hear Yanny but when I went to slide it to Laurel I never clearly heard Yanny again, the close I came was "Garry" at the far right. Repeated tries yield the same result except I do not hear Yanny when I first open the NYT slider.

      • I noticed that "hysteresis" effect, too. When you start at one end or the other, your brain locks onto one or the other sound and you keep recognizing that word past the "critical point" on the slider you lost it at in the other direction. Once you become accustomed to hearing one or the other you get biased to keep hearing it despite it trending in the other direction.

        That happened to me accidentally the other day, it was on some TV show I wasn't really paying attention to, and for the first time I heard "Laurel" distinctly, then my brain shifted to hearing "Yanny" and I rewound the DVR and all I could hear on replays was "Yanny".

        This is sort of similar to the optical illusion of the staircase that can be going up or down until you "flip" it by seeing it going the other way, or the 2 faces/vase silhouettes illusion, or the inside-out face, they all make your brain "click" or "flip" from one interpretation of the image to the other.

        I noticed this as a simple result of having the volume low on my computer speaker system. At low volume the bias is to hear the higher pitches and I heard Yanny. At higher volume, the bass is given a bigger sound stage and it was then that I was able to hear Laurel. My thought is that what you hear has more to do with the speaker quality, volume, and sound range than anything else.

      • by epine ( 68316 )

        At the 75% "yammy" position, I can mentally prime myself to hear either one, so long as my mental prime is right at the beginning.

        Once I got the hang of it, I alternated hearing Laurel/Yammy without fail several dozen times in a row. Using hysteresis and moving in small increments, I even managed to hear "Laurel" a few times at the 95% Yammy position (but it was weird and growly). Laurel is way dominant for me.

        At the 80% position (after sliding up from Laurel-ville) ,during transition from Laurel to Yammy,

      • Reminds me of the odd cases where neural networks break down and reveal surprising problems with their ability to accurately discern what they are shown. You know, where the machine is shown a bunch of pixels and responds with 85% accuracy "This is a cat."

        https://www.wired.com/2015/01/... [wired.com]

    • Nope, that's just stupid. It's blatantly "brain storm" (or "grain stall", it is really bad quality), there is nothing even remotely resempling "needle" in there.

      • by iCEBaLM ( 34905 )

        Interesting, I hear three syllables, so "green nee-dle" is how I interpret it.

      • Nope, that's just stupid. It's blatantly "brain storm" (or "grain stall", it is really bad quality), there is nothing even remotely resempling "needle" in there.

        ... I heard green needle. Played it back 10 times and I can't find brain storm in there anywhere.

    • When I first listened to it I couldn't make sense of the sounds. It actually felt like multiple sounds at once. Then I heard Yanni and later Laurel.

      • Me too, it kept changing! Probably didn't help that I already knew what to listen out for as this story is EVERYWHERE
    • It is not really any of them. Both of the samples are so filled with noise that they leave your brain to fill in a LOT.
    • by MTEK ( 2826397 )

      Brain needle

    • by muffen ( 321442 )

      Brain storm or Green needle

      If you try, "brain needle" and "green storm" also works!

    • Brain storm or Green needle:
      https://youtu.be/5pRY3wlKwm8 [youtu.be]

      Anticipate the word you want to hear and you will hear it.

      I was pretty sure I would hear "Laurel", and expected to hear "Laurel" because I heard the explanation about why people hear different things. I ended up hearing something I didn't expect to hear. Neither "Laurel" or "Yanni" but "Yenny". I do not hear the first vowel as an "a" at all but rather an "e".

    • At first I hear both at the same time. Then my brain automatically fills in "Ell"-"A"-" You"-"Are"-"E"-"Ell"---"laurel".
      I think CMU has a JavaScript implementation of their text to speech engine that resembles the Speak'N Spell.
  • It's official: (Score:5, Insightful)

    by Anonymous Coward on Friday May 18, 2018 @03:29AM (#56631344)

    According to TFS, I've been living under a rock. Or, perhaps, not everyone spends the entire day browsing Facebook and Buzzfeed.

    • by rhazz ( 2853871 )
      Honestly it's fucking everywhere. I first heard about it on the radio during my commute, then it was in the news the next morning. The dress thing was almost interesting but this just feels like society has too much free time on their hands.
    • One of the few occasions I'm glad my rock sometimes has sporadic access to the intarwebz.
  • Sorry for it being a crapbook video link :( https://www.facebook.com/phros... [facebook.com]
  • Or maybe I just do not care about these things. Like at all. No, not even a bit.

    • by Anonymous Coward on Friday May 18, 2018 @03:37AM (#56631380)

      Yet you cared enough to comment and criticize.

    • It helps to illustrate what a difficult problem voice recognition is, and the problems we will undoubtably face as these algorithms become more ubiquitous.

      Consider the case of medical transcription. Did the the doctor say you have "cancer" or "herpes"? It's important we get this right.

    • Stupid social media controversy is stupid. "Do you hear yanny or laurel ?" my answer : "it is blue and gold" ;).
    • Or maybe I just do not care about these things. Like at all. No, not even a bit.

      I hope you're not an engineer. The way the result of some design affects the minds of those people who will use it should matter to you a lot. It can mean the difference between a paycheck and a lawsuit. Anyone who's job it is to create something for someone should at least take a marginal interest in how the brain can turn their work against them.

      • by gweihir ( 88907 )

        I am a very senior engineer and also a scientist. And my paycheck is rock solid. Incidentally, this is not about "the brain", this is about how the auditory subsystem does error correction. That is a bit different and not in any way related to my work.

    • Or maybe I just do not care about these things. Like at all. No, not even a bit.

      You should probably care a little bit, since this is a legitimate scenario where two people hear different things, and might act differently as a result. I'm not going to stay awake at night over this, but it is a little bit interesting given that "yanny" and "laurel" should not sound at all alike.

      I don't particularly blame you for not paying attention to twitter, near as I can tell only people who live under rocks actually de

      • by gweihir ( 88907 )

        Let me check....no, still do not care. The thing is, people may hear exactly the same words and still understand completely different things and that already is a much, much larger problem than this non-story here. Many people are so disconnected from reality that you can explain things carefully, in different ways and with proof, and they will still understand the stupid thing they think is right. The problem in the audio-path is _minor_ compared to what people themselves do to the date they are given.

    • Tell me about it. "Unless you've been living under a rock" gets used in horrible contexts similarily to "common sense". "My political view, which half the population disagrees with, is just common sense"

      In this case, hearing about an odd audio file is something society is near universally aware of (which is what "unless you've been living under a rock..." is supposed to imply)? Some how I really doubt that.

    • Then why are you on Slashdot? It doesn't get much more nerdy than discussing the intricacies of audio compression and how the brain processes auditory input.
  • by jtgd ( 807477 ) on Friday May 18, 2018 @03:36AM (#56631372)

    One thing I haven't heard anyone mention is a third hearing for the word. When I listen to the 'Yanny' version what it sounds like to me is 'Yarry' (starts with Y, rhymes with 'Larry').

    Does anyone else hear it that way?

    • Re: (Score:2, Troll)

      Indeed. I only hear 'Yelly' or 'Yarry' when the slider is on the far right. 'Laurel' everywhere else.
      The only possible, objective conclusion is that people hearing "Yanny" are complete morons.

      • by shanen ( 462549 )

        Probably means you're an old guy with little sensitivity to higher frequencies, or that you have a really cheap computer. Unless it's both.

        What's annoying me is that I suggested the slider approach before I heard that the NY Times had created one. They should provide a numeric scale that you can use to compare your hearing to other people's or to compare one sound system to another.

      • I can hear yanny across the entire range. But if I think 'laurel' while it's to the left then I can hear laurel.

        I had to pause it, wait, and then restart to hear laurel at all to start with. Now with the slider half way to the left I can switch between the two at will.

        I used to have good HF hearing - the 15kHz whistle from CRT TVs used to annoy me. Obviously that was many years ago and I be amazed if I could still hear those frequencies.

    • by rikkards ( 98006 )

      If I slowly move through the slider on the page in the summary, I hear yarry as I move it to the right. I predominantly hear laurel all the time when slowly sliding but if just jump from the left to the right then I hear it.

    • by necro81 ( 917438 )
      I hear "Yalie", as spoken by Ned Gerblansky [youtube.com]
    • All I heard was noise and compression artifacts. I tried to make sense of it but it was too hazy and indistinct to come up with a word out of it. I do hear a 'Y' type sound at the beginning, barely.

      When a coworker showed this to me, I failed to appreciate it. It was just nonsense really. I imagine there are millions of sounds that can be made that can appear to sound one way or the other when trying to make sense of it.

      TL;DR, some people, when looking at clouds, see a rabbit while someone else looking at th

    • One thing I haven't heard anyone mention is a third hearing for the word. When I listen to the 'Yanny' version what it sounds like to me is 'Yarry' (starts with Y, rhymes with 'Larry').

      Does anyone else hear it that way?

      I hear "Yammy".

    • me too. But I also hear Laurel in the native recording. Maybe I'm mispronouncing Yanny in my head. I think of that musician - so would expect to year "yawn-ee" or even "ye-ann-e" (Like Danny but starting a Yee). Yee - Annie.?! It's an italian name :-)

      Nothing has sounded like Yanny yet. I hear what you hear - and requires imagination at that.

      But I don't get the "controversy" I've ignored it all week as it didn't seem interesting. This like "are these two colors the same?!" Yes.

  • by Cafe Alpha ( 891670 ) on Friday May 18, 2018 @04:13AM (#56631448) Journal

    Maybe if you have some horrible laptop with no base and crackly highs you might hear Yanny.

    • by jrumney ( 197329 )

      I listened to it first on my phone's crappy speaker, and still heard Laurel. So I posted on my friend's facebook that I heard Yanny, just to troll her. I figured everyone else who "heard" Yanny was doing the same.

      Even with the NYT slider, I just hear it getting distorted until right up the end it is so distorted that the L sounds at either end disappear. To me it sounds more like "Gary" than "Yanny"

      .

      On the second of the two times he says it, I hear some distinct noise in the high frequencies that comes j

    • Maybe if you have some horrible laptop with no base and crackly highs you might hear Yanny.

      I honestly thought that it was a trick the first time I listened. I had my device on cellulary service and, with pretty decent headphones heard “Yanny”. A few minutes later I was inside and on Wi-FI (with a different IP address, obviously), and it was so clearly Laurel I thought it was a completely different clip. So now I wonder if perhaps there was some issue with my cellular provider recompressing the audio or something. I have no idea.

    • by nwaack ( 3482871 )
      Umm, no. This has everything to do with computer synthesized speech and a whole host of complicated ways our brains interpret language, and practically nothing to do with the speakers themselves. And yet this is modded information, ugh.
      • It has plenty to do with the speakers, as well as the compression used. The fact that it is a computer synthesized speech has almost nothing to do with the issue. Listening to a non-compressed, non-mangled version of the computer synthesized voice has no problem discerning the word as being "laurel". People have even tracked it back to a specific site which created the synthesized voice, and discerned what software created it. The problem has everything to do with digital compression, digital to analog conv
  • No! It's Gold! No, white.... doh!
  • by reboot246 ( 623534 ) on Friday May 18, 2018 @05:08AM (#56631564) Homepage
    I heard Yanny and Laurel, but after a while I heard Jelly.
  • This country is fucked. It's YARY!
  • by aglider ( 2435074 ) on Friday May 18, 2018 @06:33AM (#56631786) Homepage

    Nonetheless I don't spend time in this foolish stuff. Just slashdot.

    • Understanding how someone can react differently to different things is not foolish, it's due diligence.

  • For Yanni's new album, "Laurel."

  • The "internet" is stupid.

    What's next, an optical illusion that looks like a vase, no, wait, it looks like two faces, no wait, it looks like a vase...

    • The most amusing thing I've seen about this was actually a picture of some loose cursive text that could be read as either Yanny or Laurel.

    • What's next, an optical illusion that looks like a vase, no, wait, it looks like two faces, no wait, it looks like a vase...

      Well we all know that one and it was a well understood phenomenon. This on the other hand..., please show me a long history of materials going back through the ages of how these two words can be misheard.

  • by ledow ( 319597 )

    And then I get people insist they can tell the difference between high-encoding-bitrate MP3s and raw audio.

    • And then I get people insist they can tell the difference between high-encoding-bitrate MP3s and raw audio.

      And I can... on exactly one MP3 I've tried so far. One particular sound effect used goes all crunchy, and sounded much better on the original CD. Even using lame. To me, that's a fairly acceptable record, since I don't have unlimited storage space.

  • I went to the Vocabulary.com site for Laurel, [vocabulary.com] and the clip there sounds totally different than the one circulating the internet. As the summary notes:

    ...the clip that you most likely heard has accentuated higher frequencies due to imperfections in the audio that were created by data compression.

    Basically, the clip was manipulated.

  • The only way you would've heard "yanny" is if you have no idea what shitty audio compression sounds like. You can use this thing [nytimes.com] and slide it all the way over to "yanny" and if you're familiar with this sort of thing it still sounds like "laurel" with awful compression.
  • Unless you've been living under a rock for the past few days

    FTFY

    Addicted to social media

  • I don't think I'd agree that this phenomenon "reveals flaws in how we listen to audio". "Flaw" is the wrong word. It's more that this reveals that "listening" is not an objective process, to people who didn't already know that.

    When you hear things, you are never hearing a sound that exists in the world. There are "sound waves", i.e. vibrations in air, that exist in the world, but your perception of a sound is not a perfect, immediate, or objective experience of those vibrations. The vibrations hit your

  • I have no idea why the summary/article calls the slider "pitch shifting":

    Pitch shifting the audio clip up seems to accentuate "laurel" whereas shifting it down accentuates "yanny."

    It's EQ shifting [wikipedia.org]. Sliding to the left is a low-pass filter, while sliding to the right is a high-pass filter.

  • so yes I live under it and enjoy it.

  • by thegreatbob ( 693104 ) on Friday May 18, 2018 @01:25PM (#56633958) Journal
    My roommate tried showing this to me, and given the context, my brain went into hyper-analytical mode. On the first listen, it sounded very much like both of the words played in different tonalities and relatively coherent modulation (as though they were rendered with some sort of speech synth), though Laurel was more understated on the speakers in use (internal TV speakers). If I had not been tainted by assumptions of what to expect, I'd most likely have heard Yanny or Ronny or similar. Psychoacoustics is fun :3, and I'm curious to how they produced the sound byte (may already be known, but I haven't had a chance to follow up on it)
    • After looking into it, it seems my assumptions were completely ill-founded, and it was produced from a recording of the word laurel. Neat stuff.
      • Also interesting that if I actually attempt to focus on the word itself, I perceive laurel very readily. If not paying attention, yanny comes through sharply.
  • And than there are those of us (like me) who have largely lost the ability to hear certain frequencies, over the years. From my childhood and into at least my teen years, I could hear a constant high-pitched whistle emitting from any CRT screen from the very second it was flipped on, and until it was powered down again. Loud and clear. From another room. Across the entire damned house. It was actually quite annoying... but I'm pretty sure my parents didn't entirely believe me, because, well, they couldn't hear it. Fast forward to today, and I have to move over to within inches of an old television to hear that same noise. (But, like the skeleton, "I remember...")

    So really, I view this "Yanny/Laurel" thing as just a slight variation on the old Mosquito noise trick based on that same premise, that teens sometimes used as adult-proof ringtones and adults sometimes use to drive away annoying kids. Kind'a makes me wonder how long it'll be before people start making even more sophisticated "Mosquito" messages, which say one thing to adults and something else entirely to kids... High pitched poop jokes, maybe? Directives to get-off-my-lawn?

    Yes-siree, the future sure is bright...

Don't tell me how hard you work. Tell me how much you get done. -- James J. Ling

Working...