Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Human speech may have a universal transmission rate: 39 bits per second

sciencemag.org

Post reply on HN
Print view View original

91–100 of 263 posts

Re: Human speech may have a universal transmission rate: 39 bits per second

#91

Post by dredmorbius »

knzhou wrote: post #49
ilamont wrote: post #15

But, he says, instead of being limited by how quickly we can process information by listening, we’re likely limited by how quickly we can gather our thoughts. That’s because, he says, the average person can listen to audio recordings sped up to about 120%—and still have no problems with comprehension. Some years ago I worked on an accessibility project for an app and website designed for people with disabilities. One…

I've known a lot of people that push podcasts, videos, and audiobooks to extreme speed. I knew a guy who'd turn video speed up to 8x so he could binge watch a season of generic anime in an hour flat. I knew a girl who'd get through paperback romance novels by scanning each page diagonally, in 10 seconds each. And here in this thread we have a lot of people bragging along the same lines. I just don't get the point. If…

Exceedingly slow narration, particularly what's normal for audiobooks, is annoying to me because it's slower than I process words. It's like walking with someone whose pace is far slower than your natural gait -- it takes more energy and concentration to slow down. It's why slow-talkers are so annoying.

This isn't "how fast can I go through this" but "what is a comfortable pace"?

So I bump the speed up, though usually fairly modestly: 1.25x - 1.5x is generally enough.

I've noticed that preferred speeds vary tremendously with the quality of the work and speaker -- high-density information and an exceedingly good speaker, and I'll slow down. Slapdash redundant content and poor speaker, I'll speed up.

The degree of polish in the production matters tremendously. I've listened to CPG Grey's YouTube videos (highly polished) and podcasts (a lot of chit-chat with his co-host). The videos work well at normal speed, or perhaps slighly sped up. The podcasts I find nearly unlistenable, though they improve at much higher speeds (1.75x - 2x).

Re: Human speech may have a universal transmission rate: 39 bits per second

#92

Post by ekanes »

zik wrote: post #73
knzhou wrote: post #49

Earlier quoted context omitted.

I've known a lot of people that push podcasts, videos, and audiobooks to extreme speed. I knew a guy who'd turn video speed up to 8x so he could binge watch a season of generic anime in an hour flat. I knew a girl who'd get through paperback romance novels by scanning each page diagonally, in 10 seconds each. And here in this thread we have a lot of people bragging along the same lines. I just don't get the point. If…

Maybe "speed readers" still receive knowledge at around 39bps - they just filter out a lot more.

I think there's something to that. You're zipping through just listening for the highlights. Similar to how you can listen to a full podcast, or often just watch highlight clips on Youtube and feel like you "mostly" enjoyed it. You got the best stuff. Obviously that isn't what some people want, but it's an interesting optimization.

Re: Human speech may have a universal transmission rate: 39 bits per second

#93

Post by dredmorbius »

JimBrimble35 wrote: post #24

I would love to see the impact of acronyms on information density. It would also be interesting to see how many bits per second the average human maximum is.

YMMV, TLAs increase s/n but leverage ROI on prefamiliarisation.

Language is symbolic, all words are pointers. Whether you collapse complexity through an Apollonian use of religious icons or through initialisations and acronyms likely matters little.

But: TANSTAAFL.

Re: Human speech may have a universal transmission rate: 39 bits per second

#95

Post by knzhou »

dredmorbius wrote: post #91
knzhou wrote: post #49

Earlier quoted context omitted.

I've known a lot of people that push podcasts, videos, and audiobooks to extreme speed. I knew a guy who'd turn video speed up to 8x so he could binge watch a season of generic anime in an hour flat. I knew a girl who'd get through paperback romance novels by scanning each page diagonally, in 10 seconds each. And here in this thread we have a lot of people bragging along the same lines. I just don't get the point. If…

Exceedingly slow narration, particularly what's normal for audiobooks, is annoying to me because it's slower than I process words. It's like walking with someone whose pace is far slower than your natural gait -- it takes more energy and concentration to slow down. It's why slow-talkers are so annoying. This isn't "how fast can I go through this" but "what is a comfortable pace"? So I bump the speed up, though usuall…

I'm totally with you, I speak at effectively 1.5x in real life, so that's what feels natural to me listening. But I get turned off if something needs to be sped up _beyond_ reasonable human speaking speed to be palatable, like 2.5x or 3x. It's like putting a pile of salt on really bland food. I get the feeling that I'm panning for gold.

Re: Human speech may have a universal transmission rate: 39 bits per second

#96

Post by bradknowles »

dfhdshfff wrote: post #71
knzhou wrote: post #49

Earlier quoted context omitted.

I've known a lot of people that push podcasts, videos, and audiobooks to extreme speed. I knew a guy who'd turn video speed up to 8x so he could binge watch a season of generic anime in an hour flat. I knew a girl who'd get through paperback romance novels by scanning each page diagonally, in 10 seconds each. And here in this thread we have a lot of people bragging along the same lines. I just don't get the point. If…

I used to watch a lot of lectures at a high speed. But I've came to realize that the faster I watch something, the faster I forget it. It is like the infromation doesn't have the time to settle in my memory, despite me understanding it. It's maybe because when things are slow, I can use the dead time to think about the implication/corner cases of what's being said.

If you really want to understand it, I would encourage you to write down notes about it by hand.

For some reason I do not yet understand, the motion of physically putting pen(cil) to paper helps ingrain that information into your brain, in a way that typing it into a computer does not.

Re: Human speech may have a universal transmission rate: 39 bits per second

#97

Post by js2 »

knzhou wrote: post #49
ilamont wrote: post #15

But, he says, instead of being limited by how quickly we can process information by listening, we’re likely limited by how quickly we can gather our thoughts. That’s because, he says, the average person can listen to audio recordings sped up to about 120%—and still have no problems with comprehension. Some years ago I worked on an accessibility project for an app and website designed for people with disabilities. One…

I've known a lot of people that push podcasts, videos, and audiobooks to extreme speed. I knew a guy who'd turn video speed up to 8x so he could binge watch a season of generic anime in an hour flat. I knew a girl who'd get through paperback romance novels by scanning each page diagonally, in 10 seconds each. And here in this thread we have a lot of people bragging along the same lines. I just don't get the point. If…

As I used to tell my boss: a man page read once thoroughly is worth more than ten-times skimming it quickly.

Re: Human speech may have a universal transmission rate: 39 bits per second

#98

Post by nickjj »

I think it also really comes down to what videos or audio you listen to. In other words, the person delivering the audio and its quality.

I listen to most non-entertainment videos / podcasts at 1.5x and I would say about 80% of them are completely ok to comprehend from a "can I easily listen to this without struggling to figure out what they are saying at a language level". But as soon as I try 2x, that drops down to maybe 30-50% because the person speaking isn't speaking clear enough, they have an accent that overpowers any type of comprehension ability on my part or the audio quality is too poor and it introduces too many artifacts.

Sometimes I ask my friends to watch the same video at 2x to see if they can comprehend it and often times they can't (but sometimes they can). We're all in the same area.

I generally find a neutral accent and very clear annunciation helps the most. I've had a bunch of people say they've watched my videos / courses at 2x without issues because apparently I have no accent which is something I've heard from a number of people in different countries where English isn't their native language. I find it interesting because I've also heard a decent amount of people say I speak very fast at 1.0x speed, so I do believe accents and annunciation has at least some role in this.

Does anyone know of any software where you can feed it an English audio sample and it spits back the number of syllables per second? Seems like a pretty cool potential ML project.

Re: Human speech may have a universal transmission rate: 39 bits per second

#99

Post by knzhou »

bradknowles wrote: post #96
dfhdshfff wrote: post #71

Earlier quoted context omitted.

I used to watch a lot of lectures at a high speed. But I've came to realize that the faster I watch something, the faster I forget it. It is like the infromation doesn't have the time to settle in my memory, despite me understanding it. It's maybe because when things are slow, I can use the dead time to think about the implication/corner cases of what's being said.

If you really want to understand it, I would encourage you to write down notes about it by hand. For some reason I do not yet understand, the motion of physically putting pen(cil) to paper helps ingrain that information into your brain, in a way that typing it into a computer does not.

Some people say writing is better because typing is "unnatural", but from an evolutionary standpoint both are extremely, equally unnatural activities. There's no reason moving a cylinder of graphite with one hand should be inherently better matched to the brain than moving squares of plastic with two.

My personal guess is that, with fast typing speeds, it's too easy to just copy things word for word. With writing, you have to at least rephrase it and reorganize it to fit the notes reasonably on the page, which forces some processing to occur. I take notes solely by typing, but I only have retention if I do it slowly and reflectively.

Re: Human speech may have a universal transmission rate: 39 bits per second

#100

Post by ckosidows »

nickjj wrote: post #98

I think it also really comes down to what videos or audio you listen to. In other words, the person delivering the audio and its quality. I listen to most non-entertainment videos / podcasts at 1.5x and I would say about 80% of them are completely ok to comprehend from a "can I easily listen to this without struggling to figure out what they are saying at a language level". But as soon as I try 2x, that drops down to…

Where are you from if people think you don't have an accent, if you don't mind me asking?
Post reply on HN