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

11–20 of 263 posts

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

#11

Post by rrss »

SketchySeaBeast wrote: post #2

>Scientists started with written texts from 17 languages, including English, Italian, Japanese, and Vietnamese. They calculated the information density of each language in bits—the same unit that describes how quickly your cellphone, laptop, or computer modem transmits information. I feel like the whole "bits" calculation is a neat way to get into the media, but not actually related to "information density". Edit: Be…

The authors use syllable conditional entropy to estimate information density, and entropy is measured in bits (or shannons, if you prefer).

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

#12

Post by Aperocky »

Mathnerd314 wrote: post #8

> 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. “It really seems that the bottleneck is in putting the ideas together.” Glad this paragraph was in the article, clears up their m…

Really depends on language, if you're writing java, you'll be putting out a lot more than that due to how stupidly verbose it is.

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

#14

Post by _nhynes »

jumelles wrote: post #7

Being able to convey emotion with tone in speech likely adds a lot more "information".

The increased information density in tonal language appears to be offset by a reduced speaking rate [0].

[0] http://ohll.ish-lyon.cnrs.fr/fulltext/pellegrino/Pellegrino_...

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

#15

Post by ilamont »

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 of the team members had low vision, and used a screen reader that must have been set to 3x or even higher. I usually listen to YouTube and podcasts at 1.5-2x and I could barely understand the audio. He seemed surprised, which indicated to me that 3x+ was the norm for people in his circle.

I wonder if his ability was trained through years of using fast screen readers, vs. a lower visual processing load leads to better audio processing, or some other explanation.

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

#16

Post by vlovich123 »

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…

Yes. Took me a while but I can comfortably understand 2x speed and now 1x podcasts seem weird like they're talking super slow. I would imagine it's something you just train even more out of necessity.

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

#17

Post by jefftk »

Aperocky wrote: post #12
Mathnerd314 wrote: post #8

> 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. “It really seems that the bottleneck is in putting the ideas together.” Glad this paragraph was in the article, clears up their m…

Really depends on language, if you're writing java, you'll be putting out a lot more than that due to how stupidly verbose it is.

> if you're writing java, you'll be putting out a lot more than that due to how stupidly verbose it is

Being "verbose" means that each letter you type communicates fewer bits of information. If the bottleneck is putting ideas together then you would expect someone writing in a more verbose language to type more letters per minute but still take a similar amount of time to communicate the idea.

In practice most Java programmers are using IDEs with good auto-completion, though, so aren't actually needing to type as many letters as you'd think.

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

#18

Post by liability »

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 typically use TTS near 2.5x (I turn it up when I'm alert and down when I'm tired.) It's definitely a learned skill; a few years ago I started at 1x and struggled even with that.

Every couple of months, take a moment to reflect on your comprehension. Is it currently easy for you to understand the audio? If yes, then crank it up a little bit until it's noticeably more difficult. Repeat this process periodically over a year or so and before you know it, it'll be set pretty damn quick.

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

#19

Post by tkfu »

I'm a bit confused, here. (I went and looked at the original paper.) They estimated information density for each of the subject languages as a whole, on average:

> In parallel, from independently available written corpora in these languages, we estimated each language’s information density (ID) as the syllable conditional entropy to take word-internal syllable-bigram dependencies into account.

But the experiment uses the same text translated into each language! Why introduce this extra variable (and source of error) of estimated language-wide information density, if you are controlling your experiment such that you have the exact same information encoded in each language? That is to say, why use an _estimated_ information density when you could measure it exactly for the texts that are being spoken? Or, conversely, why go to all the trouble of having the speakers read the same text translated into each language, if you aren't going to make use of that symmetry?

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

#20

Post by jolmg »

Aperocky wrote: post #12
Mathnerd314 wrote: post #8

> 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. “It really seems that the bottleneck is in putting the ideas together.” Glad this paragraph was in the article, clears up their m…

Really depends on language, if you're writing java, you'll be putting out a lot more than that due to how stupidly verbose it is.

I also think that in writing the bottleneck is in how fast and accurate your hand can move. I would agree that English is faster than Spanish because Spanish is more verbose.
Post reply on HN