Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[delphonso]

The movie, apparently:

Thulsa Doom: "Infidel defilers. They shall all drown in lakes of blood. Now they will know why they are afraid of the dark. Now they will learn why they fear the night."

[hector13]

The movie, apparently:

Thulsa Doom: "Infidel defilers. They shall all drown in lakes of blood. Now they will know why they are afraid of the dark. Now they will learn why they fear the night."

* hector13 goes a bit gruff

I’m Batman.

[Il Palazzo]

Thulsa Doom, that guy. He was a literal snake in the pants.

[scriver]

Oh but that evens out since the other Conan quote is a Mongol quote

[methylatedspirit]

I had a random thought, and I decided to search the interwebs for it. Apparently, the human brain "only" deals with 11 megabits (11*10^6) a second from the senses. And if I've read the article correctly, that's the raw data rate. Quote, "It is known, however, that the senses gather some 11 million bits per second from the environment.", so unless I'm reading it wrong, it's talking the raw data rate.

That sounds wrong for some reason. Just 11 megabits a second of raw data, of which 10 megabits are from the visual system? That's it? I'm probably gonna sound uneducated (because I am), but the retina has 120 million rod cells and 6 million cone cells, and most people have 2 eyes (I apologize if you don't, I'm just speaking about the general population). Even if I were to assume that rod and cone cells can only detect on/off (1 bit per cell), that would mean there's 252 megabits (252,000,000 bits/31.5 megabytes) going from the eyes to the brain each second. That's assuming a 1-bit bit depth. If I cranked it up to 8 bits per cell (which I feel is a conservative estimate, considering we can see color banding in 24-bit (8 bits per color) color visual media), it goes up to 2016 megabits (2,016,000,000 bits/252 megabytes) a second. Off by 1 to 2 orders of magnitude.

I also have doubts about their estimate for the ears. Just 100,000 bits a second? The human hearing range goes up to 20 KHz, give or take. That requires a sampling rate of at least 40 KHz (see Nyquist-Shannon theorem) to encode without aliasing. Human hearing has a maximal dynamic range of 120 dB. Let's pretend that this applies to the whole range of human hearing. This would be equal to 20-bit PCM audio. Multiply the sampling rate by the bit depth, and you get 800,000 bits (800 kilobits/100 killobytes) per second, at least if our hearing is encoded in 20-bit, 40 KHz PCM. Multiply by 2 because most of us have 2 working ears (again, general population), and that's 1,600,000 bits (1.6 megabits/0.2 megabytes) a second. Off by an order of magnitude.

I dunno about the other senses, but I suspect they're off by at least an order of magnitude too. The full table, for reference, is:

Information transmission rates of the senses
sensory system    bits per second
eyes              10,000,000
skin              1,000,000
ears              100,000
smell             100,000
taste             1,000

I don't claim to be any sort of expert. I'm just a kid who happens to know an effectively infinitesimal subset of all knowledge. I just think Encyclopedia Britannica might be a bit off with their estimates. Maybe some of you can enlighten me with how I'm wrong. I'd be thrilled if that were the case.

[bloop_bleep]

I think maybe they're counting after the thalamus does its information filtering? Yes, a lot of information is being received, but at any given time you're usually only paying attention to and dealing with a small subset of that information.

Also, our ears don't sample like microphones. They sample in the frequency domain, not the pressure domain, so they're gathering the amplitude of each audible frequency at any given moment, not the total pressure, since each hearing hair is tuned to a specific frequency. This might change your baud rate calculations.

Additionally, I'm not sure I would say each cone or rod has even 1 bit of information provided. There could well be averaging done. You don't see "a dead pixel" every time a cone or rod dies, I think, so maybe each cone or rod provides somewhere around 1/10 bits of information. That's again before thalamus filtering.

EDIT: I've read that there are about 1.2 million axons going from the eyes to the brain, per eye. That's 2.4 million axons altogether, so to get 10 megabits per second visual brain input you'd need about 4 bits or 16 signal levels per axon, which seems reasonable.

[Maximum Spin]

Additionally, I'm not sure I would say each cone or rod has even 1 bit of information provided. There could well be averaging done. You don't see "a dead pixel" every time a cone or rod dies, I think, so maybe each cone or rod provides somewhere around 1/10 bits of information.

That's not really how information theory works. A bit is a single yes or no question, so, at minimum, every receptor cell contains at least one bit of information: "is this cell firing y/n?" And in fact, it appears that individual neurons send several bits of information, though we're not clear on the details.

Which really leads to the larger point overall and the answer to meth's question: We have no idea how many bits of information the sensory organs accumulate, because we don't even really know how they work. We're not even totally sure whether the electrical impulse part of a nerve's firing even matters, as there's a minority theory that the actual signal could be sonic instead. Anyone who tells you otherwise is lying or sorely misled, even if they have Britannica in their name.

[dragdeler]

The only way that "raw data" would make sense to me is by considering the bits are nerve ends or taste buds or what have you. So more like a 11mbit processor with a clockrate of a few hundred hertz tops.

[bloop_bleep]

Yes, there's at least 1 bit when it fires, but when the data is actually collected it could be collected in aggregates of multiple sensors at once and then compressed to a smaller number of bits than if each rod or cone was individually sampled. That is, there could well be a single neuron connected to multiple rods or cones for noise reduction but that only outputs 4 bits or so. That's what I meant by the fact that we generally don't see dead pixels every time a rod or cone died; if each rod or cone visually represented its own area by itself then every time one of these cells died we'd have a glitch in our vision.

The only way that "raw data" would make sense to me is by considering the bits are nerve ends or taste buds or what have you. So more like a 11mbit processor with a clockrate of a few hundred hertz tops.

Maybe, but keep in mind that the brain is highly parallel, so give that processor a couple hundred million or couple billion cores and you can see why it can do what it does.

[Egan_BW]

Yes, there's at least 1 bit when it fires, but when the data is actually collected it could be collected in aggregates of multiple sensors at once and then compressed to a smaller number of bits than if each rod or cone was individually sampled. That is, there could well be a single neuron connected to multiple rods or cones for noise reduction but that only outputs 4 bits or so. That's what I meant by the fact that we generally don't see dead pixels every time a rod or cone died; if each rod or cone visually represented its own area by itself then every time one of these cells died we'd have a glitch in our vision.

It could be that those glitches do occur, but are too small to be consciously perceived. It's not like you're able to see every single "pixel" of your vision.

[bloop_bleep]

But if you can't notice it, then it's smaller than a pixel, isn't it? A pixel is the smallest thing you can notice, by like definition.

[Egan_BW]

It's possible to be watching a screen from far away enough that you don't notice any flaws on the scale of a single pixel, so I think not. A pixel is just the smallest visual unit, whether you notice it or not. Our mind lack the function to "zoom in" or view part of our vision in more detail, so it's possible there's more data there that we just can't really access

[scriver]

I would define pixel as "the smallest unit that can be displayed", as opposed to "smallest that can be seen"

[dragdeler]

-

[LordBaal]

Like it has been said, maybe they meant 11mb of info, information is different from data. Maybe raw data is actually larger.

At any rate 11mb per second mean more than half a giga per minute, and more than 154 gigas per hour, or about 3 and a half terabites of data per day. It does sound like a lot