4 ms·

If you pay close attention in very low level light situations, you'll notice that what you do see (or at least what I see) is actually reminiscent of the RGB gr

If you pay close attention in very low level light situations, you'll notice that what you do see (or at least what I see) is actually reminiscent of the RGB graininess of low level light photographs. That graininess is to my understanding caused by the variation of the number of photons hitting that portion of the receptors, which is unnoticeable when flooded with quadjillions of them in decent lighting. So this is a nice semi-expected result in this experiment, though it does seem to be near the limit of default human perception. I wonder how well somebody could train themselves to notice single-photon events, much like musicians can notice tiny audio events and variations.

Thrender 10y ago
Probably at this stage it's all about reducing the noise. I'm not sure what that means, maybe cooling your eyeballs?
- 50CNT 10y ago
  I'm not a medical professional, but I believe that would impede their function. Eyelids weren't designed to clear off frost.
  - baq 10y ago
    they're designed to not freeze in the first place: see all the people who've been to the poles and retained sight.
- milcron 10y ago
  Related - this was on HN just the other day. Birds have a "hyperuniform" distribution of rods and cones which is superior to our own. https://www.quantamagazine.org/20160712-hyperuniformity-found-in-birds-math-and-physics/ https://www.quantamagazine.org/20160712-hyperuniformity-foun...
  - beefield 10y ago
    Sorry for being lazy and not doing proper research, but maybe someone knows this right away. Is Sobol sequence hyperuniform? If not, is there an algorithm you can use to generate hyperuniform random variables? (Basically I am thinking that it could make Monte Carlo simulation in some circumstances a bit more effective)
    - ravar 10y ago
      If you take the 2d Fourier transform of the sequence and the low frequency components go to zero the point process is hyper uniform [1] . Poison disk noise has this property but is generally slow to generate. [1] http://arxiv.org/pdf/1408.4645v1.pdf http://arxiv.org/pdf/1408.4645v1.pdf
- limbicsystem 10y ago
  I believe that Giles Brindley (https://en.wikipedia.org/wiki/Giles_Brindley https://en.wikipedia.org/wiki/Giles_Brindley) famously did exactly this - sitting in a bath of ice water to lower his body temperature in an attempt to reduce the dark noise in his retina. I can't find a copy of the paper unfortunately... EDIT: Some details about this and his more colorful work on erectile dysunction here: http://web.archive.org/web/20150204122651/http://alignmap.com/tag/gs-brindley/ http://web.archive.org/web/20150204122651/http://alignmap.co...
rcthompson 10y ago
My understanding is that the "grain" is not necessarily caused by your eyes detecting isolated photons, but rather by the electrons in the light-absorbing molecules of your rods & cones becoming spontaneously excited by ambient heat energy rather than absorption of a photon. Either way, it shows that we can indeed consciously perceive single-molecule excitation events, whatever the cause.
- ekianjo 10y ago
  > but rather by the electrons in the light-absorbing molecules of your rods & cones becoming spontaneously excited by ambient heat energy rather than absorption of a photon Agree, it's not a digital system so there is always some noise, just like you can never get perfect silence in terms of sensory perception. Your brain has also a way to project some kind of images in your mind even when you don't see anything in complete darkness, so it's far from just being the detection of photons.
  - laumars 10y ago
    > just like you can never get perfect silence in terms of sensory perception. That's a little different because your body produces lots of noises which are only perceivable in the absence of other background noises, eg when inside an anechoic chamber[1]. [1] https://en.wikipedia.org/wiki/Anechoic_chamber https://en.wikipedia.org/wiki/Anechoic_chamber
  - zero_iq 10y ago
    It's also worth noting that senses may actually take advantage of the inherent noise to improve sensitivity of low level signals using stochastic resonance -- random resonances of noise and actual signal boosts the measured signal to a level that is detectable, thus permitting detection of signals that are actually below the (noise free) detection sensitivity. Such mechanisms have been observed in animal sensory systems, and is believed to have a role in mammalian visual processing.
suchow 10y ago
You might be interested in Elizabeth Kiersted's work with Denis Pelli (NYU) about a circumstance where people can see their own "internal" noise, based on Ramachandran and Gregory (1991)'s phantasm illusion: http://psych.nyu.edu/pelli/docs/kierstead2007intel.pdf http://psych.nyu.edu/pelli/docs/kierstead2007intel.pdf
Retr0spectrum 10y ago
It's also a perfect analog to software ray-tracing: http://chunky.llbit.se/spp-compare-1.png http://chunky.llbit.se/spp-compare-1.png The more photons you simulate/see, the less grainy it gets. Of course, the human eye has no concept of "frames", so it's not easy to directly compare the two.
- devindotcom 10y ago
  There's actually a sort of native refresh rate in the eye but it isn't exact and afaik no one has really figured out how it's coded. I think I read that there is a regular drip from one layer of the retina that provides a sort of baseline rate that deviates upwards when that pathway is excited. Don't quote me on that though...
devindotcom 10y ago
Actually, I would guess that the 'noise' you see is from the highly noisy system of retinal ganglia and such... the eye is a wonderful thing but it has a few, shall we say, quirks. Check out this diagram: http://www.corpshumain.ca/en/images/Retine_couches_couleur_(FF)_en.jpg http://www.corpshumain.ca/en/images/Retine_couches_couleur_(... Note that the light comes from the TOP and passes through all those layers before hitting the photosensitive one! So - at the Earth's surface, even on a very dark night, I think you would be receiving far, far too many photons for them to each register as a sort of noise blob. More likely it is the natural noise of your nervous system as it does all the things that cells need to do and produces the slight visible side effect that's more pronounced in darkness and low contrast. Edit: Also, since I remember this coming up in my neuro classes, this has been shown in less rigorous experimentation before: http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.html http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.h...

Thrender 10y ago

Probably at this stage it's all about reducing the noise. I'm not sure what that means, maybe cooling your eyeballs?

50CNT 10y ago
I'm not a medical professional, but I believe that would impede their function. Eyelids weren't designed to clear off frost.
- baq 10y ago
  they're designed to not freeze in the first place: see all the people who've been to the poles and retained sight.
milcron 10y ago
Related - this was on HN just the other day. Birds have a "hyperuniform" distribution of rods and cones which is superior to our own. https://www.quantamagazine.org/20160712-hyperuniformity-found-in-birds-math-and-physics/ https://www.quantamagazine.org/20160712-hyperuniformity-foun...
- beefield 10y ago
  Sorry for being lazy and not doing proper research, but maybe someone knows this right away. Is Sobol sequence hyperuniform? If not, is there an algorithm you can use to generate hyperuniform random variables? (Basically I am thinking that it could make Monte Carlo simulation in some circumstances a bit more effective)
  - ravar 10y ago
    If you take the 2d Fourier transform of the sequence and the low frequency components go to zero the point process is hyper uniform [1] . Poison disk noise has this property but is generally slow to generate. [1] http://arxiv.org/pdf/1408.4645v1.pdf http://arxiv.org/pdf/1408.4645v1.pdf
limbicsystem 10y ago
I believe that Giles Brindley (https://en.wikipedia.org/wiki/Giles_Brindley https://en.wikipedia.org/wiki/Giles_Brindley) famously did exactly this - sitting in a bath of ice water to lower his body temperature in an attempt to reduce the dark noise in his retina. I can't find a copy of the paper unfortunately... EDIT: Some details about this and his more colorful work on erectile dysunction here: http://web.archive.org/web/20150204122651/http://alignmap.com/tag/gs-brindley/ http://web.archive.org/web/20150204122651/http://alignmap.co...

rcthompson 10y ago

My understanding is that the "grain" is not necessarily caused by your eyes detecting isolated photons, but rather by the electrons in the light-absorbing molecules of your rods & cones becoming spontaneously excited by ambient heat energy rather than absorption of a photon. Either way, it shows that we can indeed consciously perceive single-molecule excitation events, whatever the cause.

ekianjo 10y ago
> but rather by the electrons in the light-absorbing molecules of your rods & cones becoming spontaneously excited by ambient heat energy rather than absorption of a photon Agree, it's not a digital system so there is always some noise, just like you can never get perfect silence in terms of sensory perception. Your brain has also a way to project some kind of images in your mind even when you don't see anything in complete darkness, so it's far from just being the detection of photons.
- laumars 10y ago
  > just like you can never get perfect silence in terms of sensory perception. That's a little different because your body produces lots of noises which are only perceivable in the absence of other background noises, eg when inside an anechoic chamber[1]. [1] https://en.wikipedia.org/wiki/Anechoic_chamber https://en.wikipedia.org/wiki/Anechoic_chamber
- zero_iq 10y ago
  It's also worth noting that senses may actually take advantage of the inherent noise to improve sensitivity of low level signals using stochastic resonance -- random resonances of noise and actual signal boosts the measured signal to a level that is detectable, thus permitting detection of signals that are actually below the (noise free) detection sensitivity. Such mechanisms have been observed in animal sensory systems, and is believed to have a role in mammalian visual processing.

suchow 10y ago

You might be interested in Elizabeth Kiersted's work with Denis Pelli (NYU) about a circumstance where people can see their own "internal" noise, based on Ramachandran and Gregory (1991)'s phantasm illusion: http://psych.nyu.edu/pelli/docs/kierstead2007intel.pdf http://psych.nyu.edu/pelli/docs/kierstead2007intel.pdf

Retr0spectrum 10y ago

It's also a perfect analog to software ray-tracing: http://chunky.llbit.se/spp-compare-1.png http://chunky.llbit.se/spp-compare-1.png The more photons you simulate/see, the less grainy it gets. Of course, the human eye has no concept of "frames", so it's not easy to directly compare the two.

devindotcom 10y ago
There's actually a sort of native refresh rate in the eye but it isn't exact and afaik no one has really figured out how it's coded. I think I read that there is a regular drip from one layer of the retina that provides a sort of baseline rate that deviates upwards when that pathway is excited. Don't quote me on that though...

devindotcom 10y ago

Actually, I would guess that the 'noise' you see is from the highly noisy system of retinal ganglia and such... the eye is a wonderful thing but it has a few, shall we say, quirks. Check out this diagram: http://www.corpshumain.ca/en/images/Retine_couches_couleur_(FF)_en.jpg http://www.corpshumain.ca/en/images/Retine_couches_couleur_(... Note that the light comes from the TOP and passes through all those layers before hitting the photosensitive one! So - at the Earth's surface, even on a very dark night, I think you would be receiving far, far too many photons for them to each register as a sort of noise blob. More likely it is the natural noise of your nervous system as it does all the things that cells need to do and produces the slight visible side effect that's more pronounced in darkness and low contrast. Edit: Also, since I remember this coming up in my neuro classes, this has been shown in less rigorous experimentation before: http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.html http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.h...