3 ms·

The Resolution of the Human Eye Tested Is 10MP

13 points by jhonovich 10y ago

ralusek 10y ago
I wonder what the other experiments were that were performed that came to such wildly different figures.
Someone 10y ago
https://en.wikipedia.org/wiki/Cone_cell https://en.wikipedia.org/wiki/Cone_cell claims (with caveats) around 100M light sensitive cells (90M rods and 5M cones), but that doesn't mean 100 megapixels for various reasons. Some differences between the human eye and a typical digital camera: - resolution varies hugely over the retina - color sensitivity varies hugely over the retina (you can't see color outside of the fovea) - the optic nerve has only about 1M ganglion cells (https://en.wikipedia.org/wiki/Optic_nerve#Structure https://en.wikipedia.org/wiki/Optic_nerve#Structure), so there is significant 'compression' very early on ("In the fovea, which has high acuity, these ganglion cells connect to as few as 5 photoreceptor cells; in other areas of retina, they connect to many thousand photoreceptors.") - bandwidth of the optic nerve is estimated at 10 megabit per second (https://www.eurekalert.org/pub_releases/2006-07/uops-prc072606.php https://www.eurekalert.org/pub_releases/2006-07/uops-prc0726...) (1) - that is very soon further compressed; I can't find decent estimates for what can be consciously seen, but am sure it is less than 1kB/second (corollary: photographic memory doesn't exist) => "The eye is not a camera" would be a better answer. (1) as with almost all of these estimates, a factor of ten is nothing.
- jhonovich 10y ago
  "The eye is not a camera" Yes, that is obviously true. The point is that the eye has the effective resolution, in terms of reading an eye chart, of a 10MP camera.
  - Someone 10y ago
    No, it doesn't. If you zoom in on those photos, you can see that the camera images entire lines of text at the same resolution. A camera can 'read' the entire chart in one snapshot. The eye doesn't work that way. For example, it is anisotropic in resolution; at the fovea, it has more 'pixels', but away from it, it has way fewer. Effect? You cannot read a line from a Snellen chart without moving your eyes. The eye doesn't take snapshots.
jsjohnst 10y ago
Curious why they used IP cameras, my guess is it had less to do with science and more to do with who they are. Reason I point this out, notice how blocky the images in general are, especially with lower light? Take any prosumer/pro digital camera of the same resolution and you'll get significantly better results, especially at 1 lux. To reword the title to be a lot more accurate, "the resolving power of the human eye is roughly equivalent to a 10MP IP camera in a well lit room". Definitely wordier and less click baity, but accurate to the "analysis" they did.

13 points by jhonovich 10y ago

I wonder what the other experiments were that were performed that came to such wildly different figures.

https://en.wikipedia.org/wiki/Cone_cell https://en.wikipedia.org/wiki/Cone_cell claims (with caveats) around 100M light sensitive cells (90M rods and 5M cones), but that doesn't mean 100 megapixels for various reasons. Some differences between the human eye and a typical digital camera: - resolution varies hugely over the retina - color sensitivity varies hugely over the retina (you can't see color outside of the fovea) - the optic nerve has only about 1M ganglion cells (https://en.wikipedia.org/wiki/Optic_nerve#Structure https://en.wikipedia.org/wiki/Optic_nerve#Structure), so there is significant 'compression' very early on ("In the fovea, which has high acuity, these ganglion cells connect to as few as 5 photoreceptor cells; in other areas of retina, they connect to many thousand photoreceptors.") - bandwidth of the optic nerve is estimated at 10 megabit per second (https://www.eurekalert.org/pub_releases/2006-07/uops-prc072606.php https://www.eurekalert.org/pub_releases/2006-07/uops-prc0726...) (1) - that is very soon further compressed; I can't find decent estimates for what can be consciously seen, but am sure it is less than 1kB/second (corollary: photographic memory doesn't exist) => "The eye is not a camera" would be a better answer. (1) as with almost all of these estimates, a factor of ten is nothing.

jhonovich 10y ago
"The eye is not a camera" Yes, that is obviously true. The point is that the eye has the effective resolution, in terms of reading an eye chart, of a 10MP camera.
- Someone 10y ago
  No, it doesn't. If you zoom in on those photos, you can see that the camera images entire lines of text at the same resolution. A camera can 'read' the entire chart in one snapshot. The eye doesn't work that way. For example, it is anisotropic in resolution; at the fovea, it has more 'pixels', but away from it, it has way fewer. Effect? You cannot read a line from a Snellen chart without moving your eyes. The eye doesn't take snapshots.

jsjohnst 10y ago

Curious why they used IP cameras, my guess is it had less to do with science and more to do with who they are. Reason I point this out, notice how blocky the images in general are, especially with lower light? Take any prosumer/pro digital camera of the same resolution and you'll get significantly better results, especially at 1 lux. To reword the title to be a lot more accurate, "the resolving power of the human eye is roughly equivalent to a 10MP IP camera in a well lit room". Definitely wordier and less click baity, but accurate to the "analysis" they did.