[WillT (OP)]

This week's question of the week is:

'What is the shutter speed of the eye?'

[alancalverd]

Given that single photons produce visible Cherenkov phosphenes, the minimum detectable event is probably of the order of 10^-18 seconds.

This is quite different from the subsequent processing time required to distinguish between events. Raster-scan television screens and mechano-optical cine projectors produce an image of continuous motion if the sampling frequency exceeds about 16 frames/second, and even when individual cine frames are clearly blurred (say taken at 40 mS intervals)  the projected "moving image" can appear as sharp as a still photograph.

LED lights driven at < 10 Hz may appear to flicker but at > 20 Hz the light output appears continuous.

[evan_au]

Some thoughts...

Aperture
The human pupil size varies from 2 mm diameter in bright light to 8 mm under dark conditions. Smartphone camera lenses are in a similar range.
However, consumer hand-held cameras have a light-collecting diameter of around 20mm, and perhaps 100mm with professional telephoto lenses. This means they can collect a lot more photons, under the same conditions.

Fastest Detectable Event
For both cameras and the human eye, extremely short events can be detected if they generate enough photons during the period when the camera aperture (or eyelid) is open. So this is not a characteristic of a camera or the eye.

Minimum Exposure Time
As Alan mentioned, the flicker fusion frequency of the human eye is around 20 frames per second for most people, so you could say that the minimum exposure time for the human eye is around 50ms.
- Some consumer cameras have a minimum exposure time of 1/1000 second, and are able to freeze very fast events which are a blur to the human eye.

Maximum Exposure Time
When making star trails, or using a tracking mount, astrophotography can use exposure times of hours.

The flicker fusion frequency suggests that the maximum exposure time of the human eye is somewhere around 5-10ms in bright conditions.
- Under low-light conditions, the human eye is able to integrate light over a longer period - but I suspect not more than a second. So you could call this the maximum exposure time.

Photon Sensitivity
In modern astronomical CCDs, about 80% of incoming photons striking the sensor will generate a detectable charge in a pixel (the rest are reflected, or strike wiring, etc).

I heard that the human eye requires two photons to strike the same cone cell within a short period to register a flash of light. This will reduce thermal noise, but also reduces sensitivity compared to a CCD camera.

Thermal Noise
Thermal noise generates lots of charges all the time in a CCD camera, so to improve the signal-to-noise ratio, visible-light astronomical telescopes are cooled to liquid nitrogen temperatures.
- Infra-red telescopes like James Webb must be cooled to <7K.
- With a non-cooled sensor, consumer-grade cameras need to collect a lot of photons to ensure that the image stands out above the thermal noise.
- To reduce thermal noise, consumer cameras typically use tricks like taking a reference exposure with the shutter closed, to measure the average thermal noise, and subtract that from the image with the shutter open. This works better with relatively bright objects which generate more than half of the charge of the thermal noise.

Colour Sensitivity
Consumer cameras are able to take colour photos under very low-light conditions, if the exposure time is long enough.

However, the human cone cells which detect colour are only effective under brightly lit conditions; in the dark, human eyesight falls back to the black-and white rod cells, with lower resolution than the cone cells of the central vision.

[alancalverd]

This means they can collect a lot more photons, under the same conditions.

Er, sort of. What matters is the brightness of the image, and for a given object this is principally determined by the ratio of aperture to focal length of the lens. This is conventionally specified as an inverse ratio f/n and for most cameras the largest aperture is around f/2, fairly close to the fully dilated human eye at f/3. Telephoto lenses bigger than f/4 are a bit unwieldy.

I heard that the human eye requires two photons to strike the same cone cell within a short period to register a flash of light.

Interesting. The same applies to photographic film. You need two visible photons to strike a single grain in a short time for a latent image to form. Under very low intensities, this causes "reciprocity failure" where the apparent sensitivity of the emulsion decreases with decreasing image brightness.  Astronomers used to "pre-fog" their plates to increase sensitivity - something of a black art probably lost nowadays thanks to  CCD imagers.  It also explains why fine grain photographic emulsions are less sensitive (typically ASA 50 - 64) than coarse-grain (ASA 1000) materials.

[evan_au]

Focal Length
The human eye has a fixed focal length of around 22mm, which gives a similar field of view to a traditional film camera with a 50mm focal length.

But camera lenses come in a wide variety of focal lengths - and a camera with a zoom lens might have a range of 60:1 in focal length (and corresponding wide range in field of view). 

Saccades
The human eye does not focus on just one image, but darts around parts of the image, allowing the high-resolution macula to dwell on different parts of the image for short periods of time. Saccades are the quick movements from one part of the image to another, occurring around 3-5 times per second, and lasting about 50ms each time. The brain suppresses the blurry image during the saccades, making them effectively unconscious and invisible to the viewer.
- The image collection time on one part of the retina is limited to 200-300ms, which could be considered the maximum "exposure time"
- But the brain is able to build up a detailed image across a number of saccades, forming a composite image like the popular "panorama" photo, or "stacking" in astrophotography.

An interesting comparison here... They suggest that the human eye can't see any more detail after looking at a scene for more than 10 to 15 seconds:
https://www.cambridgeincolour.com/tutorials/cameras-vs-human-eye.htm

[paul cotter]

Am I correct in thinking that without this movement(saccades) the receptors effectively saturate and become unresponsive?