Are bigger pixels less noisy?

Started 7 months ago | Discussions thread
D Cox Forum Pro • Posts: 29,891
Re: Are recent sensors less noisy?

JahnG wrote:

alanr0 wrote:

JahnG wrote:

Is there any consensus if pixel size influences how noisy a sensor is?

My questions are asked from the view of a ”for fun” photographer using APSC cameras and a ”1” camera, especially thinking about the noise level in low light photography.

In some DPR forums there are every now and then long threads about sensor size, pixel count and pixel size . Some experts or ”experts” write in DPR that sensors with bigger ”pixels” are less noisy, some think there is no difference.

Camera manufactors like Canon, Sony, (Panasonic) have made ”low noise” cameras with less and bigger ”pixels”. Do the manufactors not know what they are doing? (Some people in DPR say that such cameras are manufactored because the customers expect that sensors with bigger ””pixels” would be less noisy?)

In some technical magazines where also cameras are dealt with they tend to think that sensors with bigger ”pixels” produce less noise?

What do you say in this forum:

  1. Is it possible to generally say that FF sensors with fewer and bigger ”pixels” are/ are not, less noisy (to any practical degree) than FF sensors with smaller ”pixels”. (Same for APSC sensors)
  2. Some DPR members also dispute whether FF sensors (regardless of pixel count ?) are less noisy (better for low light photography) than smaller sensors. One argument is that the existing difference in noisiness between FF cameras and small sensor cameras basically depends on the fact that small sensor cameras have smaller lenses and thus the lens collects less light. According to this FF sensors and smaller sensors would be equally noisy?

As Eric Fossum notes, there are well-established practical and theoretical reasons why larger pixels are at least as noisy as smaller pixels.

For many recent cameras, the differences are typically very small, and under typical circumstances sensor read noise will be much less than the unavoidable Poisson-distributed noise associated with photon arrival statistics.

Read noise can be as low as 1 electron RMS (or even less). For comparison, if you capture (on average) as few as 25 photons per pixel in the shadow regions of your image, the Poisson standard deviation in the signal will be 5 electrons, and an additional 1e- rms will increase the noise by only around 2%.

For such a sensor, read noise becomes significant only when the signal to noise ratio due to photon statistics is already very poor.

Regarding sensor size, a larger sensor does allow you to use a lens with a larger entrance pupil and gather more light. If you are limited by the size and weight of the lens, then sensor size will not make much difference.

A much clearer trend is that sensors from 5 or more years ago are noisier than the best modern sensors. If you are interested in historical trends:

There are measurements of older sensors on Roger Clark's Clarkvision site.

The sensorgen site is no longer maintained, but the information extracted from DxO is still available on WaybackMachine.

Bill Claff includes results for recent cameras on his photonstophotos site.

So, in the link above, the Nikon D200 had a read noise of 9.1 e-, while the D500 is reported as only 1 e-.

If you are looking for quantitative evidence, it is essential to compare sensors of comparable technological maturity.

There are trade-offs between read noise, read-out speed (less of a concern with column parallel read-out) and sensor full well capacity (which impacts dynamic range). Some sensors (notably from Sony) exploit a technique developed by Aptina to switch between a high dynamic range mode at low ISO, and a low noise mode at higher ISO speeds. Bill Claff discusses this and presents measurements here.

Thank you very much for the information.

PS, You say "a larger sensor does allow you to use a lens with a larger entrance pupil and gather more light". From that I get a "foolish" thought. Might it (theoretically?) be possible to build a low light camera using big entrant lenses collecting much light and have all this light focused on a minimal 1/2.3 sensor?? In a way like an astronomical telescope.

Jahn

Just take your f/0.9 lens from your 24x36 camera and adapt it to your M4/3 camera.

I can't see any way in which a larger sensor "allows" you to use a lens with a larger entrance pupil, or a smaller sensor forbids it.

Don Cox

 D Cox's gear list:D Cox's gear list
Sigma fp
Post (hide subjects) Posted by
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow