Bigger sensor = more light

[Casper Warnich]

I've been trying to find out, why bigger sensors are more sensitive to light than smaller sensors are. But I can't quite figure it out, so maybe someone here can.

Is it all down to how big the pixels in the sensor are?
I understand that large sensors usually have bigger pixels than small ones, and that they can contain more light photons in each pixel cell.
Does that mean, that two sensors (a large and a small) with the same pixel cell sizes in theory would be equally sensitive to light? (I understand that they wouldn't have the same total of pixels)

Also, I've heard the statement that a 2/3'' camera would be better than a S35 camera for some type of work - such as ENG - because it has a deeper DOF. But if the S35 is more sensitive to light, wouldn't you be able to stop down the S35 lens to get the 2/3''s DOF, and the same amount of light sensitivity? Or is that not proportional?

Furthermore, if two sensors have different sensitivities to light, but the same native ISO, does that mean that the more sensitive sensor will have a higher exposed image at it's native ISO than the less sensitive one?

I hope this all makes sence, and that someone can educate me

Casper

 

[Jeff Kilgroe]

First of all, why do you say say larger sensors are more sensitive?

The sensitivity of a sensor comes down to the physical attributes of the sensor elements, how efficient each photosite is in registering light, how much surface area of each photosite is used, etc..

There is nothing specific out there that dictates that larger sensors and/or sensor elements are more sensitive. ...Other than some false information that tends to mislead people at times.

As for larger format cameras / sensors, even larger film formats and lenses designed to cover them, they often must deal with less light compared to smaller format cameras when comparing lenses of the same focal length.

 

[Pietro Impagliazzo]

Furthermore, if two sensors have different sensitivities to light, but the same native ISO, does that mean that the more sensitive sensor will have a higher exposed image at it's native ISO than the less sensitive one?

Isn't ISO directly related to sensitivity?

If they both have the same ISO, they'll both be equal in terms of how much light they need.

 

[Casper Warnich]

Thanks for your answer :smile:

First of all, why do you say say larger sensors are more sensitive?

The sensitivity of a sensor comes down to the physical attributes of the sensor elements, how efficient each photosite is in registering light, how much surface area of each photosite is used, etc..

Maybe "sensitive" isn't the right word. By more sensitive to light I mean it has what people call "better low light performance" - which is essentially what I am trying to understand :smile:

I know that some sensors are more efficient in extracting information from each pixel than others, but that's not what I'm interested in. I'm just trying to understand the basic math.

So suppose every sensor in my prior post has the same efficiency, would it make more sense?

As for larger format cameras / sensors, even larger film formats and lenses designed to cover them, they often must deal with less light compared to smaller format cameras when comparing lenses of the same focal length.

I'm not sure I understand this, but I think you're saying that larger sensors need lenses with larger focal length, and that those are often slower than lenses used on smaller format cameras (at least that's what I think you ment). But as I said, I am only interested in the math of the sensor / format.

 

[Jeff Kilgroe]

Isn't ISO directly related to sensitivity?

If they both have the same ISO, they'll both be equal in terms of how much light they need.

Yep... Pretty much. And if one sensor is significantly more sensitive than another, it's not going to have the same native ISO rating. But would in turn perform better at lower ISO values due to its increased sensitivity... generally speaking.

Maybe "sensitive" isn't the right word. By more sensitive to light I mean it has what people call "better low light performance" - which is essentially what I am trying to understand :smile:

Sensitive is exactly the right word. That is what sensitivity is.

I know that some sensors are more efficient in extracting information from each pixel than others, but that's not what I'm interested in. I'm just trying to understand the basic math.

So suppose every sensor in my prior post has the same efficiency, would it make more sense?

There really is no "basic math". But to simplify as much as possible, one could assume that all sensors being considered would have 100% efficiency. Therefore, every photon of light that collides with the sensor is registered and accurately measured. So then, what it all comes down to is a matter of optics and how the light transmission occurs from the source and onto the image plane inside the camera. You don't gain sensitivity by having a larger imaging plane. Doesn't work that way. If efficiency is 100%, then the real factor is getting as much light as possible to the sensor for a number of photons per square nanometer, it becomes an issue of lenses and optics.

I'm not sure I understand this, but I think you're saying that larger sensors need lenses with larger focal length, and that those are often slower than lenses used on smaller format cameras (at least that's what I think you ment). But as I said, I am only interested in the math of the sensor / format.

Focal lengths are specifically designated by their lengths in millimeters, it doesn't matter what type of lens you are using. However, larger image formats (larger sensors or film sizes) require a larger image circle projected by any given lens, this is mostly independent of focal length except that it's more difficult to make shorter focal lengths that cover larger image formats.

Because focal lengths are constant, they always represent the same angle of view per mm of area on an imaging surface. So a 50mm FF35 lens projects the exact same image onto a smaller APS-C sensor that a 50mm APS-C format lens would. This is actually a whole other area of discussion and one that gets discussed around here quite often.

But in general, lenses designed to cover larger image formats will have longer focal lengths relative to smaller formats. For example, 100mm is considered a long or "telephoto" lens on Super35 format, but it is actually a wide angle focal length when used on a medium format camera. For Super35 or 35mm cameras, 18mm is considered a wide angle, but it's a more moderate or "normal" lens when used on a 16mm or 2/3" camera.

As lenses get longer and imaging circles increase to accommodate larger image formats, it becomes increasingly less practical to produce physically larger lenses / optics to provide greater light transmission and larger aperture sizes. So while it's common to find an 85mm f1.4 lens for 35mm cameras, you are not likely to find an 85mm with an aperture much larger than about f3.5 for a 6x4.5 camera.

On the flip side, small image formats are often crippled by small optics since the primary reason to use smaller formats is to use those small optics. But it is possible to have 1/3" cameras with very "fast" lenses that gather lots and lots of light, if you don't mind lugging around a huge lens to make it happen. Think of how telescopes work -- larger telescopes are built to gather more light.

 

[Casper Warnich]

Thank you for taking to time to explain this to me.
I think I get it.


I also found the answers to my specific questions on this site:
http://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm

Like the 2/3'' DOF vs. S35:

If you don't take the different sensitivities of the sensors into consideration, but focus on getting the same DOF, then you would need a much longer exposure time for an S35 to gain the same DOF as on a 2/3''. Much longer, I think, than you could make up for by having a more sensitive sensor.

He also states that:
"No matter what the pixel size, larger sensors unavoidably have more light-gathering area. Theoretically, a larger sensor with smaller pixels will still have lower apparent noise (for a given print size) than a smaller sensor with larger pixels (and a resulting much lower total pixel count). This is because noise in the higher resolution camera gets enlarged less, even if it may look noisier at 100% on your computer screen. Alternatively, one could conceivably average adjacent pixels in the higher pixel count sensor (thereby reducing random noise) while still achieving the resolution of the lower pixel count sensor."

 

[Jeff Kilgroe]

DOF is a property of focal length and aperture size and has nothing to do with exposure time specifically. Smaller sensors have more apparent DOF because they use a smaller section of the lens image circle, which in turn greatly effects framing and size relationships. So what they're saying about S35 requiring longer exposure times is correct in a round about way. To gain the same apparent deep DOF of a smaller image format, like 2/3", you would need to close the aperture, thus letting in a lot less light, meaning longer exposure time. But to say it simply requires longer exposure time is incorrect, because there are numerous ways to achieve longer exposure time, such as filtration, which don't include a change in aperture size, thus not affecting DOF.

And yes, higher resolution sensors provide a visually less noisy image. Even if a higher resolution sensor is just as noisy as a lower resolution sensor on a per-pixel basis, the extra resolution still brings in more detail, which in turn mitigates the noise effect. That is the best way to combat noise, other than to improve sensitivity and accuracy, is to increase the amount of resolution or detail.

Larger sensors do have the theoretical ability to capture more light. Bigger area = more photons hit the imaging surface, assuming all other illumination factors are equal. However, while a larger sensor can capture more light information, it doesn't increase overall sensitivity. Total light capturing ability would be the total number of photons registered in a given amount of time. Actual sensitivity would be the total number of photons registered within a specific area of the sensor in that given amount of time.

An analogy would be to think of photons as rain drops. Then think of the sensor as an array of small sponges and each sponge is like one pixel or photosite on a sensor. The total light or rain gathering ability of your sensor would be the total of all rain drops absorbed by your sponges. The actual sensitivity would be the number of drops absorbed by each sponge. So a 200x200 array of sponges would collect 100 times as much water as a 20x20 array of identical sponges. However, each sponge would still collect the same amount of rain water. It is often argued that using larger photosites will increase sensitivity, but that is inherently false as well. The theory behind that proposition is that if you use larger sponges, let's say now we have a 100x100 sponge array that is the same dimensions as the previous 200x200 array. We still collect the same amount of rainfall, but each sponge is collecting 4X the amount of water. But once again, each sponge also accounts for 4X the area, so while each sponge may register more rain drops, we lose accuracy and still only gain the same number of rain drops per any given area of sponge surface.

Another way to put it is that larger sensors do collect more light. However this also comes with a proportionate increase in sensor surface area, therefore you're not gaining actual sensitivity by making a larger sensor or by making individual sensor elements larger. On the contrary, it's actually easier to increase light gathering ability by using larger optics and focusing down to a smaller imaging surface. For example, taking a lens that provides coverage for 65mm film and using optical correction to condense the 65mm image circle down to an 18mm image circle to cover 16mm film. Think of it as burning ants with a magnifying glass; condensing gathered light into a smaller area. This is yet another discussion and also leads to other problems or attributes of sensors and optics to consider.

With current technology, sensor performance is increasing through better sensor design. Sensors are getting larger and cheaper every day, but the increases in sensitivity are coming from the design elements themselves and not from sensor or pixel / photosite size.