Science of Sensor Size in Relation to Exposure

[Kyle Jones]

I've always been aware that a larger sensor with the same number of pixels as a smaller sensor will generally have better low light capabilities because the photosites can be larger to collect more light.

However, something I'm not clear on is what happens if you have the exact same pixel density on two sensors, just one is smaller. Will one have better low light capabilities than the other? Let's assume for the sake of argument that the sensor is full frame vs super35, on two different EF mount cameras with the exact same full frame EF lens on each. Again the sensors are identical, with the exception that one has been "cut down" to Super35 size.

Sorry if this is obvious but if it would be possible to explain the science/physics behind your answer, is really appreciate it!

 

[Mike P.]

Theoretically the same when looking at the image 1:1, but practically different because the larger sensor would inevitably be downscaled more giving a preceptibly sharper/cleaner image.

Case in point Epic vs. Scarlet (either sensor) or Weapon VV (8k Dragon) vs Weapon MG (6k Dragon)... or better yet, any RED camera at full sensor res vs cropped res...

 

[Karim D. Ghantous]

I think that it's quite a simple concept, but I'm going to try and use film to explain it.

Let's say you have to slides, one 8-perf 35mm and one 4-perf. They are of the same subject and they are of the same film stock. You project each one to fill the screen entirely. You know that the 4-perf slide is going to be more grainy, simply because you had to enlarge it more.

But you can make the 4-perf slide look less grainy (the same graininess as the 8-perf slide) simply by filling only half the screen. it will be a smaller image, but the graininess will be exactly the same.

It's not physics but it is an experiment. Or at least a thought experiment which can be verified by a physical experiment.

 

[Kyle Jones]

Theoretically the same when looking at the image 1:1, but practically different because the larger sensor would inevitably be downscaled more giving a preceptibly sharper/cleaner image.

Case in point Epic vs. Scarlet (either sensor) or Weapon VV (8k Dragon) vs Weapon MG (6k Dragon)... or better yet, any RED camera at full sensor res vs cropped res...

Okay, so there is no difference in exposure. Hmmm, what you say about shooting in cropped mode being the same makes sense...

I guess the only reason I am asking is I was recently on set and someone I was working with said "Well, large sensors don't need as much light as smaller ones" (since his A7S was at ISO 800 and my C300 Mark II was at 3200 to get the same exposure (no NDs and same aperture) and I replied, "Oh yeah? Because of the photosite size?" and he replied, no, even with the same size photosites the large sensor doesn't need as much light, it's just physics." I puzzled and decided to investigate later but I can't turn up any literature on this.

I recall that a speed booster can concentrate more light onto a smaller sensor, brightening exposure, but if the two lenses were both the same full frame lenses, none of that would be happening, right?

Any thoughts on why my C300 Mark II needed so much more light?

 

[Mike P.]

Probably the different log settings (which are different manufacturer to manufacturer) or sensitivity or the sensor (for example, RED MX is low; it's ISO800 meters closer to ISO400 compared to, say, Alexa, which appears as a full stop brighter at ISO800 even though it's a smaller sensor [though it's photosites are substantially larger).

If your in CLOG and he was in SLOG3, his would look way brighter at the same ISOs.

 

[Tom Gleeson]

he replied, no, even with the same size photosites the large sensor doesn't need as much light, it's just physics." I puzzled and decided to investigate

Unsure what he is talking about here. The only advantage of the larger sensor is the shrinking of the noise when downscaled. If anything a larger sensor needs more light as its larger area requires a larger total numbers of photons to maintain the same number of photons per square unit area. It's way easier to get a lens to concentrate light onto say a tiny B4 size sensor than it is a Full Frame sensor. The Sony A7 has superior light sensitivity due to its sensor design and not its size. The difference in the perceived exposure between the Sony and Canon is more likely due to manufacturers rubbery interpretation of ASA, sensor design and different approaches to signal processing.

As a side note I have seen people shoot camera tests with the Red camera testing sensitivity and then they shoot a cropped image and not the full sensor. This magnifies the noise therefore dynamic range is shrunk and sensitivity decreases as the noise is lifted. The VV Dragon sensor should see a slight improvement in noise when the image is downscaled which means a small improvement in both dynamic range and sensitivity when compared to a 6K Dragon image. But theres no free lunch as larger sensors need lenses that can cover a larger area with the same light intensity to see this advantage.

 

[Kyle Jones]

Probably the different log settings (which are different manufacturer to manufacturer) or sensitivity or the sensor (for example, RED MX is low; it's ISO800 meters closer to ISO400 compared to, say, Alexa, which appears as a full stop brighter at ISO800 even though it's a smaller sensor [though it's photosites are substantially larger).

If your in CLOG and he was in SLOG3, his would look way brighter at the same ISOs.

Well I guess that was it then! Thanks a ton everyone!