People are very wrong about the advantage of larger sensor

tko said:

Mark Scott Abeln said:

Double the width of a sensor, and you get four times the total light. It's simple. Same with lenses: double the front pupil diameter of a lens, and also get four times the total light. Neither one of these factors has a greater effect than the other.

Click to expand...

It isn't simple. For the SAME size lens, if you double the sensor width and redesign the lens for the same FOV, you get the same amount of light.

Click to expand...

It seems clear from Mark's context that he was referring to doubling the sensor width without changing the lens design - in other words increasing the field of view while keeping aperture and focal length constant.

Uniform luminance, 4x sensor area, so 4x more light - simple.

tko said:

The total light only goes up if 1.) The FOV goes up (you "see" more light) or 2.) The lens has a larger front diameter.

Click to expand...

Correct.

tko said:

Sensor's don't gather light. Lenses do. You can't talk about total light w/o talking about the lens assumptions made.

Click to expand...

But if you are discussing total light, you do need to take into account the sensor size, because this determines field of view.

Re-phrasing your statement:

Lenses concentrate light from the subject. Luminance in the image plane is proportional to the square of the numerical aperture (0.25 / f-number^2)

Sensors intercept light in the image plane. Total light collected is the product of exposure time, image plane luminance and sensor area.

As you say, at fixed field of view, total light is proportional to the square of entrance pupil diameter.

--
Alan Robinson

alanr0 said:

tko said:

Mark Scott Abeln said:

Double the width of a sensor, and you get four times the total light. It's simple. Same with lenses: double the front pupil diameter of a lens, and also get four times the total light. Neither one of these factors has a greater effect than the other.

Click to expand...

It isn't simple. For the SAME size lens, if you double the sensor width and redesign the lens for the same FOV, you get the same amount of light.

Click to expand...

It seems clear from Mark's context that he was referring to doubling the sensor width without changing the lens design - in other words increasing the field of view while keeping aperture and focal length constant.

Uniform luminance, 4x sensor area, so 4x more light - simple.

Click to expand...

You cannot get 4x light without change lens design! Your original lens can only provide image cover 1x area. if you want to cover 4x area with the same light flux you HAVE TO double lens diameter (or aperture) . This is what make Mark's calculation wrong. IF this is right then it is natural that by doubling sensor size you double the total light gathered. Unfortunately you cannot!

If you keep the same FOV and without redesign lens you have to move sensor to further distance (2x in this case) and result 4 time weaker light. In the end you get the same amount of light.

This is what people get lost in details.

janez said:

I was thinking about space telescopes like James Webb. Curiosity rover has 13.2 x 8.8mm sensor, not that much larger than smartphone.

Click to expand...

Typical smartphone sensor is around 4x5mm, or 20mm^2. The sensor you listed is 116mm^2, or almost 5 times greater area.

yfan said:

You cannot get 4x light without change lens design! Your original lens can only provide image cover 1x area. if you want to cover 4x area with the same light flux you HAVE TO double lens diameter (or aperture) . This is what make Mark's calculation wrong. IF this is right then it is natural that by doubling sensor size you double the total light gathered. Unfortunately you cannot!

If you keep the same FOV and without redesign lens you have to move sensor to further distance (2x in this case) and result 4 time weaker light. In the end you get the same amount of light.

Click to expand...

I think what was assumed but not stated clearly was that we assume the lens is also scaled up proportionately. For example if you have a cellphone camera with f2 lens, and then you have a FF camera with F2 lens, the amount of light goes way up. Put the cellphone lens on the FF sensor? Nobody would do that.

yfan said:

alanr0 said:

tko said:

Mark Scott Abeln said:

Double the width of a sensor, and you get four times the total light. It's simple. Same with lenses: double the front pupil diameter of a lens, and also get four times the total light. Neither one of these factors has a greater effect than the other.

Click to expand...

It isn't simple. For the SAME size lens, if you double the sensor width and redesign the lens for the same FOV, you get the same amount of light.

Click to expand...

It seems clear from Mark's context that he was referring to doubling the sensor width without changing the lens design - in other words increasing the field of view while keeping aperture and focal length constant.

Uniform luminance, 4x sensor area, so 4x more light - simple.

Click to expand...

You cannot get 4x light without change lens design! Your original lens can only provide image cover 1x area. if you want to cover 4x area with the same light flux you HAVE TO double lens diameter (or aperture) . This is what make Mark's calculation wrong. IF this is right then it is natural that by doubling sensor size you double the total light gathered. Unfortunately you cannot!

If you keep the same FOV and without redesign lens you have to move sensor to further distance (2x in this case) and result 4 time weaker light. In the end you get the same amount of light.

This is what people get lost in details.

Click to expand...

It depends on your starting point.

There are numerous adaptors allowing FF lenses to be used with micro 4/3 bodies. You can use the same 50 mm f/1.4 lens to cover 13x17.4 mm m4/3 or 24x36 mm FF. No lens redesign required, no change of image distance and magnification, but roughly twice the field of view and 4x the sensor area with the 24x36 mm sensor.

Similarly there are many C-mount lenses which can be used with 2/3 inch, 1/2 inch or 1/3 inch cameras.

For a less dramatic change in sensor size, Nikon, Canon and Pentax and Sony FF lenses can all be mounted on the same manufacturer's corresponding APS-C bodies without adapters.

--
Alan Robinson

Camley said:

yfan said:

You often hear people say that the low light performance is better with larger sensor. Manufactures and retailers all advertise saying something like “The low light performance is great due to camera's large sensor.” etc.

Click to expand...

yfan - They say that because larger sensors have better noise performance than smaller sensors due the increased amount of light collected.

Camley said:

From physics point view, sensor size actually is irrelevant in low light performance. In the low light situation, the performance is dominated by lens which is independent of sensor. The bigger the lens, the more light it gets, the better IQ.

Click to expand...

There are hundreds of posts explaining the reason for the noise advantage of large sensors. Great Bustard can tell you why you are wrong.

Click to expand...

While that is true, there's a reason that there is no such thing as a M4/3 f64 Group...

yfan said:

I do agree with you that in my original post I did not make all conditions clear and omitted many details and some may even inaccurate. I try not go into details, which I think is the reason many people have been overwhelmed and got lost. It is often discussions are drifted into secondary factors, like the other sources of noise in low light situation.

Click to expand...

I just discovered your original posting and found it "interesting" when you say "People are very wrong about the advantage of larger sensor"

Maybe with your phd in physics, you could give a reasonably complete and clear set of conditions by which you make your deductions . I personally found your various notes in this thread awkward and disjointed ... therefore confusing. In addition, some comments appear to be at odds with my non-academic understanding of theory and my practice with photography.

yfan said:

I do find few discussions are edifying and beneficial to all. Unfortunately many end up emotional and personal.

--
yfan
---
http://www.flickr.com/photos/35513150@N00
http://flickrock.com/35513150@N00

Click to expand...

--
Charles Darwin: "ignorance more frequently begets confidence than does knowledge."
tony
http://www.tphoto.ca

Seems like larger image sensors have no benefits ... if so I assume smaller image sensors have no drawbacks (or is this a one directional issue?)...

Case 1:

So I am thinking about getting me a 1" image sensor camera (as I am very wrong about the advantage of a larger sensor). Take a lot of images in low light.

So to replace the 1.4/50mm lens I need a 18.5 mm f/0,52 lens (want the same physical aperture, hence the same light grasp to take full advantage of my new image sensor). What matter to me is physical aperture to keep the flux similar).

Now - where do I get my new lens?

Case 2:

With my telescope (have one as we do not change telescopes like we change camera lenses). This telescope has fixed aperture (lens diameter) and focal lenght. When using my tiny image sensor the field of view is tiny. So getting a bigger image sensor offer a huge advantage in this real life setting - a larger field of view (despite me still not grasping the advantage of a larger sensor).

Have not got the means to get a new telescope for every image sensor - am I stuck?

Please: do NOT take this post dead seriously, just taking a look as some aspects of not knowing the advantage of different image sensor sizes... ;-)

PhotoFactor said:

janez said:

I was thinking about space telescopes like James Webb. Curiosity rover has 13.2 x 8.8mm sensor, not that much larger than smartphone.

Click to expand...

Typical smartphone sensor is around 4x5mm, or 20mm^2. The sensor you listed is 116mm^2, or almost 5 times greater area.

Click to expand...

Yes, but that is still not much larger than smartphone. I guess size is optimal regarding to IQ/mass for its misson. They could use smartphone size sensor and get similar IQ but probably with greater mass of lens.

janez said:

I was thinking about space telescopes like James Webb. Curiosity rover has 13.2 x 8.8mm sensor, not that much larger than smartphone.

Click to expand...

Random examples:

Hubble wide field camera: WFC3 8 Mp 61 x 31 mm e2v CCD

James Webb Space Telescope Near IR camera 4 Mp 37 x 37 mm HgCdTe detector

Ground-based Pan-STARRS : 1400 Mp 400 mm x 400 mm array of CCDs

--
Alan Robinson

n/t n/t

tko said:

tko said:

There are hundreds of posts explaining the reason for the noise advantage of large sensors. Great Bustard can tell you why you are wrong

Click to expand...

He won't and couldn't, because it's not wrong because a sensor can't be analyzed w/o a lens attached. That is exactly what equivalence is all about. Equivalent lenses give equivalent performance regardless of the sensor size. Please think about it.

Click to expand...

The way to compare sensor sizes alone, with size as the only variable, is to use a camera with FX and DX settings, without changing the lens, the f number or the shutter time.

Alternatively, you can just crop an image file from any sensor.

Starting from this comparison pair, you can then introduce other variables one by one. For instance, print size, lens aperture. lens focal length, or shutter time.

Introducing pixel pitch into the comparison is difficult because two sensors with the same height and width but different pitch will probably have been designed in different years, with various other changes beside pixel size. The lens and the exposure settings should be kept constant for the comparison. One could reasonably compare the Sony A7 and A7r sensors, for instance, using the same lens.

PhotoFactor said:

yfan said:

You cannot get 4x light without change lens design! Your original lens can only provide image cover 1x area. if you want to cover 4x area with the same light flux you HAVE TO double lens diameter (or aperture) . This is what make Mark's calculation wrong. IF this is right then it is natural that by doubling sensor size you double the total light gathered. Unfortunately you cannot!

If you keep the same FOV and without redesign lens you have to move sensor to further distance (2x in this case) and result 4 time weaker light. In the end you get the same amount of light.

Click to expand...

I think what was assumed but not stated clearly was that we assume the lens is also scaled up proportionately. For example if you have a cellphone camera with f2 lens, and then you have a FF camera with F2 lens, the amount of light goes way up.

Click to expand...

The number of photons goes way up, but they are spread over a much larger area. So the illuminance (photons per square mm) stays the same.

PhotoFactor said:

Put the cellphone lens on the FF sensor? Nobody would do that.

Click to expand...

It's very disappointing to see manufacturers and (some) review sites, magazines or so-called experts repeating the same popular but flawed (not to say wrong) arguments year after year after year after year after year after year after year...

Here's a thread I started on the topic back in 2007. Things have changed since... but not enough.

https://www.dpreview.com/forums/post/23033618

pidera said:

It's very disappointing to see manufacturers and (some) review sites, magazines or so-called experts repeating the same popular but flawed (not to say wrong) arguments year after year after year after year after year after year after year...

Here's a thread I started on the topic back in 2007. Things have changed since... but not enough.

https://www.dpreview.com/forums/post/23033618

Click to expand...

10 years later, and you still do not know better.

PhotoFactor said:

I think what was assumed but not stated clearly was that we assume the lens is also scaled up proportionately. For example if you have a cellphone camera with f2 lens, and then you have a FF camera with F2 lens, the amount of light goes way up. Put the cellphone lens on the FF sensor? Nobody would do that.

Click to expand...

Sort of. If you reverse the lens and put it on an extension tube you have a microscope?


-- Bob
http://bob-o-rama.smugmug.com -- Photos
http://www.vimeo.com/boborama/videos -- Videos
http://blog.trafficshaper.com -- Blog

My point is to agree with the idea that the advantage of FF sensors (as compared to smaller sensors) is in low iso rather than in high iso. Besides that I do like the fact that FF lenses have a wider range of (large) apertures available (DOF control and light gathering). So yes I know why I use FF when I use it and why not when I don't. What exactly is your point ?

pidera said:

My point is to agree with the idea that the advantage of FF sensors (as compared to smaller sensors) is in low iso rather than in high iso.

Click to expand...

Start with this: ISO is very much irrelevant to the discussion.

From the OP :

The actual advantages of larger sensor lies in two areas:

1. Resolution - the large sensor can simply fit more pixels.

2. Base ISO performance.

pidera said:

From the OP :

The actual advantages of larger sensor lies in two areas:

1. Resolution - the large sensor can simply fit more pixels.

2. Base ISO performance.

Click to expand...

Very limited view, as explained elsewhere in this thread, missing interdisciplinary connections.