Sensor size, crop factor effect aperture as well?

[can-photo]

Advertisement (gone after registration)

I seem to find contradicting opinions on the internet whether smaller sensor affects the aperture or not

 

I don't think it will as it is the same amount of light entering to the camera

 

What do you think? If it does is it same ratio and why?

[bencoyote]

I can understand the confusion. I played with the math regarding this when I first started.

 

lets take the T

23mm f/2 the pupil of that would be: 11.5mm in diameter 

35mm f/2 would be 17.5mm so obviously more light is getting through. Well it is a bit more complicated than that.

 

The 11.5mm pupil is illuminating the 18x24mm or 432mm^2 sensor while the 17.5mm sensor is illuminating a 24x36mm sensor which is 864mm^2. That is about where other evaluations seem to stop. 

 

11.5mm creates a circle of illumination but only a 2:3 portion of that is being used. 6.38x9.57mm or 61.06mm^2

17.5mm 9.70x14.55mm^2 or 141.13mm^2

 

The ratio of the pupil to the area illuminated is nearly the same.

61.06/432=0.1413

141.13/864=0.1633

 

then if you take the math one step further and compute the size of the pixels on the 16.1MP T vs the pixel size on the 23.7MP M then you find that that Leica does some pretty careful math.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[mjh]

With the f-stop number being defined as the focal length divided by the entrance pupil, the sensor size has no bearing on the aperture. Having said that, the sensor size does affect depth of field. A 50 mm Summilux-M used on the T creates an image resembling that taken with a 75 mm f/2.1 lens on the M.

[bencoyote]

With the f-stop number being defined as the focal length divided by the entrance pupil, the sensor size has no bearing on the aperture. Having said that, the sensor size does affect depth of field. A 50 mm Summilux-M used on the T creates an image resembling that taken with a 75 mm f/2.1 lens on the M.

 

Yeah that is true but I don't think that is what the OP was asking about. 

 

He was asking about "amount of light" and the "ratios". Yes of course a 23mm f/2 lens designed for APS-C will have a smaller aperture pupil than a 35mm f/2 lens designed for a full frame camera and of course the amount of light coming into the camera through that larger hole will be greater but it will have to illuminate a larger area and because of that, the ratios of the light striking the sensor will be nearly the same.

[pgk]

I seem to find contradicting opinions on the internet whether smaller sensor affects the aperture or not

 

I don't think it will as it is the same amount of light entering to the camera

 

What do you think? If it does is it same ratio and why?

 

Think of the larger sensor and a crop. Cropping the image has no effect on the amount of light striking the sensor. Different lenses with different covering power may have marginal differences, but in essence all you are effectively doing by using a smaller sensor is cropping the image which would have been produced on a larger sensor. So no effect on aperture.

[pico]

I think bencoyote is making this more difficult than necessary.

 

If I take this 240mm F/5.6 lens from the bench beside me the light admitted at any aperture is the same whether used for 4x5", 8x10" or 56mm x 56mm or 24x36mm - or any format. That's where the definition of aperture works: focal length / diameter (or as MJH correctly defined, entrance pupil). PGK makes the same point.

 

Coverage is a different issue - if a lens won't cover the format, we won't use it will we?  If we rack the lens out for close-ups, it becomes a different (longer) focal length and the F-stop has to be recalculated; in real practice we add to the aperture (smaller diameter).

 

Or go crazy and check out this 136 year-old 17.5" (444.5mm) F/5 lens. Boggling, eh?

[mjh]

Advertisement (gone after registration)

Yes of course a 23mm f/2 lens designed for APS-C will have a smaller aperture pupil than a 35mm f/2 lens designed for a full frame camera and of course the amount of light coming into the camera through that larger hole will be greater but it will have to illuminate a larger area and because of that, the ratios of the light striking the sensor will be nearly the same.

That would depend on what you mean with “the light striking the sensor”. The total amount of light striking the sensor is smaller (because the entrance pupil is smaller) while the light striking any single point on the sensor is the same (because the f-stop is the same). That is you can use the same shutter speed with both cameras but as the total amount of light reaching the sensor different, the T is at a disadvantage with regard to SNR and dynamic range.

[can-photo]

Thanks all for the replies!

 

Thanks for the math bencoyote, I understand how it works now

I can understand the confusion. I played with the math regarding this when I first started.

 

lets take the T

23mm f/2 the pupil of that would be: 11.5mm in diameter 

35mm f/2 would be 17.5mm so obviously more light is getting through. Well it is a bit more complicated than that.

 

The 11.5mm pupil is illuminating the 18x24mm or 432mm^2 sensor while the 17.5mm sensor is illuminating a 24x36mm sensor which is 864mm^2. That is about where other evaluations seem to stop. 

 

11.5mm creates a circle of illumination but only a 2:3 portion of that is being used. 6.38x9.57mm or 61.06mm^2

17.5mm 9.70x14.55mm^2 or 141.13mm^2

 

The ratio of the pupil to the area illuminated is nearly the same.

61.06/432=0.1413

141.13/864=0.1633

 

then if you take the math one step further and compute the size of the pixels on the 16.1MP T vs the pixel size on the 23.7MP M then you find that that Leica does some pretty careful math.

[pico]

Thanks all for the replies!

 

Thanks for the math bencoyote, I understand how it works now

 

Sure you do.

Not by his explanation.

[jank]

I seem to find contradicting opinions on the internet whether smaller sensor affects the aperture or not

 

I don't think it will as it is the same amount of light entering to the camera

 

What do you think? If it does is it same ratio and why?

 

As long as we understand the aperture being just the ratio of a focal length to its physical size, the smaller sensor size only means cropping the picture , understanding that as cutting it e.g. by scissors to smaller size.

One can easily see that the exposure remain same.

We only cheat ourselves of the cut-off part of the image.

However, the perceived depth of field changes. And that seems to be  the confusing part.

Jan

[adan]

The basic equation of aperture size vs. aperture number (f/whatever) starts working before the light even enters the lens, because of the field of view.

 

In other words, (and for the moment we will stick with a constant crop factor) - point your 50mm f/2 lens at a subject area of 4 feet by 6 feet. The total light reaching your lens will be reflected from 24 square feet (what the lens can include in the picture). Of which your f/2 lens can only make partial use - that proportion that will fit through the f/2 aperture of 25mm.

 

Now - switch to a 25mm f/2 lens. Your subject area is now, from the same distance, 8 feet by 12 feet. Four times the area. Four times the light. Which goes through a hole 1/2 the diameter (f/2 = 25/2 = 12.5mm) and 1/4 the area. 4 x 1/4 = 1 - so you get exactly the same volume of light. 1 = 1.

 

Crop your sensor, and you get (for APS) 1/2 the sensor area that needs light, and 1/2 the subject area (field of view) to reflect light. So again, 1/2 = 1/2. Same number of photons per mm² - same exposure.

 

(Before you get excited - remember that a 1.5 crop is linear lengths - the AREA is divided by 1.5 x 1.5, or 2.25, so a bit less than half the area.)

 

This is WHY we use those screwy f/numbers instead of raw diameters in mm or inches. An f/1.4 lens captures and passes the same amount of light, whether it is a 21mm f/1.4 with an opening of 15mm, taking in half the sky - or a 75mm f/1.4 with an opening of 53mm, taking in (light from) only a small part of the sky.

[adamdewilde]

To answer your question bluntly. Nobody on the internet that I've seen/read has it quite right since they miss a few key points.

 

Your topic and your question are two different things BTW. There are to many variables for me to tell you yes or no unless you were to rephrase your question. I can give you some basic truths below.

 

All camera manufactures lie about ISO. It's no longer standard. The only real value in ISOs is now to compare ones noise against another manufactures noise given the same T/stop lens and same shutter speed. I've done it with a number of cameras.. It's a silly boring thing to do and requires studio lighting. I would rather not expose my findings as it would start flame ware and generally things aren't to far off, nothing more then +/- 2 stops from the cameras I've tested.

 

Aperture values are calculated based lens design and not amount of gathered light. T/stops is what you need to determine a lenses true light gathering value. You would be very surprised how wildly varied T/stops and F/stops can be.

 

Now that we've got that sorted I can get into slightly more detail about what I think your question is.

 

Say I take a FF lens like a 50/2 it's T/stop is maybe exactly T/2. That lens on a FF sensor will let in the same amount of light as that lens on a APS-C sensor. Assuming you're shooting a evenly light scene. (I'll explain this more down below)

 

Say you take the same lens and this time you shoot a person. The person will have the same bokeh behind them no matter the camera (assuming the sensors don't render wildly different). The issue will be that on the APS-C sensor the person is cropped (obviously). But why is this important. Because in order to have the persons full face in the camera, you would then have to step back quite a bit. Thus you are changing the point of focus. Thus you are changing the rendering and bokeh of the lens. 

 

That's it. That's the only difference. (back to above) There is nothing causing more or less light to hit the sensor. Any of that WASTED light from the crop is not wasted because it doesn't mater. It's not the total sum of the light hitting the sensor. It's the sum of the light hitting each pixel. Thus light "absorbing" is in question. Not light transmission. And that would then ultimately depend on the sensor. 

 

In general do APS-C sensors have inferior light gathering capabilities? Maybe. But that has nothing to do with the lenses. And their crops.

 

So a rule of thumb would naturally be to multiply the crop factor by the F/stop. That may give you an idea of the bokeh. Though frankly the way lenses are designed it's not even a fair or accurate assessment of bokeh anymore. Unless you're comparing a FF lens on a FF and crop body. And remember it doesn't really matter much for low light abilities because your APS-C lens could have a better transmission value then your FF lens.

[bencoyote]

I think that the big misunderstanding here is a perspective. Let's not consider it from the perspective of constant lens and different sensors but rather from equivilent field of view with different lenses and different sized sensors. Personally I hate the idea of crop factor which is 35mm centric and increasingly irrelevant as we have an ever increasingly diverse range of sensor sizes from cell phones up to various medium format sizes.

 

the point that I was trying to make and my understanding of what the OP was asking was for a roughly equivalent field of view is the light gathering capability of f/2 the same between a 23mm on an APS-C sized sensor and a 35mm lens illuminating 35mm sensor. The answer is the ratio of the reflected light gathered to the area of sensor are illuminated is about the same. The variation between f-stop and T-stop was not considered in my quick and dirty analysis.

 

For the purposes DOF a 23mm is a 23mm and a 35mm is a 35mm and crop factor really doesn't matter except in as much as people using smaller sensor cameras tend to think of a 23mm as a 35mm because they have approximately the same field of view and 35mm as a 50mm for the same reason. I did not think that was the essence of what the original poster was asking.

 

As for the variation in ISO all camera makers do seem to lie. Some more so than others. I did a little testing of an Olympus E-M1 and the Leica T at a maker fair and the Oly departed considerably while the Leica seemed to be close to specification. It seems to be related to the fact that Oly was trying to compete with DLRs on specs while the Germans tend to believe in more objective meausured truth even if their spec sheets don't look quite so impressive. This was not an exhaustive analysis, just a quick demonstration test in a booth by someone who had made equipment to demonstrate this kind of thing. Think science fair project but by an adult.

 

Lastly a big factor in the noise vs sensor size question has a lot to do not with sensor size per se but with pixel size on the sensor. It is not exactly a linear ratio because of certain things having to do with semiconductors but in general a bigger pixel size is able to gather more photons. The fact that the T has a 16MP sensor while the M has a 24MP sensor partly makes up for the smaller sensor size. In essence they traded resolving power for light gathering capability. If I remember correctly it is 4 microns on the T vs 6 on the M. Assuming the same technological generation of sensor, If the T had used 6 micron pixels it would have about the same dynamic range and ISO range as the M but would only be about a 10MP camera.

[adan]

For the purposes DOF a 23mm is a 23mm and a 35mm is a 35mm and crop factor really doesn't matter...

 

Almost. What the lens projects is the same regardless of format, and "a 23mm is a 23mm."

 

But once you factor in enlarging the image (whether a print, or to a given dimension for the web), then the DoF (which is dependent on enlargement) will change slightly.

 

Take a picture on FF with a 23mm lens @ f/2. Blow its 1" by 1.5" original size up 10x (10 x15" print or web image) and there will be slightly blurred points that don't show as visible blur.

 

Take a picture with the same lens and same distance on the crop sensor, and you have to enlarge it 15x to get the same final image area. Some "barely sharp" areas in the previous final image, will now be visibly blurry.

 

I.E. A 23mm @ f/2 on cropped format will not have the same DoF as a 35 @ f/2 on FF - but it also won't have quite the same DoF as a 23mm @ f/2 on FF, in the same final image size.

 

If you made "contact prints" of the images - at the original sensor size (1 x 1.5" and .67 x 1") - you'd see identical blur, with different cropping. If you enlarged the FF image to 10 x 15, and the cropped image to 6.77 x 10 (same 10x enlargement for both) - you'd see identical blur. Truly just like cropping in the enlarger.

 

There's math to figure just how different the final image DoF would be from two formats and the same lens and the same final picture size - like, the aperture times the crop. So a 23mm @ f/2 with a 1.5 crop, would have the DoF (final image size) of a 35mm @ f/3 on FF.

 

A 23 using f/8 on APS sensor, will have the DoF of a 35 @ f/12 (FF), assuming your framed prints are the same size.

[adamdewilde]

I think that the big misunderstanding here is a perspective. Let's not consider it from the perspective of constant lens and different sensors but rather from equivilent field of view with different lenses and different sized sensors. Personally I hate the idea of crop factor which is 35mm centric and increasingly irrelevant as we have an ever increasingly diverse range of sensor sizes from cell phones up to various medium format sizes.

 

the point that I was trying to make and my understanding of what the OP was asking was for a roughly equivalent field of view is the light gathering capability of f/2 the same between a 23mm on an APS-C sized sensor and a 35mm lens illuminating 35mm sensor. The answer is the ratio of the reflected light gathered to the area of sensor are illuminated is about the same. The variation between f-stop and T-stop was not considered in my quick and dirty analysis.

 

For the purposes DOF a 23mm is a 23mm and a 35mm is a 35mm and crop factor really doesn't matter except in as much as people using smaller sensor cameras tend to think of a 23mm as a 35mm because they have approximately the same field of view and 35mm as a 50mm for the same reason. I did not think that was the essence of what the original poster was asking.

 

As for the variation in ISO all camera makers do seem to lie. Some more so than others. I did a little testing of an Olympus E-M1 and the Leica T at a maker fair and the Oly departed considerably while the Leica seemed to be close to specification. It seems to be related to the fact that Oly was trying to compete with DLRs on specs while the Germans tend to believe in more objective meausured truth even if their spec sheets don't look quite so impressive. This was not an exhaustive analysis, just a quick demonstration test in a booth by someone who had made equipment to demonstrate this kind of thing. Think science fair project but by an adult.

 

Lastly a big factor in the noise vs sensor size question has a lot to do not with sensor size per se but with pixel size on the sensor. It is not exactly a linear ratio because of certain things having to do with semiconductors but in general a bigger pixel size is able to gather more photons. The fact that the T has a 16MP sensor while the M has a 24MP sensor partly makes up for the smaller sensor size. In essence they traded resolving power for light gathering capability. If I remember correctly it is 4 microns on the T vs 6 on the M. Assuming the same technological generation of sensor, If the T had used 6 micron pixels it would have about the same dynamic range and ISO range as the M but would only be about a 10MP camera.

 

 

 

I agree with most of this

I'm not exactly sure what the OP was trying to figure out to be honest. I just thought I'd bring up some points people rarely discuss that do impact results from apples to oranges comparisons.  

[can-photo]

What I was trying to figure out whether the crop sensor affects the aperture or not as the title suggests.

 

think of it like I am using a light meter , and reading the values to compare my or camera's exposure, and whether I need to correct or not, whether that depends on the ff lens, or sensor. I wasn't really trying to get to the dof, or perspective. It is more correct exposure with fast or slow. and understanding technical details of that.

 

Thanks for all the answers.

[adamdewilde]

What I was trying to figure out whether the crop sensor affects the aperture or not as the title suggests.

 

think of it like I am using a light meter , and reading the values to compare my or camera's exposure, and whether I need to correct or not, whether that depends on the ff lens, or sensor. I wasn't really trying to get to the dof, or perspective. It is more correct exposure with fast or slow. and understanding technical details of that.

 

Thanks for all the answers.

 

 

In that case, read my long post once again.

In general there shouldn't be any darkening so to speak if you take a 50/1.4 FF lens and stick it on a crop body.

 

Though if you're comparing a 50/1.4 FF lens to a 32/1.4 APS-C lens. Then it has more to do with transmission of light then anything else a crop would introduce. And those T-stops are NEVER posted by the manufacture and a few sites that include "approximate" T-stops rarely measure them properly (but act as decent guidelines none the less). 

 

Aperture multiplied by crop factor is only for DOF approximation. And as mentioned above, smaller sensors in general may have poorer light gathering capabilities thus the manufacturers may lie about ISO values. Thus results may vary.

 

This is easily proved in the following situation.

 

I can take my Leica SL or even Nikon body, shoot my FF lens, then turn on crop mode and shoot again. Nothing happens to the image as far as going darker or lighter when a crop is introduced. BUT If I move forward and backwards to approximate the same crop. Obviously perspective will change, and that might affect the light coming into the sensor. Though generally speaking it won't. 

 

 

Bottom line. A 56/1.2 lens is faster then an 85/1.4 lens if the T-stops are accurate to the F-stops. Therefore the 56mm 1.2 lens lets in more light. Though again because you are forced to step further back, it creates a different perspective and more DOF, but never mind that since it wasn't your original question.