How does crop mode or cropping in post impact subject isolation/DOF?

[bleirer]

So I think I know the answer to this question, or at least have an opinion, but I find it interesting. If you shoot in crop mode, or even crop in post, for depth of field purposes do you have to divide your f number by the crop factor?

 

[DRwyoming]

So I think I know the answer to this question, or at least have an opinion, but I find it interesting. If you shoot in crop mode, or even crop in post, for depth of field purposes do you have to multiply your f number by the crop factor?

If you maintain the same camera distance to your subject for the same focal length lens at the same actual aperture, then yes the effective aperture scales with the crop factor from a DoF perspective but most of the time when framing a scene with a full frame vs crop camera (including as you say cropping in post) you either change position relative to the subject to frame the scene appropriately or change lens focal length to frame the scene appropriately.

IOW, it gets tricky in the real world as full frame vs crop aren't the only variables. If you fix all but one, say shoot from the same subject distance with one focal length and aperture and only vary the crop factor you can come to one set of conclusions but if you take into account that to frame a given scene or achieve a certain main subject size in the frame with the full frame vs crop camera then focal length or working distance or both typically change and you can reach different conclusions as to the impact of cropping.

So in a general way, if you stay in the same position and shoot with the same lens at the same aperture and switch to a crop (crop camera, in-camera crop or cropping in post) you'll frame a different scene and have a different main subject size in the frame with the cropping but DoF will decrease with cropping. But if you stay in the same position and change focal length to frame the same scene and maintain the same main subject size in the fame and keep the aperture the same the DoF will increase with the crop. Then add the variable of changing shooting position (working distance to main subject) and it gets even harder to figure out the DoF impact of cropping and of course as per your previous thread the perspective will change.

If you play with these things in an online DoF calculator like this it's easy to see there's no simple answer as it depends a lot on how a given photographer will actually work and what they're trying to achieve: https://www.photopills.com/calculators/dof

 

[bleirer]

If you maintain the same camera distance to your subject for the same focal length lens at the same actual aperture, then yes the effective aperture scales with the crop factor from a DoF perspective but most of the time when framing a scene with a full frame vs crop camera (including as you say cropping in post) you either change position relative to the subject to frame the scene appropriately or change lens focal length to frame the scene appropriately.

IOW, it gets tricky in the real world as full frame vs crop aren't the only variables. If you fix all but one, say shoot from the same subject distance with one focal length and aperture and only vary the crop factor you can come to one set of conclusions but if you take into account that to frame a given scene or achieve a certain main subject size in the frame with the full frame vs crop camera then focal length or working distance or both typically change and you can reach different conclusions as to the impact of cropping.

So in a general way, if you stay in the same position and shoot with the same lens at the same aperture and switch to a crop (crop camera, in-camera crop or cropping in post) you'll frame a different scene and have a different main subject size in the frame with the cropping but DoF will decrease with cropping. But if you stay in the same position and change focal length to frame the same scene and maintain the same main subject size in the fame and keep the aperture the same the DoF will increase with the crop. Then add the variable of changing shooting position (working distance to main subject) and it gets even harder to figure out the DoF impact of cropping and of course as per your previous thread the perspective will change.

If you play with these things in an online DoF calculator like this it's easy to see there's no simple answer as it depends a lot on how a given photographer will actually work and what they're trying to achieve: https://www.photopills.com/calculators/dof

Thank you, nice answer and food for thought.

One scenario one sees often is the idea of shooting macro at a greater distance with the hopes of avoiding a narrow dof, and instead cropping later. I'm pretty sure you covered that situation and that turns out to be a 'no free lunch' kind of thing, since the cropping would take back the dof that shooting from farther away initially had.

 

[Strodav]

Your question is confusing to me, but I believe this is what you may be looking for. IF you want to get the same results with a crop sensor camera as with a full frame camera then the following 3 formulas apply:

Cropped focal length = FF focal length / crop factor
Cropped aperture = FF aperture / crop factor
Cropped sensor ISO = FF ISO / crop factor ^2

Again, if you have a FF camera and crop sensor camera next to each other shooting the same subject and you wan to get the exact same look / result from both cameras, then the formulas apply.

 

[bleirer]

Your question is confusing to me, but I believe this is what you may be looking for. IF you want to get the same results with a crop sensor camera as with a full frame camera then the following 3 formulas apply:

Cropped focal length = FF focal length / crop factor
Cropped aperture = FF aperture / crop factor
Cropped sensor ISO = FF ISO / crop factor ^2

Again, if you have a FF camera and crop sensor camera next to each other shooting the same subject and you wan to get the exact same look / result from both cameras, then the formulas apply.

That is good info. Actually what got me thinking about it was the discussion about standing in one place with the same focal length but switching to crop mode or cropping in post. I believe in that case one loses depth of field equivalent to dividing by the crop factor.

 

[Strodav]

That is good info. Actually what got me thinking about it was the discussion about standing in one place with the same focal length but switching to crop mode or cropping in post. I believe in that case one loses depth of field equivalent to dividing by the crop factor.

So, two cameras side by side one FF one DX each with the exact same, say, 85mm f/1.8 portrait lens at 10 feet from a gorgeous model in a beautiful setting. The first thing you will notice is that your subject fills more of the frame with the DX camera, so if you want to frame the same shot, you will have to move the DX camera further from your subject. The next thing you will notice is the DOF is different. Say your have chosen f/5.6 on both cameras. Using the Cambridgeincolour.com DOF calculator the FX camera will have a DOF of about 1.5 ft, the DX camera will have a DOF of about 1 ft. Say, your ISO is at 800 to get a good exposure. After you take the shot and pixel peep on your monitor, you will notice a bit more noise in the shadows of the DX camera because the pixel density of the DX sensor is higher than the FX sensor so it captures less light resulting in more noise. If you wanted to get the same SNR you would need to use 800 / 1.5^2 = 355 (so use ISO 320) on the DX camera to get the same noise level as on the FX camera.

If you just take the shot with the FX camera and crop by 1.5x in post, the subject will fill more of the frame, the DOF will not be any different because is was set when you clicked the shutter button, but the apparent noise increases with the size of the print because you are magnifying the noise when making a larger prints.

 

[bleirer]

So, two cameras side by side one FF one DX each with the exact same, say, 85mm f/1.8 portrait lens at 10 feet from a gorgeous model in a beautiful setting. The first thing you will notice is that your subject fills more of the frame with the DX camera, so if you want to frame the same shot, you will have to move the DX camera further from your subject. The next thing you will notice is the DOF is different. Say your have chosen f/5.6 on both cameras. Using the Cambridgeincolour.com DOF calculator the FX camera will have a DOF of about 1.5 ft, the DX camera will have a DOF of about 1 ft. Say, your ISO is at 800 to get a good exposure. After you take the shot and pixel peep on your monitor, you will notice a bit more noise in the shadows of the DX camera because the pixel density of the DX sensor is higher than the FX sensor so it captures less light resulting in more noise. If you wanted to get the same SNR you would need to use 800 / 1.5^2 = 355 (so use ISO 320) on the DX camera to get the same noise level as on the FX camera.

If you just take the shot with the FX camera and crop by 1.5x in post, the subject will fill more of the frame, the DOF will not be any different because is was set when you clicked the shutter button, but the apparent noise increases with the size of the print because you are magnifying the noise when making a larger prints.

I believe if I took the shot with FX then cropped 1.5 in post, the dof would stay the same if I left the cropped image small, but upon viewing both side by side at the same size I have magnified the blur of the cropped image. I've changed the circle of confusion and now have less DOF.

 

[Strodav]

I believe if I took the shot with FX then cropped 1.5 in post, the dof would stay the same if I left the cropped image small, but upon viewing both side by side at the same size I have magnified the blur of the cropped image. I've changed the circle of confusion and now have less DOF.

You cannot change the circle of confusion or DOF in Post Processing as it is set by the lens, aperture, and camera (pxiel density) at the time you take the shot. The circle of confusion comes from the cone of light rays not coming into perfect focus at the sensor, not anything to do with looking at the image file on a monitor or printing it. All you can do in post is magnify the softness in the image, perhaps due to the circle of confusion, or you can minimize it by shrinking the image. If you print it, the closer the viewing distance the more resolution you can see magnifying the softness in the image. The further the viewing distance the less resolution the eye can see. Note that detail / resolution is different than sharpness. Sharpness can be improved in post, detail / resolution is set at the time you press the shutter button. Then there's the whole airy disk diffraction thing.

 

[Maljo]

Same camera, same lens, same aperture, same subject distance -> same depth of field.
You are just cropping the center of the image.

 

[Tom Reynolds]

Ok, I have:
1) D-500/533.34fx lens 20.8 mp (1.5crop factor)
2) Canon R7/500 lens 32.4mp (1.6 crop factor)
3) Oly OM-1/400mm lens 20MP (2.0 crop factor)

Per equation 1, all three are an 800mm FX equivalent.

What you are saying is that:
1) the highest pixel density will have the lowest detail?
2) the larger crop factor will have more or less depth of field?
3) the larger crop factor will need larger or smaller aperture or higher/lower ISO for the same brightness.

What I am trying to determine is the penalty the 4/3 sensor has in comparison with the Canon crop sensor in relation to the D-500.

Tom

 

[bleirer]

I saw this article which seems to support of the idea that cropping in post alters the dof. The circle of confusion is dependent on the viewing size and viewing distance as well.

Sensor Size, Perspective and Depth of Field

photographylife.com

 

[DRwyoming]

Same camera, same lens, same aperture, same subject distance -> same depth of field.
You are just cropping the center of the image.

If you just cropped and then had a smaller central portion of the larger image then that makes sense but DoF historically is compared at the same output image size. When you resize the cropped central portion of the image to the dimensions of the uncropped image you enlarge the slightly out of focus circles of confusion and the DoF changes relative to the uncropped image.

There's roughly 100 years of good research describing DoF and it was known even way back then that if you change media size (e.g. go from 120 roll film to 8"x10" sheet film) and otherwise frame the same scene the DoF and or aperture required to achieve a given DoF changes with the media size. That's no different than today's discussions of DoF vs sensor size or cropping in post. But one important key to understanding this is to evaluate DoF at the same output file size and not just view the output from a smaller media size camera as a smaller image.

As posted previously if you want to explore how DoF changes with cropping and different combinations (e.g. same field of view with a different lens choice, same or differing aperture, same distance to main subject or changing working distance) including cropping you can play with various combinations here: https://www.photopills.com/calculators/dof If you do, you'll see that DoF definitely changes with cropping if you leave everything else the same.

 

[bleirer]

Ok, I have:
1) D-500/533.34fx lens 20.8 mp (1.5crop factor)
2) Canon R7/500 lens 32.4mp (1.6 crop factor)
3) Oly OM-1/400mm lens 20MP (2.0 crop factor)

Per equation 1, all three are an 800mm FX equivalent.

What you are saying is that:
1) the highest pixel density will have the lowest detail?
2) the larger crop factor will have more or less depth of field?
3) the larger crop factor will need larger or smaller aperture or higher/lower ISO for the same brightness.

What I am trying to determine is the penalty the 4/3 sensor has in comparison with the Canon crop sensor in relation to the D-500.

Tom

Neither crop factor nor pixel density affect exposure. The light per unit area of the sensor is the same no matter the size. However larger sensors can have better low light performance and dynamic range. I think this is related to the noise formula given earlier.

The pixel density doesn't impact depth of field as far as I know. It's the subject distance, the equivalent focal length, the aperture, and the crop factor that have an impact.

If you physically move back using the crop camera to get the subject the same size in the frame as the full frame camera then the crop camera will have greater depth of field at the same aperture. This is because of the subject distance changing. If you don't physically move, the crop camera will have less depth of field because to view them side by side the blur of the cropped image is enlarged.

 

[Palmor]

What I am trying to determine is the penalty the 4/3 sensor has in comparison with the Canon crop sensor in relation to the D-500.

Tom

The ”penalty” is that the m43 sensor has less light coming into it so for the same ISO it will be ~2 stops worse ISO performance than FF.

There are ways to help offset that penalty. If you have access to faster glass for M43 than FF so you can lower the ISO or if your are shooting more static subjects the IS of M43 is very good, allowing you to shoot lower shutter speeds than you may with FF.

I don’t worry about the DOF differences as much, but there are online calculators that you can use to determine if the DOF differences will be an issue for you. For some the advantage is on FF but for other it’s M43.

 

[EricBowles]

That is good info. Actually what got me thinking about it was the discussion about standing in one place with the same focal length but switching to crop mode or cropping in post. I believe in that case one loses depth of field equivalent to dividing by the crop factor.

Keep it simple. Consider this using the same camera, same focal length and same distance to the subject. In that scenario, the original image, original image cropped to DX proportions in post, and the same image made with the camera set to DX Crop mode will all produce exactly the same DOF. The cropped in post vs. DX mode will be the same framing as well. The full frame image will have the same DOF, but the field of view would be different to the apparent DOF may seem different but it is the same.

If you zoomed tighter with the full frame image to the point where the framing is the same, your focal length will be 1.5 times the original focal length. In that case the DOF of the full frame image will be shallower, and you'll have much more subject isolation. You will maintain that shallower DOF with full frame for all focal lengths and levels of zooming if the actual focal length is longer than 67% of the DX focal length.

There are a lot of different aspects to this. For example, with good light and reasonable shutter speeds, if you can frame the image with less than a 1.5 crop, full frame will use the longer focal length to fill the frame and have a shallower DOF. Once you get to a 1.5 crop or deeper, they are the same. But it may be a little easier to focus on the subject if the subject is smaller in the frame by switching to DX crop mode because you would get additional viewfinder magnification and the AF boxes would be slightly larger.

I think a lot about that middle ground between FX and a crop to DX. This is a range where FX usually has an advantage in noise and DOF. Of course this can vary with specific camera models, resolution, ISO level, etc. I've also found that high resolution sensors don't have a full advantage of cropping to DX proportions without an adverse impact because those high resolution sensors need the benefit of downsizing to reduce noise. So we can't both use high ISO levels and crop deeply with a high resolution camera unless the final output is small or for web only.

 

[bleirer]

Keep it simple. Consider this using the same camera, same focal length and same distance to the subject. In that scenario, the original image, original image cropped to DX proportions in post, and the same image made with the camera set to DX Crop mode will all produce exactly the same DOF. The cropped in post vs. DX mode will be the same framing as well. The full frame image will have the same DOF, but the field of view would be different to the apparent DOF may seem different but it is the same.

If you zoomed tighter with the full frame image to the point where the framing is the same, your focal length will be 1.5 times the original focal length. In that case the DOF of the full frame image will be shallower, and you'll have much more subject isolation. You will maintain that shallower DOF with full frame for all focal lengths and levels of zooming if the actual focal length is longer than 67% of the DX focal length.

There are a lot of different aspects to this. For example, with good light and reasonable shutter speeds, if you can frame the image with less than a 1.5 crop, full frame will use the longer focal length to fill the frame and have a shallower DOF. Once you get to a 1.5 crop or deeper, they are the same. But it may be a little easier to focus on the subject if the subject is smaller in the frame by switching to DX crop mode because you would get additional viewfinder magnification and the AF boxes would be slightly larger.

I think a lot about that middle ground between FX and a crop to DX. This is a range where FX usually has an advantage in noise and DOF. Of course this can vary with specific camera models, resolution, ISO level, etc. I've also found that high resolution sensors don't have a full advantage of cropping to DX proportions without an adverse impact because those high resolution sensors need the benefit of downsizing to reduce noise. So we can't both use high ISO levels and crop deeply with a high resolution camera unless the final output is small or for web only.

The part I think I disagree with is the part about the same full frame camera cropped to DX or shot in DX mode. Of course it's exactly the same shot, similar to taking an 8x12 print and a pair of scissors and cutting out the 4x6 center. The dof is the same until you either view it closer or enlarge it to be a size equal to the original. At that point the dof of the crop is less. Advanced dof calculators all use the image size and viewing distance to calculate the circle of confusion.

Here is a quote from that article I cited earlier that I think explains it clearly:

"Smaller Sensor = decreased depth of field (if identical focus distance, physical focal length and physical f-number)
When you put photographs from two cameras next to each other to compare them, you are typically looking at these images at the same size. However, the image sensors that generated these two images may be very different in size. For example: the iPhone has a sensor that is less that one seventh the size of a 35mm full-frame DSLR in each of its dimensions. This means that the physical image that was projected onto the image plane of the iPhone was magnified by a factor more than seven times more than the DSLR’s image so that it could be displayed at the same size in the side-by-side comparison in this post.

This magnification magnifies everything – also imperfections and blurring in the projected image. This means that, at the same distance from your subject, at the same physical focal length and aperture setting, a camera with a smaller sensor will have shallower depth of field than the one with a larger sensor. The images will have the same perspective, but different fields of view (framing), so it is a bit of an apples and oranges comparison. However, the result is real, and goes contrary to common knowledge and what one might have expected!"

Source:

Sensor Size, Perspective and Depth of Field

photographylife.com

 

[bleirer]

I also think that same notion applies to noise. A formula for noise was mentioned earlier where the more one crops the more noise is evident. But if one just cropped and kept looking at it small from the same distance it would have to be the same noise because it's the exact image just with part cut off. Like holding your hands out to block the outer part of the image, it doesn't get noisier by doing that. But once it gets resized to be equal to the original is when the noise is seen. That or viewing it at a closer distance, same idea.

Same idea with DOF I think. Crop but keep it small and view at the same distance, no change. But resize and you see the difference.

 

[DRwyoming]

I also think that same notion applies to noise. A formula for noise was mentioned earlier where the more one crops the more noise is evident. But if one just cropped and kept looking at it small from the same distance it would have to be the same noise because it's the exact image just with part cut off. Like holding your hands out to block the outer part of the image, it doesn't get noisier by doing that. But once it gets resized to be equal to the original is when the noise is seen. That or viewing it at a closer distance, same idea.

Same idea with DOF I think. Crop but keep it small and view at the same distance, no change. But resize and you see the difference.

Yup, comparisons of noise or DoF should be made at the same output image size.

 

[NorthernFocus]

I also think that same notion applies to noise. A formula for noise was mentioned earlier where the more one crops the more noise is evident. But if one just cropped and kept looking at it small from the same distance it would have to be the same noise because it's the exact image just with part cut off. Like holding your hands out to block the outer part of the image, it doesn't get noisier by doing that. But once it gets resized to be equal to the original is when the noise is seen. That or viewing it at a closer distance, same idea.

Same idea with DOF I think. Crop but keep it small and view at the same distance, no change. But resize and you see the difference.

Yup, comparisons of noise or DoF should be made at the same output image size.

Exactly so. This topic gets way more complicated than it needs to be due to people studying/discussing the theory rather than comparing final results. In theory for a given lens at a given aperture and range DOF is narrower with a cropped sensor. But in practice it works the opposite way because shooting with cropped sensors/crop mode/cropping in post means increased range for a given FOV. Using a DOF calculator to compare FF vs cropped sensor possibilities for a given lens requires varying the range with all else remaining equal. That's why the classic argument is that FF sensors produce shallower DOF than cropped sensors. For a given (uncropped)FOV they do because the subject is closer or a longer lens is needed. Both of which result in shallower DOF.

 

[Whiskeyman]

Depth of field is basically an optics/lens and distance to subject issue. If all you change is a camera's sensor capture area (full-frame to other than full-frame), the depth-of-field (range of acceptable focus) is the same. (If switching between camera bodies, differences in sensor physics may be an issue.)

The field of view changes.

The amount that you can enlarge either image is the same, if sensor pixel density is constant.

However, the depth of field does not change, nor does the lens focal length or lens apertures.

 

[DRwyoming]

However, the depth of field does not change, nor does the lens focal length or lens apertures.

The DoF certainly does change if you view the resulting images scaled to the same size as you should for a direct comparison. When you enlarge the cropped image to the size of the uncropped image the circles of confusion grow in size and the DoF changes.

Check any DoF calculator and keep everything the same except switch from a full sensor body to a crop sensor body you'll see the near and far focus distances as well as the total DoF change but that change is based on the century old convention to compare DoF at a common output image size. If all you do is crop the center of the image and leave it as a smaller image then of course DoF doesn't change but by convention that's not how DoF for different media sizes is compared and that dates back to the sheet film days but is just as valid when talking about sensor sizes and cropping.

Online Depth of Field Calculator

Depth of field and hyperfocal distance calculator

www.dofmaster.com

Depth of Field (DoF) calculator | PhotoPills

This Depth of Field (DoF) calculator will help you get the creative control on the design of your photos. Decide the zone of sharpness you need to tell the story you want and get your message across.

www.photopills.com

 

[bleirer]

This first one here shows how circle of confusiion varies with image size and viewing distance as well as our visual acuity. The 2nd incorporates all into the calculator. I believe we can safely substitute image size for print size and still be good.

Circle of Confusion (CoC) Calculator | PhotoPills

Would you like to understand Circle of Confusion (CoC)? and how it changes with sensor size, maximum photo print dimension, photo viewing distance and viewer’s visual acuity. This Circle of Confusion calculator will work out the numbers for you.

www.photopills.com

Advanced Depth Of Field (DoF) Calculator | PhotoPills

Get the total control over depth of field. This calculator allows you to use an adjusted Circle of Confusion to calculate the adjusted depth of field values according to your desired print size, viewing distance, camera sensor size and viewer’s visual acuity.

www.photopills.com

 

[bleirer]

I think the normalizing for equal image size is a common thing to get an apples to apples comparison. Photons to photos does it. Steve did it in his article on crop vs full Frame. I think some of dxo's tests do it.

 

[DRwyoming]

I think the normalizing for equal image size is a common thing to get an apples to apples comparison. Photons to photos does it. Steve did it in his article on crop vs full Frame. I think some of dxo's tests do it.

Yeah, it's a convention that's been used for at least a century to make head to head DoF comparisons of images captured on very different media sizes. But yes, more recently it's been adopted as the standard way to compare other things like image noise from different size sensors or different resolution cameras. Things become really hard to compare if you don't normalize output size.

 

[NorthernFocus]

Depth of field is basically an optics/lens and distance to subject issue. If all you change is a camera's sensor capture area (full-frame to other than full-frame), the depth-of-field (range of acceptable focus) is the same. (If switching between camera bodies, differences in sensor physics may be an issue.)

The field of view changes.

The amount that you can enlarge either image is the same, if sensor pixel density is constant.

However, the depth of field does not change, nor does the lens focal length or lens apertures.

All theoretically correct but not meaningful in real world application. For practical photography comparing output images with the same FOV and format is what matters and why it has been the method of comparison for decades.