MACRO question

[digibeet]

Does the crop factor affects the depth of field?

[nosquare2003]

An excellent article to read:

http://bobatkins.com/photography/technical/digitaldof.html

[DaveG]

An excellent article to read:

http://bobatkins.com/photography/technical/digitaldof.html


I think that Mr. Atkins misses something when he says that the same lens on a 35mm vs 10D camera would have a different depth of field.

How does the lens know that its being used on a different camera? It doesn't make sense.

Try a set up like this: Put a 50mm on a 35mm camera and place it two meters away from the subject. Now the distance between the camera/subject must not change. Select any aperture you'd like. The results will be a a loose crop and X amount of depth of field.

Now you place the same lens on a 10D. Nothing else has changed except that I have a tighter crop. Why would the depth of field change? It's the same lens and the lens has only ONE amount of depth of field to give based on a constant aperture/subject distance. How could it be otherwise? If I took the 35mm image and cropped it to the same size as the 10D I would have two identical images, right down to the EXACT same depth of field.

Field of view changes with the format, but depth of field is constant to the lenses' focal length. I've only wished that the normal lens for my 8x10 camera - a 300 mm Nikkor - gave me the same depth of field as a 50mm lens would have on a 35mm camera. Unfortunately it gives me EXACTLY the same depth of field as 300mm lens designed for 35mm, which is probably one of the reasons that it stops down to f128 (diffraction be damned!).

[Cadwell]

Read the section in that article on "Circle of Confusion" - then it might make sense to you...

DoF is dependant on the "Circle of Confusion", the subject distance, the lens focal length and the aperture.

So with constants Subject distance = 20ft, lens = 50mm, fstop = f/2.8

On my 1D Mark II body, DoF = 6.45ft
On my 10 body, DoF = 5.29ft

The difference? A 1.3x sensor on one and a 1.6x sensor on the other....

Calculation made with the Online DoF Calculator (http://www.dofmaster.com/dofjs.html)

[DaveG]

Read the section in that article on "Circle of Confusion" - then it might make sense to you...

Subject distance 20ft, lens 50mm, fstop f/2.8

On my 1D Mark II body, DoF = 6.45ft
On my 10 body, DoF = 5.29ft

The difference? A 1.3x sensor on one and a 1.6x sensor on the other....

Calculation made with the Online DoF Calculator (http://www.dofmaster.com/dofjs.html)


Sorry, if you crop the images the same way the circle of confusion point will be identical too.

[Cadwell]

Sorry, if you crop the images the same way the circle of confusion point will be identical too.

But you're not cropping anything... the term "crop" is only relevant when you are comparing the size of a dSLR sensor with a 35mm frame... it plays no part in a DoF calculation.

[steven]

Circle of confustion is a value associated with the camera?
and has not direct relation to the sensor size?

[Jon]

Circle of Confusion refers to the maximum amount a point can be blurred on the film/sensor and still appear to the viewer as a point on subsequent enlargement. Depth of Field tables assume a value for this for a standard enlargement (actually, print size, so the degree of enlargement is going to differ with different sensor/film formats) and viewing distance. So, if you use the same lens, at the same point and same aperture, to take pictures on two different formats, the DoF on the image that's undergone less enlargement will be greater. If you crop a 35 mm frame so you're only using the same area that a 10D would, the DoF in the resultant prints would be the same. BUT that's not how most people use their lenses. Most people try to compose using as much of the sensor area as possible, so their prints from 35 mm would show greater DoF than their prints from 10D if they used the same lens, the same aperture, and stood at the same spot. Of course, if their visualization of the final image compels them to step further back with the 10D, or to use a wider lens with the 10D than they'd use with the 35 mm to capture the same subject area, the DoF is going to be different from either of the other previously-assumed conditions.

[DaveG]

Circle of Confusion refers to the maximum amount a point can be blurred on the film/sensor and still appear to the viewer as a point on subsequent enlargement. Depth of Field tables assume a value for this for a standard enlargement (actually, print size, so the degree of enlargement is going to differ with different sensor/film formats) and viewing distance. So, if you use the same lens, at the same point and same aperture, to take pictures on two different formats, the DoF on the image that's undergone less enlargement will be greater. If you crop a 35 mm frame so you're only using the same area that a 10D would, the DoF in the resultant prints would be the same. BUT that's not how most people use their lenses. Most people try to compose using as much of the sensor area as possible, so their prints from 35 mm would show greater DoF than their prints from 10D if they used the same lens, the same aperture, and stood at the same spot. Of course, if their visualization of the final image compels them to step further back with the 10D, or to use a wider lens with the 10D than they'd use with the 35 mm to capture the same subject area, the DoF is going to be different from either of the other previously-assumed conditions.


So depth of field varies based on enlargement too?

If I shoot a portrait with a 50 mm lens with a full 35mm camera I will move in and the shorter distance will reduce the depth of field that I'd get with the 50 on a 20D. But my point all along was that the depth of field from the same lens on ANY format will not change if the aperture and subject to distance diatance doesn't change. As for seeing less depth of field in a larger print that can be taken to the point where with a big enough print NOTHING is in focus.

[Jon]

So depth of field varies based on enlargement too?

If I shoot a portrait with a 50 mm lens with a full 35mm camera I will move in and the shorter distance will reduce the depth of field that I'd get with the 50 on a 20D. But my point all along was that the depth of field from the same lens on ANY format will not change if the aperture and subject to distance diatance doesn't change. As for seeing less depth of field in a larger print that can be taken to the point where with a big enough print NOTHING is in focus.

Yep, and also viewing distance. For any given lens, focussed on any given distance, and at any given aperture, any other single point (not focussed on) will have a constant circle of confusion regardless of the size of the film or sensor used to record it (neglection optical aberrations). If you make a contact print, and view it from "normal viewing distance", it'll look sharp through a much larger area than if you enlarge it 4x, and view it from the same distance. That's because your eye gets into the picture (you should pardon the expression) as well. If your eye can't resolve the enlarged circle of confusion, you'll see it as a point. When you enlarging the image, you'll see more points become blurs/disks because you've magnified the disk that's the original circle of confusion by the same amount you've magnified the entire picture. If you stand further back from the final print, you're effectively applying a reduction to everything, so the depth of field will appear greater. So anything you do to magnify the output image size will reduce DoF, while anything you do to reduce the output image size will increase DoF.

So, the circle of confusion, which is recorded on the film, or sensor, is constant with constant focal length and distance. The depth of field, which is what you're able to see in the output image, is also affected by the magnification applied to the original output image and by the distance you observe it from. If you viewed a 30" by 40" enlargement and an 8" x 10" enlargement of the same shot, from the same distance, don't you think you'd see more blur in the larger print?