Help me understand these DoF calculators

[Great Bustard]

I went to http://www.cambridgeincolour.com/tutorials/dof-calculator.htm to try the dof calculator and what perplexed me was that by changing the format from 1.5 crop to full frame and keeping all other parameters the same I get a deeper dof.

I have always thought that using the same lens at the same aperture and focusing at the same distance would yield the same dof no matter what sensor was put behind the lens.

@Great Bustard - you seem to be ignoring the OP's point (bolded above).

Actually, the first scenario given by GB is exactly what the OP asked: same focal length and shooting distance, relative aperture (aka f-number); with the printed output sized proportional to the sensor size -- in that case the OP's understanding would be confirmed; depth of field would be the same.

However, if the print size in both cases is the same -- which is how photos would be typically viewed and how they would be properly compared; and which is the assumption built into the DoF calculator -- then the depth of field yielded by the smaller sensor would be shallower.

GB is pointing out the the evaluation of depth of field requires accounting for all of the variables, including not just the shooting parameters, but all of the steps to producing a final, viewable image.

Can anyone help me understand this?

The easiest way to help you understand what's going on is to tell you that if you took a photo of the same scene from the same position with the same focal length and relative aperture using cameras with different sensor sizes,

I agree except for the relative aperture because that deviate from the OP's issue.

The relative aperture is the f-number. Although the OP used the vernacular "aperture" it is best to be clear whether one means the actual aperture size, or the that size relative to the focal length, which is what is more commonly meant. GB was being clear about that.

printed the photos out at a size proportional to the sensor size (e.g. 1.5x at 12x18 inches and FF at 18x24 inches), and cropped all the photos to the same framing, then the DOFs would be the same.

Why all the fuzz about printing when printing will introduce another unwanted element? Just do a side-by-side (split screen) of unretouched photos is easier. The two photos I posted above will just do as well. Those are unretouched, not cropped. The filesize was reduced to accommodate DPR filesize upload requirement.

Assuming you mean your earlier post, it isn't clear what point you are making with those. But if images taken with different sensor sizes but identical shooting conditions (including focal length, relative aperture, and subject distance), and are then displayed at the same size and viewing distance, then the one from the smaller sensor will exhibit less depth of field. Is that what you were illustrating?

Alternatively, photos of the same scene taken from the same position with the same framing and aperture diameter (e.g. the aperture diameter for 100mm f/2 is 100mm / 2 = 50mm) displayed at the same size and viewed from the same distance will have the same DOF for all systems.

Aperture diameter is irrelevant for three reasons: (1) almost physically impossible to measure the aperture diameter (2) The lens has already marked aperture ring (3) changing the aperture setting is not in-line with the OP.
I posted your suggestion at least two hours ahead of your posting. Look above your post.

Well here you are kind of off the rails. The actual aperture diameter (not the relative aperture) is what directly determines depth of field. As GB points out, if those are normalized (and everything else, including framing, is the same), then depth of field will be the same. Understanding that is crucial. Incidentally, one need not measure the actual aperture diameter, is is easy and convenient to determine it from the known focal length and relative aperture, as GB shows.

For example, photos of a scene from the same position taken at 100mm f/4 on 1.5x and 150mm f/6 on FF will have the same DOF if displayed at the same size.

Again, using two different Aperture and FL does not address the spirit of the OP.

The OP wanted to better understand how the various parameters work interdependently to influence depth of field, including cases where the depth of field is the same (as he expected) and why they would be different, as he discovered from the DoF calculator. GB's discussion is 100% on point.

For a more in-depth (and technical) explanation, see here.

Dave

Perfectly stated -- nothing more for me to add.

 

[HelplessCorgis]

gee I always thought that DOF would be the same going from crop to full frame, if all other variables were the same. Learn something new everyday...

 

[Michael Fryd]

gee I always thought that DOF would be the same going from crop to full frame, if all other variables were the same. Learn something new everyday...

It depends on which variables you keep the same.

If you keep angle of view, and aperture diameter the same, then DOF will be the same.

100mm at f/2 (50mm aperture diameter) on a crop and 150mm at f/3 (also 50mm aperture diameter) on a full frame will yield the same angle of view, same depth of field, and he same picture. If you keep shutter speed the same (and adjust ISO to compensate) you will also likely get the same overall image noise.

The reason there are disagreements over Depth of Field and sensor size, is that different people have different ideas of what it means to keep everything else "the same".

 

[fredflintstone]

I went to http://www.cambridgeincolour.com/tutorials/dof-calculator.htm to try the dof calculator and what perplexed me was that by changing the format from 1.5 crop to full frame and keeping all other parameters the same I get a deeper dof.

I have always thought that using the same lens at the same aperture and focusing at the same distance would yield the same dof no matter what sensor was put behind the lens.

Can anyone help me understand this?

My understanding of what's going on is that, for a start, using the same lens (e.g. a 50mm) will give different angles of view on different sensor sizes but it will give the same DoF at the sensor. However, the factor you haven't considered is the degree of enlargement for screen display or for printing.

DoF is a perceptual matter. Lets assume that under given viewing conditions the average human eye will perceive a dot 0.1mm across as a small disc rather than as a point. Anything smaller than this will look like a point. An image is made up of such points.

If you take a sharply focused image of a thin line, for example, and print it to a size (say 8"x12" or 20x30cm) where the line is less than 0.1mm thick then it will look sharp. If the line is slightly out of focus at the sensor then it will be wider at that stage, so you won't be able to print quite so big and have a sharp looking line.

To print an 8"x12 from FF requires that you enlarge the whole sensor image by a factor of 8x. To get the same size print from an APS-C sensor you need to enlarge 12x.

If the FF image has a line 0.01mm across it will still look sharp enlarged by 8x (0.08mm wide) . Meanwhile the APS-C image will look fuzzy because it will be enlarged by 12x , making it 0.12mm wide -- enough for the human eye to see the blur.

All this has nothing to do with the fact that you get different angles of view.

--
Albert
(The one in France)
There is no such thing as a professional camera.
A 'pro' or 'professional' photographer is someone who earns money from photography. It is not some sort of measure of quality or expertise.
Cameras don't get paid.

I always assumed DoF was relative. I don't know an exact definition, but say that the DoF is where the resolution degrades by 100% (the imaged points double in diameter) from what it is at the plane of focus (the distance at which your lens is focused). So if you enlarge the print (whether it is FF or crop sensor), the resolution (the x0.707 width of the point spread function) of the image at both the plane of focus and the 100% degradation planes will both be degraded by the same amount, but the relative amounts (ratios) will stay the same.

If it is absolute, as in your example, if you take the FF image and enlarge it to say 16"x24", the line will now exceed the 0.1mm thick criteria, and will not look sharp. So the DoF of the FF image has now changed just by enlarging it further. I don't buy it.

Can anyone give an exact definition of DoF?

--
Stone Age (former film) Photographer.
Yabadabado!

 

[Adrian Tung]

Can anyone give an exact definition of DoF?

From Canon EF Lens Work III, Optical Terminology & MTF Characteristics:

Loading…

software.canon-europe.com

Depth of Field:

The area in front of and behind a focused subject in which the photographed image appears sharp. In other words, the depth of sharpness to the front and rear of the subject where image blur in the focal plane falls within the limits of the permissible circle of confusion.

And from the same document,

Permissible Circle of Confusion:

The largest circle of confusion which still appears as a “point” in the image. Image sharpness as sensed by the human eye is closely related to the sharpness of the actual image and the “resolution” of human eyesight. In photography, image sharpness is also dependent on the degree of image enlargement or projection distance and the distance from which the image is viewed. In other words, in practical work it is possible to determine certain “allowances” for producing images which, although actually blurred to a certain degree, still appear sharp to the observer. For 35mm single lens reflex cameras, the permissible circle of confusion is about 1/1000~1/1500 the length of the film diagonal, assuming the image is enlarged to a 5”×7” (12 cm × 16.5 cm) print and viewed from a distance of 25~30 cm/0.8~1 ft. EF lenses are designed to produce a minimum circle of confusion of 0.035 mm, a value on which calculations for items such as depth of field are based.

 

[Michael Fryd]

...I always assumed DoF was relative. I don't know an exact definition, but say that the DoF is where the resolution degrades by 100% (the imaged points double in diameter) from what it is at the plane of focus (the distance at which your lens is focused). So if you enlarge the print (whether it is FF or crop sensor), the resolution (the x0.707 width of the point spread function) of the image at both the plane of focus and the 100% degradation planes will both be degraded by the same amount, but the relative amounts (ratios) will stay the same.

If it is absolute, as in your example, if you take the FF image and enlarge it to say 16"x24", the line will now exceed the 0.1mm thick criteria, and will not look sharp. So the DoF of the FF image has now changed just by enlarging it further. I don't buy it.

Can anyone give an exact definition of DoF?

"Depth of Field" is the distance between the minimum and maximum subject distances which appears to be acceptably focused in the final print.

Note that this is an attribute of both the final print and the viewing conditions, not the image itself.

For instance, certain objects may appear acceptably sharp when viewing a print from 5 feet away, but when you move in to 2 feet away they don't look focused. This is an example of how Depth of Field can change with viewing distance.

Depth of field calculators make assumptions about typical viewing circumstances. They provide good estimates of Depth of Field for those circumstances.

.

Here's an example of how size affects sharpness. Consider this slightly out-of-focus image of a semi circle.



The image has been printed in 6 different sizes. To most people the smallest version on the right appears sharper than the larger version. Yet, they are the same image, just printed at different sizes. This demonstrates that apparent sharpness is dependent on viewing size, and is not a fixed property of the image.

 

[fredflintstone]

Can anyone give an exact definition of DoF?

From Canon EF Lens Work III, Optical Terminology & MTF Characteristics:

http://software.canon-europe.com/fi...0_EN.pdf?_ga=1.105905031.543256715.1484022656

Depth of Field:

The area in front of and behind a focused subject in which the photographed image appears sharp. In other words, the depth of sharpness to the front and rear of the subject where image blur in the focal plane falls within the limits of the permissible circle of confusion.

And from the same document,

Permissible Circle of Confusion:

The largest circle of confusion which still appears as a “point” in the image. Image sharpness as sensed by the human eye is closely related to the sharpness of the actual image and the “resolution” of human eyesight. In photography, image sharpness is also dependent on the degree of image enlargement or projection distance and the distance from which the image is viewed. In other words, in practical work it is possible to determine certain “allowances” for producing images which, although actually blurred to a certain degree, still appear sharp to the observer. For 35mm single lens reflex cameras, the permissible circle of confusion is about 1/1000~1/1500 the length of the film diagonal, assuming the image is enlarged to a 5”×7” (12 cm × 16.5 cm) print and viewed from a distance of 25~30 cm/0.8~1 ft. EF lenses are designed to produce a minimum circle of confusion of 0.035 mm, a value on which calculations for items such as depth of field are based.

Thanks, but I have just been reading on Wikipedia, something I should have done before opening my mouth (seems like I did read this info some time ago, just forgot).

So what I get out of this so far (more reading to do), it is based on the 'acceptable circle of confusion' criteria, which is based on visual perception of the largest circle size that is still perceived as a point, given a standard print size (Wikipedia uses 8"x10") viewed at a standard distance. So if you make a large print of the same image, the DoF decreases! And here all these decades I thought it was a relative thing, that the print size didn't matter. Not sure those DOF markings on legacy lenses (and some modern ones) are really of much use then.

 

[dsjtecserv]

Can anyone give an exact definition of DoF?

From Canon EF Lens Work III, Optical Terminology & MTF Characteristics:

http://software.canon-europe.com/fi...0_EN.pdf?_ga=1.105905031.543256715.1484022656

Depth of Field:

The area in front of and behind a focused subject in which the photographed image appears sharp. In other words, the depth of sharpness to the front and rear of the subject where image blur in the focal plane falls within the limits of the permissible circle of confusion.

And from the same document,

Permissible Circle of Confusion:

The largest circle of confusion which still appears as a “point” in the image. Image sharpness as sensed by the human eye is closely related to the sharpness of the actual image and the “resolution” of human eyesight. In photography, image sharpness is also dependent on the degree of image enlargement or projection distance and the distance from which the image is viewed. In other words, in practical work it is possible to determine certain “allowances” for producing images which, although actually blurred to a certain degree, still appear sharp to the observer. For 35mm single lens reflex cameras, the permissible circle of confusion is about 1/1000~1/1500 the length of the film diagonal, assuming the image is enlarged to a 5”×7” (12 cm × 16.5 cm) print and viewed from a distance of 25~30 cm/0.8~1 ft. EF lenses are designed to produce a minimum circle of confusion of 0.035 mm, a value on which calculations for items such as depth of field are based.

Thanks, but I have just been reading on Wikipedia, something I should have done before opening my mouth (seems like I did read this info some time ago, just forgot).

So what I get out of this so far (more reading to do), it is based on the 'acceptable circle of confusion' criteria, which is based on visual perception of the largest circle size that is still perceived as a point, given a standard print size (Wikipedia uses 8"x10") viewed at a standard distance. So if you make a large print of the same image, the DoF decreases! And here all these decades I thought it was a relative thing, that the print size didn't matter. Not sure those DOF markings on legacy lenses (and some modern ones) are really of much use then.

You are exactly correct -- if the print size is larger (and the view distance stays the same) then the depth of field would be less. Since it would be impossible to predict the ultimate print size and viewing distance at the moment of capture (and the same image could be printed at different sizes anyway), it has always been necessary to standardize on print size and viewing distance that was considered the most common. The depth of field scales on lenses were (and are) based on those assumed viewing conditions (as well as assumed film/sensor size). And yes, if the actual print is different or the viewing distance is different, then those DoF scales are invalid.

It's always been that way, even back in the film days. We just didn't have the internet to allow us to talk about it as much!

Dave

 

[tony field]

Can anyone give an exact definition of DoF?

From Canon EF Lens Work III, Optical Terminology & MTF Characteristics:

http://software.canon-europe.com/fi...0_EN.pdf?_ga=1.105905031.543256715.1484022656

Depth of Field:

The area in front of and behind a focused subject in which the photographed image appears sharp. In other words, the depth of sharpness to the front and rear of the subject where image blur in the focal plane falls within the limits of the permissible circle of confusion.

And from the same document,

Permissible Circle of Confusion:

The largest circle of confusion which still appears as a “point” in the image. Image sharpness as sensed by the human eye is closely related to the sharpness of the actual image and the “resolution” of human eyesight. In photography, image sharpness is also dependent on the degree of image enlargement or projection distance and the distance from which the image is viewed. In other words, in practical work it is possible to determine certain “allowances” for producing images which, although actually blurred to a certain degree, still appear sharp to the observer. For 35mm single lens reflex cameras, the permissible circle of confusion is about 1/1000~1/1500 the length of the film diagonal, assuming the image is enlarged to a 5”×7” (12 cm × 16.5 cm) print and viewed from a distance of 25~30 cm/0.8~1 ft. EF lenses are designed to produce a minimum circle of confusion of 0.035 mm, a value on which calculations for items such as depth of field are based.

Thanks, but I have just been reading on Wikipedia, something I should have done before opening my mouth (seems like I did read this info some time ago, just forgot).

So what I get out of this so far (more reading to do), it is based on the 'acceptable circle of confusion' criteria, which is based on visual perception of the largest circle size that is still perceived as a point, given a standard print size (Wikipedia uses 8"x10") viewed at a standard distance. So if you make a large print of the same image, the DoF decreases! And here all these decades I thought it was a relative thing, that the print size didn't matter. Not sure those DOF markings on legacy lenses (and some modern ones) are really of much use then.

I just compared using DOF Master, and the lens marking for DOF on two leica lenses. They pretty well match the calculations. In other words, based on the DOF default assumptions for COC, image size, and viewing, the indicated numbers reflect this nominal reality.

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

 

[Matsu]

Can anyone give an exact definition of DoF?

From Canon EF Lens Work III, Optical Terminology & MTF Characteristics:

http://software.canon-europe.com/fi...0_EN.pdf?_ga=1.105905031.543256715.1484022656

thanks for the link, a very helpful glossary +1

 

[Michael Fryd]

...So if you make a large print of the same image, the DoF decreases! ...

Maybe.

If you make a larger print, you may very well view it from a longer distance, and the Depth of Field might not noticeably change.

If you make an 8x10, you may view it from 2 to 3 feet away. If you make a 16x20, you may stand back and view from a comfortable 4 to 6 feet away. At 32x40, you may step back to 8 to 12 feet. If you adjust your viewing distance is correlated to the print size, the Depth of Field may not change much with print size.

.

However, it turns out that the LCD preview on the back of the camera is rather small for the viewing distance. The Depth of Field you see on that LCD screen, is likely greater than what you will get in your final print.