Depth of Field (Bokeh) lesson

[petrochemist]

The longer MM a lens is, for a given distance at same aperture, the DOF for the longer lens will be thinner.

Technically this is complete garbage.
MM has no meaning here at all, what you meant I think was the Focal length which is usually measured in mm - changing the letters to caps completely changes their meaning. Technically there is a length unit of Mm, which is a megameter, 1 billion times the length of a millimeter. I never met anyone using this unit perhaps because of the ease of confusion.
Millimeters would also be suitable units for the aperture (if not measured as a ratio) and the DOF.
Your incomplete maths describes some of the aspects controlling DOF but isn't really much help.

The DOF formula is (with all distance measurements should be done in the same units):
DOF = 2 *aperture * circle of confusion * subject distance²/focal length².

Aperture being the f-stop (a unit-less number derived from the focal length of the lens divided by diameter of the opening.) not the physical diameter of the opening itself.

'circle of confusion' being the distance between two dots that the viewing system can JUST resolve - it's dependent on the viewing size but with digital systems is often taken as being the pixel size on the sensor. If the camera is kept the same it can be simply considered a constant for comparing the other aspects. Which is probably a good thing as it promotes so many technical arguments!

The other bits should need no explanation.

 

[Ysarex]

I show up to take that photo with my phone camera and you bring a FF DSLR. Click and click we take the same photo. Infinity is a limiting factor. Exposure is a limiting factor. The dirt road on the left is a limiting factor (looks bad) and the campground on the right is a limiting factor (ruins the composition). So we take the same photo and at any given f/stop value hyperfocal distance is closer to me than to you, yes because I have a shorter lens focal length, BUT ALSO because the circle of confusion is a smaller value for my camera than yours and that's because my camera has a smaller sensor (see chart above).

Joe

Joe, the circle of confusion is not a function of the sensor, it is not caused or defined by the sensor but the lens and primarily the diameter in mm of the effective aperture. Any camera with a smaller sensor uses a shorter focal length lens to take "similar" photos, (let's drop the "same" tag as I see what you mean). To maintain the same exposure you use the same f-stop which as a ratio of the focal length to effective aperture diameter means any similar photo shot on a smaller sensor with the same exposure is always using a narrower effective aperture. This is what affects the circles of confusion and gives the greater dof on smaller sensors.

If we make the assumption that two sensors have the same fov when one is shot with a 50mm lens and the other with a 35mm lens and that to maintain the same exposure (shutter speed and ISO the same) we use f4 on both. Then the image shot through the 50mm lens is shot through an effective aperture with a diameter of 12.5mm, and the image shot through the 35mm lens is shot through one of 8.75mm. With the fov being the same the smaller effective aperture limits the angle of light entering the lens making it more parallel and thus giving the greater dof. Exactly the same thing happens when we put the 35mm lens on the larger sensor, we get greater dof because of a shorter focal length and how it affects effective aperture while maintaining exposure. This effect is not in this case anything to do with sensor size but focal length, the optics and the change in the circles of confusion are the same in both cases.

It's only when you specify "same photo with same fov" that you are forced to change sensor size in proportion to focal length and it's this condition alone that brings sensor size into the equation.

You mean the condition of taking photos with a camera. So according to you DOF is the same for all format cameras and they can all be used to take photos with DOF that appears the same. I'd like to see you make that work. The fact is that to take a photo you have a camera in your hand. Your ability to render DOF is in part a function of the size of that camera's recording media. How "direct" the effect doesn't matter in the end when the finished photo is viewed.

Take the OP's original photo. He wanted to claim the extreme shallow DOF was due to his long 150mm lens. But my 5.2mm lens took a photo with even shallower DOF. DOF is a function of magnification and f/stop when you simplify the equation. And when you simplify the equation note it no longer contains lens focal length as a "direct" determinant -- (just like sensor size is not "direct") since magnification is a function of focal length and subject distance. But FL is there and can't be removed. And you also can't remove sensor size from the determination of Coc. Although both values are not "direct" you can't complete a valid DOF calculation without a numerical value that represents focal length and you can't complete a valid DOF calculation without a numerical value that represents sensor/film size. Show us otherwise. Show us a DOF calculator or set of DOF equations that don't contain a value that represents the size of the sensor/film. If Coc is in the equation it is representative in part of the sensor/film size; you cannot claim otherwise.

The relationships that govern dof work just the same when we do not take photos with the same fov.

The chart you show defines circles of confusion used to calculate dof for various format sizes based solely on image magnification (the size defined in your chart is magnified x number of times) to a 10"x8" print to be viewed at a certain distance. It is because you magnify the image made by the smaller sensor more than that of a larger sensor that the circle of confusion must be smaller to create the same dof in the 10"x8" print.

Duh... And that is in fact how DOF is defined by our discipline -- viewed on a finished 10x8 print from a normal viewing distance. We have long since agreed on that -- no other definition makes practical sense. Modifications of the calculations exist if you want to run them through for a larger print size. You can find online calculators that accommodate (Cambridge in Colour) but you can't find a DOF calculator that doesn't use a value for Coc that is determined in reference to the sensor size. Show us otherwise.

Joe

 

[Braineack]

Grabs popcorn, I love coc deniers

using tapatalk.

 

[beagle100]

OK

 

[Derrel]

Background Blur and Its Relationship to Sensor Size

Background blur and its relationship to sensor size
*********

 

[Derrel]

Keep in mind that depth of field is NOT the same thing as background blur!

Small-sensor cameras, like say the iPhone 7, or compact point & shoot digital cameras, can produce almost no real "background blur". Background blur is to a high degree, correlated with the actual, phusical WIDTH of the light-admitting aperture. The physically wider the aperture, the more blur that can be obtained; this becomes readily apparent when you use a 300mm f/2.8 lens, and shoot half-body portraits of a 6 foot tall person, using f/2.8 as the aperture value; the background on a half-body person is tremendously blurred. using an iPhone 7 at f/2.8, with anhalf-bopdy person shown in a standing shot, from the belt up to the top of their head, the SAME f/2.8 aperture produces a background that is not very blurred at all.

 

[jcdeboever]

I need to smoke a doobie... anyone got one?

 

[Tim Tucker]

So according to you DOF is the same for all format cameras and they can all be used to take photos with DOF that appears the same.

No, the image formed at the sensor of light recording media is the same for the same focal length, f-stop and subject distance. Not the finished image.

DOF is a function of magnification and f/stop when you simplify the equation. And when you simplify the equation note it no longer contains lens focal length as a "direct" determinant -- (just like sensor size is not "direct") since magnification is a function of focal length and subject distance.

When you simplify the equation then dof is a function of how parallel the light rays from a point object are passing through the lens, which is in turn a function of subject distance and aperture diameter, (minus the amount you have to magnify the image). Because we use exposure as constant then we use f-stop, the ratio of the focal length to aperture diameter and focal length affects it because it changes the effective aperture diameter. You can only remove focal length by using aperture diameter not f-stop, but of course that means that exposure must then vary across formats.

I did not remove sensor size from the equation, I said:

There are four variables that affect dof; distance to subject, aperture, focal length, and image magnification. Focal length is irrespective of sensor size as the only variable here is image magnification for final viewing. A 150mm lens has the same dof on any sensor before you magnify the image

But you said:

BUT ALSO because the circle of confusion is a smaller value for my camera than yours and that's because my camera has a smaller sensor (see chart above).

I am saying that in the statement above you confuse a correct understanding of dof.

Now you can say that sensor size affects dof and you'd be correct. But to say that a smaller sensor has greater dof because sensor size is part of the equation mis-understands that the role of smaller sensors in the equation is to decrease dof through extra magnification, not increase it. The smaller sensor does not have a smaller COC, the smaller sensor needs a smaller COC in order to produce the same dof in the magnified image.

The increase, or smaller COC created at the sensor, is entirely a function of the optics and is independent of sensor size. That you generally use shorter focal length lenses with smaller sensors, (and smaller effective aperture diameters) is why they generally have greater dof. If you were to use sensor size as a way of controlling dof then you may come across the flaws mentioned above, and find that the effects of magnification of the image to final viewing size is quite small compared to the effects of using shorter focal lengths with smaller effective aperture diameters.

 

[Ysarex]

I did not remove sensor size from the equation, I said:

There are four variables that affect dof; distance to subject, aperture, focal length, and image magnification. Focal length is irrespective of sensor size as the only variable here is image magnification for final viewing. A 150mm lens has the same dof on any sensor before you magnify the image

But you said:

BUT ALSO because the circle of confusion is a smaller value for my camera than yours and that's because my camera has a smaller sensor (see chart above).

I am saying that in the statement above you confuse a correct understanding of dof.

And I'm saying that in this statement, "A 150mm lens has the same dof on any sensor before you magnify the image" you confuse a correct understanding of DOF. DOF is not defined at the sensor or film. There isn't any DOF on the sensor. DOF is defined in the final image.

Now you can say that sensor size affects dof and you'd be correct.

That is all I said. Glad to see you agree. I said, "Don't forget the size of the recording media. All else held equal to take the same photo with cameras of different sensor size, there will be a DOF difference due to sensor size."

And you said, "Not really. The difference in dof is due to the magnification of the image to the final viewed size. With all being equal then a 150mm lens will produce the same circles of confusion regardless of the sensor size, it's just how much you magnify it in the finished image, (and how that relates to viewing distance)."

Which confuses a correct understanding of DOF. DOF is not defined at the sensor or film. DOF is defined in the final image which includes magnifying the image to the standard final print size. Coc therefore is specified differently for different size formats.

Trying to define DOF at/on the sensor or film is what trips people up and causes confusion.

But to say that a smaller sensor has greater dof because sensor size is part of the equation mis-understands that the role of smaller sensors in the equation is to decrease dof through extra magnification, not increase it. The smaller sensor does not have a smaller COC, the smaller sensor needs a smaller COC in order to produce the same dof in the magnified image.

You're nit picking in a salvage attempt. Yes, the Coc is not a property of the sensor. There is no DOF calculation that has a variable X = sensor size. I know that. But there is no DOF equation that does not have a variable c = Coc and in all of those equations that value for c is set relative to the sensor size. And that's what I mean when I say you can't remove sensor size from the equation. The smaller sensor needs a smaller Coc that's fine. Point is you must use that smaller Coc for each progressively smaller sensor to do a valid DOF calculation.

The increase, or smaller COC created at the sensor, is entirely a function of the optics and is independent of sensor size. That you generally use shorter focal length lenses with smaller sensors, (and smaller effective aperture diameters) is why they generally have greater dof. If you were to use sensor size as a way of controlling dof then you may come across the flaws mentioned above, and find that the effects of magnification of the image to final viewing size is quite small compared to the effects of using shorter focal lengths with smaller effective aperture diameters.

Never said a thing about which variables had the greater or lesser effect, that's off topic.

I think it comes down to this: You want to brush aside sensor size as accounted for in "image magnification" and consider image magnification as something occurring after DOF is determined. You think of DOF as defined at the sensor. I see that as confusion. I want to stick with tradition and be true to the equations we use to calculate DOF and that define DOF on the final print -- the way our discipline has done it since before we were both born. There are no standard equations to calculate DOF at the sensor. You can't find a DOF calculator that does that. Go to DOF Master or Cambridge in Colour and use the DOF calculators and they incorporate a value for Coc relative to the sensor size. Download one for your phone -- does the same thing.

Surely there's no argument that you get generally deeper DOF from smaller sensor cameras and shallower DOF from larger sensor cameras. Same was true with film and our traditional DOF equations reflect that. I'm a traditional kind of guy. So when understanding this equation:

I want to be able to say that the value of c changes with the size of the sensor which therefore appropriately reflects the phenomena that smaller format cameras have deeper DOF and vise versa. Sensor size therefore is a DOF determinant factor since you can't calculate DOF without it.

Joe

 

[OGsPhotography]

Good thread, I am a circle of confusion now .

 

[Tim Tucker]

In my very first example I threw out your condition of fov being constant and introduced focal length as being constant, and guess what happens? Dof decreases with smaller sensors. Now this is not the whole story, just as your examples are not the whole story, but your thesis that smaller sensors produce greater dof does not hold true in all cases. All I've been saying, and I think I've been quite clear about it, is that it's the use of shorter focal length lenses associated with smaller sensors and the use of smaller aperture diameters to maintain exposure, not the sensor size.

I want to be able to say that the value of c changes with the size of the sensor which therefore appropriately reflects the phenomena that smaller format cameras have deeper DOF and vise versa. Sensor size therefore is a DOF determinant factor since you can't calculate DOF without it.

Just because a smaller sensor requires a smaller value for c does not mean that sensor size determines dof. c is not dof or sensor size it is the limit for the size of the coc and nothing else. COC is entirely derived from the optics.

Smaller sensors actually have a negative effect on dof, they decrease it. It's only when you reduce the aperture diameter or increase subject distance that dof increases, if you simply reduce sensor size and hold focal length, subject distance and aperture constant then dof decreases with a smaller sensor.

1) If you have a set up where all you change is the camera body for one of a smaller sensor (the optics, exposure and subject distance remain the same), the dof decreases with the smaller sensor.

2) If you change the lens for a shorter focal length, (sensor size, exposure and subject distance remain the same), then dof increases.

Now if you do both, (change the sensor size to a smaller one and change the lens for one of a shorter focal length but maintain fov, subject distance and exposure constant), then the extra magnification needed for the smaller sensor decreases dof and the use of a shorter focal length (and the smaller aperture diameter to maintain exposure) increases it, but does so to a much larger extent. The size of the COC does not change with sensor size, it only changes with subject distance, f-stop and focal length, (or subject distance and aperture diameter, but please note that to maintain exposure it is much better to use f-stop and therefore it is subject distance, f-stop and focal length). The only effect sensor size has is in the magnification to the final reference print, and smaller sensors need greater magnification.

I cannot be clearer.

(Seeing as the value of c is directly derived from the sensor size it is no surprise that it changes with sensor size, x + y = z then z - y = x, it's nothing to do with the lens!)

 

[Ysarex]

In my very first example I threw out your condition of fov being constant and introduced focal length as being constant, and guess what happens? Dof decreases with smaller sensors.

No need to guess, of course that's obvious. There is no point in comparing different photos. If you put a 165mm Angulon on an 8x10 camera and shoot the Grand Tetons that photo will have more DOF than a macro photo of a fly taken with a 165mm lens on a 35mm camera. Duh... so what.

The only comparison that makes any logical sense is to take the same photo.

Now this is not the whole story, just as your examples are not the whole story, but your thesis that smaller sensors produce greater dof does not hold true in all cases.

It's not a thesis it's proven and accepted fact. Here's a reference: "'Same picture' for both formats
When the 'same picture' is taken in two different format sizes from the same distance at the same f-number with lenses that give the same angle of view, and the final images (e.g., in prints, or on a projection screen or electronic display) are the same size, DOF is, to a first approximation, inversely proportional to format size (Stroebel 1976, 139)." Depth of field - Wikipedia

Notice the reference to Stroebel. That would be Leslie Stroebel and his definitive text View Camera Technique. I'm not going to type the whole paragraph so I took a snapshot (page 130 5th edition):

All I've been saying, and I think I've been quite clear about it, is that it's the use of shorter focal length lenses associated with smaller sensors and the use of smaller aperture diameters to maintain exposure, not the sensor size.

Chicken and egg argument and you're claiming the egg doesn't matter which is irrational.

I want to be able to say that the value of c changes with the size of the sensor

But you can't, because it doesn't.

Then you need to change the Wikipedia article on Circle of Confusion and get this chart out of there:

And then you need to contact Dan Fleming at DOFMaster and get him to remove this page: Circles of Confusion for Digital Cameras Because they're changing the Coc value predicated on the size of the format.

c is not dof or sensor size it is the limit for the size of the coc and nothing else. You cannot equate something that is entirely derived from the optics as being a function of the sensor.

I never claimed Coc was a function of the sensor. I know better. I said that in an equation used to calculate DOF the value of Coc is selected predicated on the size of the recording media. I remain correct. I also said you can't find a DOF calculator that doesn't work that way. I posted this earlier:

Please note the two different values for Circle of confusion. They are different because the sensor sizes are different, an FX and DX camera. Please provide a link to an online DOF calculator that doesn't do that.

Smaller sensors actually have a negative effect on dof, they decrease it. It's only when you reduce the aperture diameter that dof increases, if you simply reduce sensor size and hold focal length and aperture constant then dof decreases with a smaller sensor.

1) If you have a set up where all you change is the camera body for one of a smaller sensor (the optics, exposure and subject distance remain the same), the dof decreases with the smaller sensor.

Two different photographs and so no comparison being made -- no meaning.

2) If you change the lens for a shorter focal length, (sensor size, exposure and subject distance remain the same), then dof increases.

Two different photographs and so no comparison being made -- no meaning.

Now if you do both, (change the sensor size to a smaller one and change the lens for one of a shorter focal length but maintain fov, subject distance and exposure constant), then the extra magnification needed for the smaller sensor decreases dof and the use of a shorter focal length (and the smaller aperture diameter to maintain exposure) increases it, but does so to a much larger extent.

Same photo and deeper DOF in the final print from the smaller sensor camera. Yep, that's what I've been saying.

The value of c does not change with sensor size,

In all equations used to calculate DOF the value of c is changed when the size of the sensor is changed. That's obvious from the screen shots above. Please provide a link to a DOF calculator that doesn't do that. That doesn't mean that Coc is a function of sensor size. It doesn't mean that sensor size alters the physics of optics. It simply means that we can't calculate DOF without a value that references the size of the sensor. An accommodation in the equation is made to account for the size of the sensor. Please provide a link to an online DOF calculator that doesn't do that.

Same hang-up, you're trying to calculate DOF at the sensor. It's not defined that way. You can't calculate DOF without including a value that is predicated in part on the size of the sensor. As such changing the size of the sensor changes the result of the calculation. Show us a calculator that doesn't do that.

Pragmatically one more time: Take the "same photo" with two cameras of different sensor size and the photo from the smaller sensor camera will have deeper DOF. You just said that above and that's what I originally said, and the equations that are used to do that calculation include a value predicated on that sensor size. Show us a DOF calculator that doesn't do that.

I didn't say Coc is a function of sensor size. I didn't even say that Coc changes optically when we change the sensor size. We change the value of Coc in the equations when we change sensor size. We get more accurate and practically useful results from the equations when we do that. We don't calculate DOF any other way -- can't calculate DOF without a value referenced to the sensor size and so sensor size is a DOF determinant.

Joe

it only changes with subject distance, f-stop and focal length, (or subject distance and aperture diameter, but please note that to maintain exposure it is much better to use f-stop and therefore it is subject distance, f-stop and focal length). The only effect sensor size has is in the magnification to the final reference print, and smaller sensors need greater magnification.

I cannot be clearer.

 

[Derrel]

I was at the library for a couple of hours earlier this week. I picked up a book by a well-known photographer, and in it I found a little graphic that had some distilled wisdom. In this graphic, he said that for "storytelling pictures", apertures of f/16 and f/22 were often the most useful.

For single-themed or single-subject images, he suggested apertures of f/4 and f/5.6 were often useful.

For subjects where depth of field was not normally a significant factor, he suggested f/8 and f/11.

In another area of the book, he mentioned that when shooting a lens wide-open, "What you see through the viewfinder is basically, what you will get."

 

[Ysarex]

I was at the library for a couple of hours earlier this week. I picked up a book by a well-known photographer, and in it I found a little graphic that had some distilled wisdom. In this graphic, he said that for "storytelling pictures", apertures of f/16 and f/22 were often the most useful.

For single-themed or single-subject images, he suggested apertures of f/4 and f/5.6 were often useful.

For subjects where depth of field was not normally a significant factor, he suggested f/8 and f/11.

In another area of the book, he mentioned that when shooting a lens wide-open, "What you see through the viewfinder is basically, what you will get."

Yes, and thank you for that reminder.

One of the reasons I'm toughing this one out is that it's really useful to have general rules that hold true. One of those general rules is that cameras with smaller sensors/film generally produce photos with deeper DOF and cameras with bigger sensors/film generally produce photos with shallower DOF. It's not a huge difference when stepping from one camera to next size, but it is a substantial difference say between an M4/3 and old sheet film camera. This is a helpful general rule to know and to know the rule requires that we identify the role of the sensor/film size in determining DOF. It's helpful to be able to state the rule simply and directly. This one is a winner in that regard.

Joe

 

[Derrel]

I should probably have mentioned, the famous photographer mentioned above shoots mostly with APS-C Nikons, and FX Nikons. The book has mostly "new" work in it, from the Nikon D300s (APS-C), and then the Nikon D3x, and finally, the Nikon D800e or D810 (all three of which are 24x36mm sensor sized). The majority of the images were made with a 24-85mm Nikkor lens; a 105mm Nikon macro lens, and the Nikon 70-300 VR lens, all three of which are pretty common lenses, affordable by the majority of enthusiasts.

I was not too surprised to see many,many good pictures made with the 24-85 zoom lens, and with the light-in-weight 70-300mm VR lens.

While many people fuss over using f/16 and f/22 due to softening caused by diffraction, my personal experiene is that the deeper depth of field that can be obtained by stopping down to f/16 or f/22 on a 24x36mm sensor, or a smaller sensor like APS-C, is in many ways, sort of a theoretical worry more so than the disastrous "ruiner of pictures" so many internet folks seem to classify the small f/stops as being! I totally, totally understood what he meant by suggesting f/16 and f/22 for "storytelling pictures".