[D4rknezz]

Hey guys,

Is there a way to calculate how blurry a background will be? (I can't believe I am just thinking of all these questions now, by the way)
I mean, I have a DOF calculator. It tells me that if I am using...a 50mm lens, and the subject is...err..3 meters away, i'd get sharp images up to ...whatever...5meters away...then blurry afterwards.

This is all good - except it doesnt tell me how blurry. Is it pleasing thin, or is it still there kind of sharpness? I want to make my subjects pop right out of the background, but I think the background needs to be blurry enough that almost no detail smaller than ..say...7 inches is preserved. How do i get this with any consistency?


Especially when taking family photos. Consider that i am not in a studio, and the distance to the background varies depending on where i am, but in general, we dont have a large expanse of empty background in the city. Is cheating in PP the only way to do this then? go as large an aperture as I can afford, hopefully 2.8, use a focal length as long as I can afford, hopefully longer, and pray that the background is far enough to pop my subject out?

Also, would a different focal length for an equivalent composition yield the same amount of dof? for example...if I had my prime 50mm with me, i stand 2 meters away from the subject and get everything in frame, and render a pleasing dof. Imagine i swap out the 50 with a 125, stand 7 meters away (just a stab in the dark) and get the same things in frame. Do i get the same dof? Considering that the ratio of the distance between me and the subject and the subject and the background has now changed....

Thanks guys.

[vonBaloney]

Use optical preview to see for yourself. Or digital preview, or just shoot a lot of tests til you get a feel for it. Focal length of the lens has a huge effect, as well as distance to subject. Someone else can explain the math...

[Na Horuk]

Yeah, different focal lengths will give different apparent results, and aperture will have an effect on it as well.
The general rule is that lower f-stop, closer focusing, and tele lenses will give more apparent blur. This isn't the whole math, but it is good enough. The actual blur (or bokeh) will depend on what lens you are using (as different lenses will produce different bokeh, flares, and other distortions). Some make a nervous, jagged blur, others a creamy one, some make swirly bokeh (like the russian Helios 44-2), and some even donut bokeh (mirror lenses).
Also, look up zone focusing, this might help you out in some cases.

[D4rknezz]

Thanks always wondered about those dials - didnt know the name for it. I see a lot of calculator, but really, because of the different variables involved when in location, i was hoping there'd be a math somewhere for me to plug in. I mean...put a 50-135 in front of your lens and at any given moment your dof is different. Sigh.

[twitch]

Google DOF calculator and it will tell you how blurry it will be. However this is one thing that is easy to figure out yourself. Blurriness of the background (with all else held the same esp. distance to subject) is most dependant on the aperture size (focal length / aperture f-stop). So a DA 55-300 @ 300mm wide open will give you a blurrier distant background than, say, a 50mm f/1.4 wide open becuase the aperture is bigger (52mm vs 36mm).

There's an excellent article here on this topic http://www.bobatkins.com/photography/technical/bokeh_background_blur.html

[Lowell Goudge]

The thing with using a DOF calculator is it really tells you how much front to back distance you have, for a dot to remain a dot.

What the OP really wants to know (in my opinion) is two things, first of all, how big a circle will an out of focus dot appear in his image, because blurred backgrounds are really just continually overlapping big circles, and second, and perhaps not easy to identify, is the quality of the OOF circle, i.e. hard edged, soft edged, uniform intensity varrying intensity etc... the first could be calculated using a derivation of a DOF calculator, the second is a lens specific quality.

Whiel the first may be possible, I think using either live view or DOF preview is a better bet. BUT if you are really looking for blurred backgrounds just make sure you have separation between subject and background, and select backround that does not have large hard regular shapes.

[D4rknezz]

Thanks guys. The combined information here is great.

I wasnt aware of the aperture size being a function of focal length / f-stop, i kind of thought f-stop is aperture. Great new info. As well as Zone-Focusing. Lowell was correct in what he surmised what I wanted - and I think the answer to it would be figuring out what kind of lens do I want to use as my walk around with zone focusing and aperture size - get a prime of that focal length and then always take a picture at a specific distance subject/background ratio. That would keep the consistency of my pictures.

Of course, when distance can't be manipulated thats when everything goes wrong

Now here is a related question, related to the aperture size :

If using a 50mm 2.8 i need to be 2 meters away from a subject to get a field of view of half the body, the background is 6 meters away from the subject. (36mm aperture size)

|__|______| (camera, subject, background)


Because I wanted a greater aperture size, I switched to a 135mm 2.8 (49 mm aperture size). however, i now need to stand 5 meters away to get the same field of view

|_____|______|


Wouldnt the DOF of the 50mm , despite being a smaller aperture size, be better in this case (because of the relative separation between subject to background is bigger)



Some photographers do this like this is science, if I see their work. There must be a method to their madness .

[Lowell Goudge]

Ok here is a simple idea.

Consider that for an ASP-C sensor (pentax DSLR) that depth of field is defined as a circle of confusion of 0.020 millimeters (20 microns) What this means is that a point expands to 20 Microns at the limit of acceptable focus. As I said earlier, you could in theory project this back to what ever circle diameter you want at some distance, based upon determining some acceptable degree of blurriness, from tests. But regardless, that distance will be geometrically linked to your depth of field calculation.

Now you change lenses, yes the 135 needs to be further away to get the same field of view as the 50mm. this is because for any lens when focusing much further away than focal length, the net is

Image size = subject size x focal length / distance

but look at the depth of field of the 135mm lens, it is much shallower, so you don't need as much distance behind the subject to project the circle of confusion to the same size as you did with a 50mm lens.

Without trying to prove it, just from a "gut instinct" I would think that you could simply plot this using LOG paper for distance and draw a straight line projection from the perfect focus point, through 20 microns at the limit of depth of field behind the subject, and continue on. if you plot this also for the 50mm lens, and then take a third point for the 50mm lens at the known background distance this would give you the projected circle size, for your blurry background, this circle size should be the same for the 135 to have the same foreground vs background separation.

If you can't follow this, let me know and I will try to graph the concept.

[tuco]

It seems any calculation of "blurriness" is going to be a number as apposed to an answer like "not blurry, "blurry", "really blurry" and "extremely blurry". You'd still have to per-visualize what that's going to look like anyway so why not just use the camera to tell you that.

[Lowell Goudge]

It seems any calculation of "blurriness" is going to be a number as apposed to an answer like "not blurry, "blurry", "really blurry" and "extremely blurry". You'd still have to per-visualize what that's going to look like anyway so why not just use the camera to tell you that.

thats why I suggested that until you determine the degree of blurriness for one lens, aperture and foreground / background spacing all you have is a calculation, but if you have one specific degree of blurriness in mind, and know the background distance that produced it with one lens, it should be possible to predict the new shooting geometry for any other lens. I think that is the OP's point, how to set up.

Thinking further on this idea, I can see perhaps a set of charts, perhaps 3 in total, with focal length on the X axis, background distance on the y axis and a series of lines radiating out of the specific focal lengths for portraits, head shots, half bodies and full bodies, (each line from one focal length would have a different slope as a function of aperture) so you could, once you know one shooting setup, change lenses, and move the background to get the same effect quickly

[D4rknezz]

@ Taco

Hahah its true. Except I think it works best only when I have several primes. With several zooms, testing becomes difficult : The variable changes all the time. OO Look! A bird! (135mm zoom, bg at infinity...). Ooo! Look! A phone booth! (50mm zoom, bg at 2 feet away...)...
If i kind of get the math, I figure I can restrict my focal length choice to a specific size, and my background choices too. Also...sometimes its just timing. I am at my day work till dark...then at the weekend...I shoot in studio or do post processing. Too much working with camera to experiment with it! LOL.

@Lowell... I think I asked beyond my capability to comprehend Please walk me through I think what you are saying is that despite the relative difference between me - subject - bg, for the same field of view i should still get the advantage of the longer zoom because of the size of the aperture. I still have in my mind that there must be a limit to this, like a 1000 mm zoom vs 50mm , if trying to get the same fov.

Anyway, my confusion below, bolded :

Ok here is a simple idea.

Consider that for an ASP-C sensor (pentax DSLR) that depth of field is defined as a circle of confusion of 0.020 millimeters (20 microns) What this means is that a point expands to 20 Microns at the limit of acceptable focus. As I said earlier, you could in theory project this back to what ever circle diameter you want at some distance, based upon determining some acceptable degree of blurriness, from tests. But regardless, that distance will be geometrically linked to your depth of field calculation.
To clarify, the 20 microns is at the acceptable focus, so everything beyond 20 microns is considered out of focus?

Now you change lenses, yes the 135 needs to be further away to get the same field of view as the 50mm. this is because for any lens when focusing much further away than focal length, the net is

Image size = subject size x focal length / distance

What is this image size, and how is this calculated? Is it an image size by MP? or by inches? So if the subject is 6ft tall, focal length is 50mm, distance is 5 meter, what do i get for an image size?

but look at the depth of field of the 135mm lens, it is much shallower, so you don't need as much distance behind the subject to project the circle of confusion to the same size as you did with a 50mm lens.

Without trying to prove it, just from a "gut instinct" I would think that you could simply plot this using LOG paper for distance and draw a straight line projection from the perfect focus point, through 20 microns at the limit of depth of field behind the subject, and continue on. if you plot this also for the 50mm lens, and then take a third point for the 50mm lens at the known background distance this would give you the projected circle size, for your blurry background, this circle size should be the same for the 135 to have the same foreground vs background separation.

(i tried, i can't follow this ...particularly about taking a third point for the 50mm to get a projected circle size.

If you can't follow this, let me know and I will try to graph the concept.

Thanks so much.

[Lowell Goudge]

@ Taco

Hahah its true. Except I think it works best only when I have several primes. With several zooms, testing becomes difficult : The variable changes all the time. OO Look! A bird! (135mm zoom, bg at infinity...). Ooo! Look! A phone booth! (50mm zoom, bg at 2 feet away...)...
If i kind of get the math, I figure I can restrict my focal length choice to a specific size, and my background choices too. Also...sometimes its just timing. I am at my day work till dark...then at the weekend...I shoot in studio or do post processing. Too much working with camera to experiment with it! LOL.

@Lowell... I think I asked beyond my capability to comprehend Please walk me through I think what you are saying is that despite the relative difference between me - subject - bg, for the same field of view i should still get the advantage of the longer zoom because of the size of the aperture. I still have in my mind that there must be a limit to this, like a 1000 mm zoom vs 50mm , if trying to get the same fov.

Anyway, my confusion below, bolded :




Thanks so much.

Not sure this helps but here goes.

The vertical axis is distance, with the subject at perfect focus, i.e. circle of confusion = 0.

Lets assume you find that for the subject at 10 feet for a 50mm lens, the background is out of focus to the extent you wish, with the background 5 feet back, at say 15 feet, and the lens at F8. That has the projection of the circle of confusion at .04mm in diameter.

What you want to do now, is to shoot the same subject with the same field of view, using a 135mm lens. First of all, since the focal length is 2.7x that of the 50mm, the subject must be 27 feet away.

if you look at the set of lines coming out of the 27foot point on the Y axis, these represent the different apertures of the 135 lens. you want to pick an aperture that will give you the same circle of confusion projection of 0.04 mm so if you look at the grid, and the vertical grid line at 0.04 for circle of confusion, any line cutting this, will for the corresponding lens, give you the Aperture and distance to background for the same level of blurriness. Agail, if you take the line for F4, it shows the background just under 32 feet away or 5 feet behind the subject.

What I found interesting here is that if you keep the shooting aperture the same, you achieve the same degree of blurriness with the same subject to background distance, irrespective of focal length, as long as you always shoot the same field of view at the subject, i.e. maintain the same image magnification. Maybe that's why it is so easy to move back and shoot with a different focal length, there is no calculation. If F4 and a certain setback of background works, moving back proportional to change in focal length changes nothing (at least in the first order. Actually, longer focal lengths have slightly less DOF as a function of magnification, for any aperture. you see this playing with a DOF calculator, as I did for generating the data here

[Wheatfield]

Personally, I think the best way to get one's head around this is to go out and take pictures. This is a visual medium. I don't really think any amount of Googling DOF calculators or looking at graphs and charts is going to give a person a good visual idea of what happens to out of focus areas.

[Lowell Goudge]

Personally, I think the best way to get one's head around this is to go out and take pictures. This is a visual medium. I don't really think any amount of Googling DOF calculators or looking at graphs and charts is going to give a person a good visual idea of what happens to out of focus areas.

I don't disagree. it was simply an idle thought that something could be learned by looking at the results.

What would be interesting is to see if the DOF calculator shows the correct thinking, that blurriness is constant function of subject to background distance for any aperture. and the same subject magnification. that is a rule of thumb that if correct is worth learning, and it is probably faster playing with a calculator than going out and shooting, at least to see if there is a possible trend,

[D4rknezz]

Not sure this helps but here goes.

The vertical axis is distance, with the subject at perfect focus, i.e. circle of confusion = 0.

Lets assume you find that for the subject at 10 feet for a 50mm lens, the background is out of focus to the extent you wish, with the background 5 feet back, at say 15 feet, and the lens at F8. That has the projection of the circle of confusion at .04mm in diameter.

What you want to do now, is to shoot the same subject with the same field of view, using a 135mm lens. First of all, since the focal length is 2.7x that of the 50mm, the subject must be 27 feet away.

if you look at the set of lines coming out of the 27foot point on the Y axis, these represent the different apertures of the 135 lens. you want to pick an aperture that will give you the same circle of confusion projection of 0.04 mm so if you look at the grid, and the vertical grid line at 0.04 for circle of confusion, any line cutting this, will for the corresponding lens, give you the Aperture and distance to background for the same level of blurriness. Agail, if you take the line for F4, it shows the background just under 32 feet away or 5 feet behind the subject.

What I found interesting here is that if you keep the shooting aperture the same, you achieve the same degree of blurriness with the same subject to background distance, irrespective of focal length, as long as you always shoot the same field of view at the subject, i.e. maintain the same image magnification. Maybe that's why it is so easy to move back and shoot with a different focal length, there is no calculation. If F4 and a certain setback of background works, moving back proportional to change in focal length changes nothing (at least in the first order. Actually, longer focal lengths have slightly less DOF as a function of magnification, for any aperture. you see this playing with a DOF calculator, as I did for generating the data here

You must have meant F8 there...right?
So a couple questions :

- This contradict the earlier statement by taco about aperture size affecting the COC - and aperture size is a function of focal length/f-stop. Because 135/8 > 50/8 - and yet here you are saying given the same fov they will have the same COC. can we reconcile these two opinions?
- Which dof calculator gives you this information, please? The dof calculator i use only kind of say the areas that will stay sharp. No calculation on coc for a given focal length like what you just showed...


@Wheatfield - I agree to an extent, and in reality, its probably more practical to just test out. However, we do learn the usage of aperture , shutter speed, etc by theory. So i think its not too far fetched to be wanting to know the theory behind something a little more complex - especially if a formula can be derived out of it that allows me to look at a scene, punch in total distance between me and the background, and know which focal length i need to choose and where the subject needs to stand.

I may not get the formula I wanted, but even the question itself has given me more understanding on some stuff. Besides, if I know what to test specifically in theory, i will get better result than just going out there shooting. So far i havent been closer to my understanding of DOF than when i started shooting a year and a half ago - even though I've gotten some awesome (subjective ) pictures with dof in it. But...when i see them... how far was the subject? how far was the subject to the background? I forgot all these things. Could i have gotten a better result if I had used a different lens? No clue. I want this as a science, and not just luck. And when I come home with a 1000 shots and go through them, i rarely have the clue of what i did right at that particular moment....