Bokeh when approaching hyperfocal distance

Is there a rule of thumb for the extent of blurring you get out of the background (at a given distance) as your subject approaches the hyperfocal distance.

For example, if I need to be 150 feet away from am subject to allow the subject to fill the frame, and I am shooting at f/4.0, the range in focus is 115-214 feet, how far away will the background need to be to get a nice creamy background? 1000ft, 2000ft, 5000ft?

Now repeat this thought experiment for 300ft, 400ft, and 500ft. With 500ft being the hyperfocal distance everything should be in focus, no matter how far away the background is.

Heya,

Depends on focal length. F4 doesn't tell us anything if it's at 16mm or 600mm, as it changes depth of field to two extremes.

Your thought experiment, I'm assuming, is with a telephoto where depth of field is going to be very limited, because at wider angles, depth of field would be to infinity/horizon with a wide angle at F4 at 150 feet.

Very best,

For your original question, where your subject is at or very close to the HFD, the simple answer is that you will NEVER get the background to look blurred. Hyperfocal distance is the focus distance that is required to give you a depth of field that runs from HFD/2 to infinity. Remember that the HFD varies based on both the focal length of the lens, and the aperture size. Once you are focused at or beyond the HFD then the background will never seem to be blurred.

When it comes to the specific distances that you are mentioning then it will depend on how close to the HFD actually is to the distance you are focused at, and also to some degree what constitutes infinity. If you are shooting with a 50mm lens a subject at 150 feet can pretty much constitute being at infinity already, let alone the HFD. Shoot with a 500mm lens, and focus at 150' and even down at f/8 you are going to pretty much obliterate a far distant background. I have the Sigma 150-600 C, and at 600mm and f/6.3, the maximum aperture, if I focus at 150' and then try to have the lens AF on a subject away at the horizon, the AF will not operate, since I have the do not AF if focus not possible option set. The background is so defocused that the AF cannot measure the change.

Alan

There is a formula that computes 'quantity of blur' (not 'bokeh' per se, which is the 'quality of blur'...but let's not get overly distracted by the fact that there is so much improper common usage of the word). But the quantity has a 'distance to background' dependency which makes any rule of thumb not generally possible. One can generalize that when the distance is somewhat 'far away', the quantity of blur is related to the 'aperture diameter' (FL/ fstop = diameter)

This graph exemplifies the issue of quantifying blur as a simple rule of thumb

MalVeauX wrote in post #18311146
Heya,

Depends on focal length. F4 doesn't tell us anything if it's at 16mm or 600mm, as it changes depth of field to two extremes.

Darn, I meant to say at 135mm focal length. Sorry I forgot that part of the question.

s1a1om wrote in post #18311259
Darn, I meant to say at 135mm focal length. Sorry I forgot that part of the question.

135mm f/4, focused on an object at 150', frames an area 26' wide (about 8m) at that distance. DOF is 120'-600'...but the hyperfocal distance is at 598'. If your focus is at hyperfocal distance, your subject at 150' is NOT in focus! So your question posed does not all tie together. Nevertheless, here is a graph of the background blur...

Wilt wrote in post #18311193
But the quantity has a 'distance to background' dependency which makes any rule of thumb not generally possible.

It looks like you have some sort of program for calculating this. Do you have a reference for the equations? I wouldn't think it would be that difficult to program up something similar for my own reference.

I thought the the quantity of blur was a function of focal length, aperture, distance to focus point, and distance to background?

Wilt wrote in post #18311270
135mm f/4, focused on an object at 150', frames an area 26' wide (about 8m) at that distance. [/IMG]

Specifically, the reason for the question comes from thinking about air to air photography of light aircraft (wing span ~28 ft). Most photos you see online seem to be taken with smaller apertures (>f/8) so everything, including the background is fairly sharp. I think part of the reason for this is that the photos are usually taken in bright sunlight so it's necessary to use a narrower aperture to get a proper exposure around 1/60th (for good prop blur).

But, I was wondering if it would be possible to get a different look/feel to the air to air shots by using an ND filter to allow for wider apertures. Based on the charts you posted, it seems like the function asymptotically approaches a value at large distance to background ratios. At 135 mm and f/4, the asymptote appears to be at .42, which based on the other chart you posted with 50mm and f/1.4 (something I can relate to), it doesn't seem like enough to really get a nice blur. Obviously f/2.8 would be have a larger effect, but I'm not sure if that would be enough either.

Wilt wrote in post #18311270
135mm f/4, focused on an object at 150', frames an area 26' wide (about 8m) at that distance. DOF is 120'-600'...but the hyperfocal distance is at 598'. If your focus is at hyperfocal distance, your subject at 150' is NOT in focus! So your question posed does not all tie together.

I did a bad job posing the question. I do understand what you note above. My question was assuming that quantity of blur was a function of both distance to subject and distance to background, I would think that as the subject approaches the hyperfocal distance the distance to background would need to be larger to get a similar quantity of blur.

ie. the blur 10 ft behind a subject that is 10 ft from the camera would not be the same as the blur 10 ft behind a subject that is 100 ft from the camera.

s1a1om wrote in post #18311393
It looks like you have some sort of program for calculating this. Do you have a reference for the equations? I wouldn't think it would be that difficult to program up something similar for my own reference.

I thought the the quantity of blur was a function of focal length, aperture, distance to focus point, and distance to background?

Specifically, the reason for the question comes from thinking about air to air photography of light aircraft (wing span ~28 ft). Most photos you see online seem to be taken with smaller apertures (>f/8) so everything, including the background is fairly sharp. I think part of the reason for this is that the photos are usually taken in bright sunlight so it's necessary to use a narrower aperture to get a proper exposure around 1/60th (for good prop blur).

But, I was wondering if it would be possible to get a different look/feel to the air to air shots by using an ND filter to allow for wider apertures. Based on the charts you posted, it seems like the function asymptotically approaches a value at large distance to background ratios. At 135 mm and f/4, the asymptote appears to be at .42, which based on the other chart you posted with 50mm and f/1.4 (something I can relate to), it doesn't seem like enough to really get a nice blur. Obviously f/2.8 would be have a larger effect, but I'm not sure if that would be enough either.

I did a bad job posing the question. I do understand what you note above. My question was assuming that quantity of blur was a function of both distance to subject and distance to background, I would think that as the subject approaches the hyperfocal distance the distance to background would need to be larger to get a similar quantity of blur.

ie. the blur 10 ft behind a subject that is 10 ft from the camera would not be the same as the blur 10 ft behind a subject that is 100 ft from the camera.

I do not think of an algebraic equation with four variables to be a 'rule of thumb' which average folks can easily memorize and compute in their heads, unless I am dealing with a bunch of mathematicians or physicists!

The graphing program that I used... http://asklens.com ….8-on-a-0.9m-wide-subject


Assuming you shot with 200mm lens at f/8 vs. 200mm lens at f/2 with a -4EV ND filter, hyperfocal at f/2 is 800m away, at which your framed area is is 96m x 144m wide, so unless your target was a small airplane it would not fill much of the frame! That is why I stated that your question is not well thought out...one would NOT ever focus 'approaching hyperfocal' if shooting with 200mm f/2 lens and a -4ND filter to permit that wide open aperture...if the alternative were to use f/2 vs f/8 at hyperfocal distance, the DOF zones are 400m-Infinity vs. 160m-Infinty, !

To humor your question and assume you are shooting a small airplane, shooting at an object 800m meters away frames an area 144m wide (the graph shows 14m only because the program has too small of a window for three digits), the DOF is 30m with 200mm at f/2. Here is what the blur quantifies, vs. the blur at f/8 when focused at 110m...

Something like a waterfront boardwalk that curves around a bay or similar can be a great place to play around and better understand the depth of field and out of focus blurring qualities of your lenses. (Bonus points to the site if it happens to have lights as well so that you can study how point light sources are distorted and changed as they go out of focus.) While reading and math are good starting points to learn and get your head wrapped around things, eventually you are going to want to sit down in the real word and take a look at what is actually happening in the photos taken with the lens and how out of focus elements are rendered. Set your camera down on a tripod if possible, and then slowly work your way through the scene changing your focus a little bit with each shot.

Assuming you're working with digital and not film, then it is rather trivial to sit there and take hundreds of photos in short order with a slight nudge of your lens's focus ring while the auto focus is turned off. Make sure your exposure is decent, then run through the same scene with roughly evenly spaced focus differences all the way from one end of the lens's focus range to the other. Take the results home, load them up and spend time studying them on a large screen. (As a bonus, you also get to study how much 'focus breathing' your lens does, which is where a lens may change focal length based on how near/far it is focused.)

A key factor (for anyone really new who might be reading this thread) to remember when getting your head around optics and depth of field is that there is not really a volume of space that is 'in focus', and one that is 'out of focus' (or two if you account for in front/behind the in focus), but rather a thin plane that is 'in focus', and a long gradient running toward you and away from that plane of "Slightly less in focus".

If you had a mathematically perfect lens mounted to a mathematically perfect sensor that allowed infinite resolution, then there would be a single thin plane somewhere within the scene where any point of light would fall as a perfectly focused point of light on the sensor. Even just a billionth of a micron forward or back and the point of light would be distorted and 'out of focus', and the farther from this plane the point in the scene is, the more distorted the point becomes on the sensor.

The "Depth of field" concept in photography works because our lenses and sensors aren't mathematically perfect, and we're not looking at the information on an accuracy scale anywhere close to that. The error of a few microns becomes too small to tell apart from the 'true' in focus point. "Depth of field" calculators are more about "Reasonably in focus" vs "the error is becoming visible", but the point at which the error "becomes visible" is rather fuzzy and relies on your final print. (Pun only vaguely intended.)


As far as artistic use of bokeh and how far behind the subject you need the background to be for it to look nice? Well, that depends on the nature of the background - what kind of highlights and the shape of the lighting you have in the scene - and the lens. "The more the better" is often a decent rule of thumb, but as in all things photographic "It depends" comes into play. Separation of at least half the camera to subject distance is probably not a terrible place to start.

And don't forget: A boring background without texture/contrast is probably going to give you boring looking out of focus elements in an image.


Another important thing to remember: You do not need the focus plane of the image to fall perfectly on the primary subject to have the subject in focus... With care you can focus in front of the subject while keeping them within an 'acceptable range of focus', and let the farther background be more so out of focus than it otherwise would have. It does however mean that you will limit the enlargement you can make while still hiding the focus error in the subject.

If you are happy with the option of post processing the shot, and aren't restricted from using editing to get your final result, then you also can potentially use a workflow that involves blending multiple exposures. One of my friends often does shots where they will use rapid burst fire while pulling focus back from their subject. (Something which takes some skill and practice with to maintain reasonable sharpness and not introduce excessive motion blur or misalignment.) They will then go back through all the photos from the burst of a given 'shot', and take the first photo with the sharpest subject and flip through them till they have the focus distance that gave the best background, then blend the two in Photoshop.
- Enlarging the in-focus-subject shot by a slight amount can make it easier to hide the out-of-focus-subject behind it.

I thought I would try to quantify the size of the blur for a reasonable sized air to air shot of a light aircraft, using a 135mm lens, on a 35mm format body. Given the average size of a single engine light aircraft I allowed for framing an area of 15m by 10m, which would require shooting at 56.3m (185') which I think is actually a very reasonable distance for this type of work. I am also going to assume that the subject aircraft is at an altitude of 1000m (3280') AGL.

With those parameters the blur disk for a perfect, infinitely small, point source will be 0.425% of the image width at f/2, and 0.11% at f/8, based on using the website Wilt referenced. In terms of pixels, if we assume a sensor with 5700 pixels on the long edge (21.6 MP) the blur disks will be (rounded up to the next full pixel) 25 pixels at f/2 and 7 pixels at f/8. In a 18×12 print those blur circles would become 0.077" at f/2 and a mere 0.020" at f/8. In real terms, when dealing with a small point with actual size, the blur will make the feature seem that much larger in total, since this is dealing with diameters, so only half the diameter is added to each side of the actual feature. Also remember that the area each side of a feature will be blurred into each other, so the interaction of features can become complex, but I guess here we are moving from blur sizes, into the realms of true Bokeh, and the subjective evaluation of the blur.

Most folks will probably still be viewing an 18×12 at about 12", but I cannot recall the correct visual acuity limits for the viewer regarding the perceived DoF. I don't doubt the values that Wilt has quoted for the HFD for a 135mm lens, but you are not particularly close to that when shooting at these sensible looking distances.

Alan