Hyperfocal Distance - How to adjust it ?

[PatrickMe]

Hi everyone,
I have two wide-angle lenses and before I go to my question, let's say the f/stop is at f/11.

One is the Nikon DX 10-24mm

where its distance meter shows a max of only 0.5M and then the beautiful Infinity symbol.
Q1. my focal length stays at 10mm the hyperfocal distance shall be 0.45M, and at 12mm there goes the 0.65M? Where shall I go where the lens' max is 0.5M? Does it means I am restricted only to use the only and only focal length of 10mm?

The other is the NIkon FX 24-70mm

and its meter is 2M and also that Infinity symbol.
Q2. Focal length of its widest 24mm and the hyperfocal distance calculator said 1.75M and the same as the above 10-24mm, I can go further no more.

I hardly can believe my understanding in this matter but I really don't know what to do with it. I have gone thru the thread posted by Mark's Hyperfocal distance-turorial misleading of 15/11/2013 but I couldn't catch up with all those technical terms and knowledge therein.

Can anyone please enlighten me in this respect.

Your help and advice will be highly appreciated.

Thank you.

Hyperfocal is the numeric value of your DOF and very, very useful and quite simple to use. (Calculation is made according to recording media size/lens analog or digital)

As for me, I am shooting lens fall-off and never above sweet spot. f/stop to film plane gives you your focus point and in return it show your near & far (in sharp range vs blur) the only draw back is that you have to move your film plane (camera) to that exact point/distance, now the plus is you are using the full potential of the quality of your glass at specific f/setting and secondary effects in your image composition.

I am using a shooting scope to evaluate distances and a smartphone hyperfocal app, as you can see nothing difficult to put in practice.

Hope this makes sense to you

[Polar01]

Lukas as to you post #19, about print anything of size how does 15'x32" work for you, when I display the following image of course it is printed at 360ppi not stripped down to post here, the comment I get are how did you get it so sharp all the way from the very front to the back. They ask if I did a number of shots to focusing on different areas and then combining them into one image. I tell them no, however I did use 5 shots, 4 of which are used to create the spillage of water over the all to get the effect I wanted, all settings are the same for the complete set of images.
In case you are interested setting are f/4, 28mm, focus distance 21ft, near focus 10.3ft, far infinity

Cheers: Allan

[PatrickMe]

Hi Allan

This is a nice picture, don't get me wrong, but not relevant to hyperfocal!

[t6b9p]

I think hyperfocal distance has its place in digital photography but requires some modification. As stated several times, the original DoF tables/charts/lens markings are based on an old standard derived around World War II. The CoC was based on film printed at 8x10" and so is somewhat out of date (understatement). I also agree that if the horizon is important then the hyperfocal method falls short (literally).

By the way, whoever coined the term Circle of Confusion was not kidding. It wasn't until I started building my own DoF/CoC/Diffraction excel charts, that the whole inter-relationship of, defocus blur, diffraction blur, CoC requirement and DoF, became much clearer.

These days we tend to print bigger which requires a smaller CoC.
A smaller CoC is achieved in the image by minimizing defocus blur and diffraction blur.
Reducing defocus blur requires a higher f/stop, BUT a higher f/stop produces more diffraction.

A compromise must be made between high f/stop for less focus blur and low f/stop for less diffraction. The resulting "optimum aperture" unfortunately tends to be on the low side for landscapes, resulting in too narrow a DoF. The only real solutions to this appear to be, high resolution panoramas (relaxes the CoC requirement), tilt lenses (extends the DoF at the "optimum" aperture for dynamic situations) or focus stacking (extends the DoF at the "optimum" aperture for static situations).

There is one app that looks pretty good (if only it had an input for wavelength for IR users) http://www.georgedouvos.com/douvos/O...s_Stacker.html
The lack of wavelength input is what spurred me to start experimenting in excel.

[Polar01]

Well Patrick, I set up on a very small knoll, used f/4, then using that smartphone app, entered settings, it generated a focus distance of 21 feet, focused on something that distance away, so everything from 10.3ft out to infinity should be in focus. Is that not how hyperfocusing works? You will have noticed that the image is a ratio of near 1x2 not the usual 2x3 which means the image has been cropped, as I was on a small knoll, I positioned the camera downwards at a location approx. that 21ft away, then brought the camera up level, that gave me an area greater than I needed for the foreground to be in focus. As for the shots of the water spilling over the wall I did not have to change any setting from the original setup. As Lukas was talking about not being able to get a good image of a 16x24 print, well this is 15x32.
Again as this image was setup, taken using the hyperfocus method, then printed as a large image I believe that there is a relevant of the thread.

Cheer: Allan

[t6b9p]

This is a nice picture, don't get me wrong, but not relevant to hyperfocal!

Perhaps you could explain.

In case you are interested setting are f/4, 28mm, focus distance 21ft, near focus 10.3ft, far infinity

At f/4 you were obviously considering the avoidance of diffraction issues. I was expecting at this aperture that you would be shooting a 36MP DSLR and was surprised to find that it was 24MP. Did you select f/4 using a single pixel dimension for your CoC requirement?

[xpatUSA]

By the way, whoever coined the term Circle of Confusion was not kidding. It wasn't until I started building my own DoF/CoC/Diffraction excel charts, that the whole inter-relationship of, defocus blur, diffraction blur, CoC requirement and DoF, became much clearer.

Well said. There's even a school of thought that prefers to use multiples of pixel pitch rather than considering anything else. I have to admit that blindly following a rule-of-thumb invented by Zeiss et al goes against my nature. Or even looking up stuff on the ever-convenient internet, for that matter.

The internet tells me the CoC for my MFT camera is .015 or 0.016mm depending on who you believe. However, I calculated that, for viewing on my monitor NEC LCD1990SX, it's more like 0.030mm - double the DOF, whoopee!

[ajohnw]

Out of interest the DOF scales on older lenses are calculated on the basis of roughly 1/1400 of the diagonal of the format. Curiously this results in a max COC that is roughly twice the size of the one needed for the 10x8 pint viewed at 10in / 250mm which is based on the idea that the human eye has an angular resolution of 1 arc min. The doubling is down to lens aberration and what that means is that the lens manufacturers will strive to maintain this resolution. Actually better European manufacturers used a much finer limit on the DOF scales.

On top of this some lenses are better than others and most lenses have significant differences in performance across the frame.

That was film. Digital is rather different. For a start interpolation software is used to create a colour image in the first place. There are various styles of this software and it''s difficult to judge what effect it will have on circles of confusion. Then there is post processing sharpening in particular. Then there is what I call jpg type software. Providing the detail in the shot is captured well there won't be any problem scaling it up. Problems arise when the COC is so big that it can't capture all of the detail in the shot.

There is no harm in using DOF calculators but in practice people may find they have a better or worse DOF than expected even before post processing. Actually from a lot of film shots in the past I don't really agree with them. What if I wanted to produce a 20x16. I didn't find the scales on the lenses much use for that and the distance scales were often out as well. That's why cameras have view finders and focusing aids.

Somebody mention 1 pixel resolution. A 4 pixel block is used to generate a colour image in a RGGB arrangement. Some people reckon that 2 pixels is the max colour resolution of a digital camera but for a variety of reason that is unlikely to be met in practice. In theory they can photograph a 2 pixel black and white chequered board as it appears on the sensor accurately, Maybe some one would like to try it and post a shot.

Diffraction effects also mentioned are another thorny area complicated by how camera lenses tend to be designed and the digital effects. One of the problems is that contrast at the diffraction limit is very low, so low that it's generally not considered to be of much use to a photographer unless their camera is pointing at stars. There is a very good page on this subject here where the contrast for BLACK and WHITE are shown in 3D at Rayleigh's limit. The contrast level of most interest to photographers is 50% not the around 8 at the limit.

http://www.luminous-landscape.com/tu...solution.shtml

Some people look at resolution tests and suspect that cameras recognise that they have taken a shot of a test chart and behave accordingly. They might just do that with jpg's or even raw processing software but on the other hand they should recognise it automatically as they are black and white. If they were mid grey and white a simplistic view is that at the resolution where black and white has 50% contrast they would be at 1/4 grey. Actually the white gets less white and the grey gets less grey. A considerable difference. 1/8th grey 1/16 etc. The depth of field available at these resolutions is also rather small.

John
-

[Polar01]

I am afraid I have made an error, when I looked at the data for the image I posted, I had stated that it was shot at f/4 whereas it was shot at f/14, that difference means that it is likely that I focused on something about 10 to 12 feet away asopposed to something 21ft away as stated above in post #25. Again my mistake, however that is not that important as the shot was setup based on using the hyperfocus method to get the greatest amount of the image in focus.

Cheers: Allan

[muls]

Hi Patrick
Hope the below link is of some help.
http://www.dofmaster.com/dofjs.html
Cheers:Murali

[ajohnw]

The basic calculation for hyperfocal distance is (F^2/NC)+F where F is the focal length, N is the aperture and C is the required diameter of the maximum circle of confusion. So that COC will be caused by objects at infinity when the lens is focused at this distance The near limit is very close to 1/2 of that. The measurements need to be in the same units so if some one wanted to use feet the COC would have to be converted to feet or worked out in mm and converted etc. The fact that the near limit is 1/2 the hyperfocal distance is why some people focus 1/3 of the way into a shot and choose an aperture. As I seem to be under estimating DOF on micro 4/3 I am going to have to rethink some of what I do mostly at shorter distances.

Dofmaster uses a COC of 0.015mm for 4/3, 0.023mm for full frame and 0.019mm for APS. The 0.023mm is what Leica used to use. My understanding is that Canon for instance used the larger number I mentioned for lens markings when they had them on film lenses. I'm not sure if they still provide depth of field info on their web site. I haven't checked if dofmaster uses a fraction of the sensor size but it should.

The problem with doing the calc in the head is that NC is always a small number.

Going back to the 1/3 way in it's fairly easy to see why hyperfocal distance is a bit of a problem when distances aren't accurately measured. Using Dofmaster I set up F16, micro 4/3, 45mm FL. The hyperfocal distance comes up at 8.48m. Near is 4.24m. If I put a distance of 8.45 in the top section near comes out at 4.2m and far comes out at roughly 2.2km. The near is more or less the same as given in the hyperfocal distance section but there is a big difference in the far - infinity is a funny number anyway. All I need to to achieve this result is to be 30mm out in gauging the distance that I focus too. If I happen to focus 0.25m short the far limit comes down to 275m. These errors are caused by under estimating the distance. If I over estimate the distance far stops at infinity and near is no longer 1/2 the distance It's a bit less. The far limit is past infinity. If I focus on something 10km away the near limit is roughly 8.5m, 5km roughly the same, at 500m it's down to 8.3m, 10m 4.6m.

Taking another example an m 4/3 8mm lens at F8. The hyperfocal distance is 0.54m, near 0.27m and far infinite. If I still focus at 500m near goes to 0.53m, 100m 0.53, 10m 0.51, 1m 0.35, 500mm. Say I go too close and focus at 500mm whoops near is 0.26m but far is down to 6.45m just for being 40mm out. Going to F9 cures that - another touchy area.

I'd hope people will realise that these figures mean that the calculations are difficult to use in practice. Where they can be useful is deliberately blurring back grounds to make subjects stand out. Final size does have to be accounted for then and a tape measure is needed. The idea is to leave the background recognisable but blurred by a controlled amount. This mimics human vision and the eye will automatically go to the subject which in this case is probably best placed near central. The reason final size comes into it relates to say it's calced for 10x8, small leaves in the background and a viewing distance of 250mm - that's unlikely really but if the shot finished up in a magazine at 5x4in it wont be viewed at 125mm/5in. Basically it's a bit of a judgement call. Maybe we should switch to A4,5 etc but a lot of this stems from sizes around 10x8in.

: I have double checked these numbers but do your own as a check. If you read CinC's tutorial I vaguely remember noting a comment that calculated numbers for this area shouldn't be taken too much to heart. Basically it's best to go out and use the camera. It doesn't matter what it's pointed at to find out what it does.

John
-