DoF - different formats?

[q.g._de_bakker]

Andrew is correct. We can compare DoF between formats. And as far as the different theories and calculations go, we can get an approximate idea about DoF's dimensions using them.

<br><br>But though you can compare DoF between multiple formats (ignoring differences in assumptions made when different formats are considered. And assuming that all of these assumptions cancel out anyway. Except, that is, the assumption that the assumptions will all cancel out, disappear, itself of course ;-) ), DoF itself is still elusive and illusory, just because of those assumptions.<br>We should not forget that, and use DoF as far as we can and is usefull. But we should also be aware of the tricks it can play on us (or rather the tricks we play on ourselves when we assume DoF is some well defined and easy to grasp thing.)

[James G. Dainis]

When I told a friend that I normally shoot with my 8x10 view camera lens at f/64 He replied, "Wow, you must get some really great depth of field". Not so. DOF for f/64 on 8x10 is about equal to about f/16 on 35mm.

[ty_mickan]

<p>Hyperfocal distance is an outdated technique for landscape use (zone focusing is very much useful technique),and it will usually yield mediocre results. it is a formula that was determined in the 20's by the company now known as Leica. It may have provided adequate results back then, with older uncoated spherical lenses, and 80 year old emulsions, but today it just doesn't cut it. We now use the latest emulsions and digital sensors, and normally now we discard the enlarger as well. All this adds up to a Circle of Confusion size that needs to be alot smaller than the 'international standard' of 1/30mm (for small format, and 1/1500 of the diagonal film/sensor size for anything larger). <em><strong>Usually</strong> </em>, for landscapes, infinity focusing will provide the best compromise. And there is always going to be a compromise. A lens size of around 4mm (give or take 1mm) will usually give the best results, whilst taming diffraction limited spot size. This applies to small, medium or large format. Actually the easiest way to decide on aperture is to assess a scene, decide what the smallest item is in the scene that you want to render acceptably sharp, and adjust your aperture size to the same size. at inifinity, everything that is the size of the aperture will be rendered with the same resolution, diffraction aside. To simply focus half way between the closest and farthest points of required 'sharpness' is not the answer either.<br>

Hyperfocal Scene<br /><img src="http://farm3.static.flickr.com/2752/4122610111_1513189ef6_o.jpg" alt="" width="700" height="466" /><br>

Infinity Scene<br /><img src="http://farm3.static.flickr.com/2677/4122611671_6d05259c95_o.jpg" alt="" width="700" height="466" /><br>

Hyperfocal Foreground<br /><img src="http://farm3.static.flickr.com/2710/4123380632_1377fe51dc_o.jpg" alt="" width="700" height="466" /><br>

Infinity Foreground<br /><img src="http://farm3.static.flickr.com/2704/4122610905_406fea7933_o.jpg" alt="" width="700" height="466" /><br>

Hyperfocal Mid<br /><img src="http://farm3.static.flickr.com/2797/4122609825_30c10a2ce8_o.jpg" alt="" width="700" height="466" /><br>

Infinity Mid<br /><img src="http://farm3.static.flickr.com/2502/4123382452_24ba891684_o.jpg" alt="" width="700" height="466" /><br>

Hyperfocal Far<br /><img src="http://farm3.static.flickr.com/2783/4122609143_db42ef2b3f_o.jpg" alt="" width="700" height="466" /><br>

Infinity Far<br /><img src="http://farm3.static.flickr.com/2663/4122610445_caa34f0a7e_o.jpg" alt="" width="700" height="466" /><br>

It is hard to pickup on the original shots, and I haven't cropped in the far background, but the silo in either shot is rendered the same. Even though you may think that the inifity focus may render it sharper, diffraction has limited the spot size. The lens used here was a Leica-M Summilux 1.4/35mm ASPH @ F9.5, giving me a lens size of around 3.7mm. If I had have used F5.6, the diffraction limited spot size would have been smaller, however the foreground detail would have slightly suffered. The foreground was more important in this shot than the silo. I perhaps therefore could have used f11, which would have improved the foreground at the expense of the silo. If this were a critical shot, I may taken one shot at each of the two apertures.</p>

[James G. Dainis]

Mr. Mickan,

 

The differences shown are amazing. I never would have thought that infinity focusing would be sharper over all distances than hperfocal focusing.. Perhaps the word "acceptable focus" often used with hyperfocal distance should be clearly defined. The hyperfocal focus examples you show are far more not as acceptable as the infinity focus examples.

[q.g._de_bakker]

A simple but very good rule to follow is to forget about hyperfocal focussing. Focus on whatever needs to be sharp, the subject of your photo. Use the aperture to control how blurry the rest will appear.<br>Simply remember that there will only be one thing really sharp, and that is where you have put focus on. The rest will not be.<br>Mathematic formulae do not know what your subject is, what has to be sharp. So do not let mathematics decide what in your scene should be sharp and what not.<br><br>James, the infinity focus crops of Ty0s examples aren't all sharper than the hyperfocal crops. The hyperfocal distance used apparently was pretty close, the infinity distance (obviously) is not. So the foreground in the hyperfocal image is sharper than in the infinity image.<br>Given the particular subject, hyperfocal turned out better than infinity focus. Better still would have been to focus on the foreground gravestones/monuments, though using hyperfocal distance apparently was not far off.

[AlanKlein]

<p>But the subjects, especially in a landscape photo, are often close and distant. So which to focus on? I look at the range I need in focus far and distant. The figure the aperture that would cover that range then lower it by one stop for good measure. If I follow your advice, what subject would I focus on?</p>

[lex_jenkins]

<blockquote>

<p>"Hyperfocal distance is an outdated technique for landscape use (zone focusing is very much useful technique),and it will usually yield mediocre results. it is a formula that was determined in the 20's..."</p>

</blockquote>

<p>The research and math date back to the 19th century and the math and physics haven't changed. If we're seeing significant differences between results with our equipment and the expected results based on DOF charts and the hyperfocal setting, the most likely reason is the physical limitations of manufacturing tolerances in lenses.<br>

<br>

Wide angle lenses, particularly for miniature format 35mm film and smaller format cameras, often have a short focus rotation or "throw", from minimum focus to infinity. With autofocus wide angle lenses the throw can be extremely short, sacrificing hyperfocal accuracy in favor of quicker AF performance. And the throw tightens significantly toward infinity. Typically 95% of the entire focus throw rotation occurs from minimum focus to 10 ft. From 10 ft to infinity usually takes up less than 5% of the remaining rotation. With a 50mm "normal" lens the rotation from around 10 ft to infinity may range from 1/2" to an inch - plenty of room for precise alignment. But with wide angle lenses the remaining focus throw from 10 ft to infinity may be only 1/4" or 1/8" - not enough room for precise alignment. And that's with manual focus lenses. With AF wide angle lenses the rotation from 10 ft to infinity may be even smaller.<br>

<br>

Add to that any error, however slight, in the manufacturer's DOF scale, and the problem is compounded.<br>

<br>

So if we're seeing "better" results setting a wide angle lens to infinity rather than the hyperfocal setting, it may be due to a combination of a slightly inaccurate DOF scale and the extremely short rotation toward infinity.<br>

<br>

Often when I'm trying to maximum DOF with wide angle lenses I'll do the same thing Ty described - I'll just rotate the focus to the infinity stop and set the aperture at f/8 to f/16. It's repeatable and consistent, and easier in dim lighting, especially with the camera mounted on a tripod at eye level where I'd need to change my alignment to see the DOF scale on the lens. I'm not going to do that after I've carefully leveled the camera, checked it with a bubble level, etc. So I'll rack the lens to infinity, stop down and usually find the results acceptable unless the primary subject is within 50 yards or so of my position.<br>

<br>

Acceptable sharpness always involves a compromise between focus, diffraction, viewing distance and print/reproduction size on the monitor, slide projection screen, etc.<br>

<br>

And the same challenge applies to infrared photography. The IR mark on the lens isn't always accurate enough for critical focus, especially shooting wide open and up close. It's a rough guideline but demands some tests to determine the precise IR shift.</p>

[q.g._de_bakker]

The math and the physics may not have changed, but the quality of both lenses and recording media have. Remember that DoF depends on the difference between sharp and unsharp being not very noticeable, and that that not only depends on how unsharp the unsharp bits are, but also on how sharp he sharp bits are.<br>Zeiss publiced a short article in an issue of their Camera Lens News that concluded that though the research (from the 1920s, consisting of having groups of people judge a large amount of small prints) that put the assumptions into the mathematics was seriously outdated, the way people use cameras and focus makes even the inaccurate scales on lenses and old formulae more than precise enough. Sadly. Why make good lenses, good film and sensors, if people waste the quality they offer? Use tripods and take care focusing (and that includes not using hyperfocal distances, i.e. not letting maths decide what should be in focus), or you waste quite a lot!<br><br>Even a landscape should have a point of interest, a subject, a feature that is the focus of the image. Focus on that. A picture that lacks focus, that for that reason needs to be sharp from 2 inches to infinity, leaves the viewer wondering what he or she is supposed to look at, i.e. is not a strong picture.

[AlanKlein]

<blockquote>

<p>Even a landscape should have a point of interest, a subject, a feature that is the focus of the image. Focus on that. A picture that lacks focus, that for that reason needs to be sharp from 2 inches to infinity, leaves the viewer wondering what he or she is supposed to look at, i.e. is not a strong picture.</p>

</blockquote>

<p>That's an aesthetic opinion that doesn't usually apply for landscapes taken with medium and wide angle lenses as shown below. There are often close , medium and far elements that produce a better picture when all are in focus. A quick zone approach to maximize the range and get pretty close to hyper-focal is to reduce the aperture and then focus about 1/3 up the picture, not 1/2 way as was suggested in someone else's post. The 1/3 works for any distance range you want to be in focus. </p>

<p><a title="Crooked fence by Alan Klein, on Flickr" href=" src="https://farm4.staticflickr.com/3690/11021823226_c91ca5eac0_z.jpg" alt="Crooked fence" width="640" height="523" /></a></p>

[q.g._de_bakker]

I think it does apply, Alan.<br>In the example you posted i see a fence a field and some trees. I don't know what the subject is, what i am supposed to see, what i am shown. Why am i looking at that picture?<br>The fence and trees and intermediate field touch, are not separated, making it as a pure visual (i.e. assuming no meaningful subject) thing not very clear, more a jumble of lines and similar tones. Whether that makes a pleasing picture depends on taste, i guess.<br>The image would be much stronger if it did have a focus, it it said 'look at this fence!' and showed a fence with a non-distinct backdrop.<br><br>That concerns having a photo of something vs a photo in which something (also) is present. The thing about DoF and hyperfocal, letting mathematics decide that neither fence nor trees, but some dried out blade of grass in the middle distance is the thing that deserves to be in focus is another. But a lot has been said about that already, so i leave it at that.<br>Just a quick note about the thirds or halves ratio of distribution of DoF: that ratio is not fixed, it changes with focussing distance. Close up it tends towards an equal distribution, in the middle distances it is more like thirds, and at the hyperfocal distance, there is no such ratio at all.

[AlanKlein]

<p>QG: The 1/3 is a rough estimate for quick focusing and hyperfocal without calculations. However, there are always DOF considerations, leaving aside the aesthetics since people can have differing views on that.</p>

<p>Let's stick with the technical about DOF. Let's assume for the moment, that I only want the fence to be in focus, as you suggested, with the background not in focus. Since the fence starts at three feet close and runs about 15 feet back, you still don't have an exact point of focus. You still have to figure DOF from the front on the left to the back of the fence on the right. You still have to consider what f stop will give you that range and which will leave the background out of focus. And then set the focus point for the DOF to catch the whole fence.</p>

<p>That shot was MF 6x7 with a 50mm lens. So these are the calculations for just getting the fence in focus.<br /> Focal length (mm): 50mm <br /> Selected f-stop f/22</p>

<p>Subject distance 4.5 ft</p>

<p>Depth of field<br /> Near limit 2.62 ft<br /> Far limit 15.9 ft<br /> Total 13.3 ft</p>

<p>In front of subject 1.9 ft (14%)<br /> Behind subject 11.4 ft (86%)</p>

<p>Hyperfocal distance 6.21 ft<br /> Circle of confusion 0.06 mm<br /> Calculator: http://www.dofmaster.com/dofjs.html</p>

<p>It's interesting that there's only about 1 3/4 feet between a focus point that only gets the fence in focus (4.5 ft) to the hyperfocal distance of 6.21 ft that will get everything in focus. There's around a 1/7 ratio of what's in focus before and after the 4.5 ft focus point.</p>