Depth of field - is it linear?

[William W]

Following on . . .

. . . I was taught many years ago was that a good, safe approximation for minimum viewing distance is the diagonal of the print size.

Noted the word "minimum" (my underlined and bold).

Cine & Television editing, tethered screens etc use a typical distance of about 1.5 ~ 2 times the screen's diagonal.

The last studio we had we edited / post produced our Still Images at viewing about 1.8 the screen's diagonal: Clients viewed the final product on big screens at a distance of 1.5 the diagonal.

Previously (in Print and Film days) we would show a 24 x 20 inch sample on an easel viewed at about 3' (36") - that's really close to the diagonal but for a good print, that viewing distance can have great impact.

WW

[William W]

. . . If so, why don’t DOF tables or calculators (e.g. Paulo Renato’s) incorporate these variables?

I have not looked closely at Paulo Renato’s calculators; but I assume they reflect DoF Tables.

The answer to your question is firstly historic and secondly implicit in the fact that the DoF Chart/Tables are charts/tables for a SPECIFIC viewing condition.

Those conditions are defined by the viewing distance (about arm's length) and the dimensions of the (un-cropped) print (about 10x8 inches).

The logic is: having the Image Size and the Viewing Distance a STANDARD then the resultant Depth of Fields (note PLURAL) will be useful as COMPARATIVE MEASUREMENTS

The salient point is, DoF Charts were designed to allow the Artist/Technician access to COMPARATIVE MEASUREMENTS.

WW

PS: We are getting more rain in Sid Knee than you are in Dar Win.

[Pippan]

Yes but we've had 1392mm so far this calendar year in Dar Win.

PS. Thanks for the DOF info. I hadn't even considered the effect of cropping.

[William W]

I think I found what I was looking for in this - http://paulorenato.com/index.php/151

Maybe what you were looking for: but it is a flawed document in its first general conclusion.

viz (my bold and underline for emphasis of erroneous commentary):

First some definitions are in order. In photography, depth-of-field (DOF) is loosely defined as the region in focus for a given photograph. The main factors affecting DOF are the size of the camera sensor, the lens focal length and aperture, and the distance to the subject. Here's how this works. With a given camera, if what you want is to get as shallow DOF as possible (i.e. a blurred background as commonly seen in portraits) then you want to:

- Get closer to your subject

- Use a wider aperture (e.g. f/2 is better than f/8)

- Use a long lens (e.g. 100 mm is better than 24mm)

(ibid)

That is an incorrect conclusion for all general shooting distances for Portraiture.

Focal Length does not matter.

Provided the Camera Format; the Aperture and the FRAMING of the SHOT remain constant then the Depth of Field will also remain constant.

Note that Depth of Field is related to the 'quality of' BACKGROUND BLUR, (aka 'Bokeh') but they are NOT the same thing.

WW

[William W]

Off Topic - I beg indulgence:

Yes but we've had 1392mm so far this calendar year in Dar Win.

PS. Thanks for the DOF info. I hadn't even considered the effect of cropping.

yeah.. I was just talking this fortnight... it's driving me mental

[William W]

. . .Thanks for the DOF info. I hadn't even considered the effect of cropping.

You are welcome: DoF is one of my favourite pastimes.

WW

[Pippan]

Sooo, can I assume framing of the final image also takes into account any cropping?

[george013]

Something I’ve been wondering. Given that, as I understand it, truly sharp focus is only at an infinitesimally thin plane, are the generally proffered limits of ‘sharpness’ in DOF tables and calculators really only ‘acceptable’ sharpness, and do they assume some standard image magnification (say 8” x 10”) and viewing distance (say 2 feet). Does a larger print (say 24” x 36”) or a closer viewing distance result in a thinner DOF at the same camera settings, since even slightly out of focus areas have been magnified to the viewer and would be more visibly OOF? If so, why don’t DOF tables or calculators (e.g. Paulo Renato’s) incorporate these variables?

Those variables are incorporated in the coc. By choosing a sensor size you set the coc. This is supposed to be valuable for a A4 print viewed at arm length, or something like that.

The background is that one can see a circle of 0.36mm at a distance of 1 meter as a point. I hope I have those figures right. A FF sensor is 24x 36mm. When printed on a A4 it must be enlarged about 12 times. Calculating back it means that the maximal circle on the sensor is 0.36/12 0.03mm. That's the coc of a FF camera. A DX will be 1.5 times smaller, so the enlargment for a A4 print will be 1.5 times higher. It becomes 0.02mm.

If you crop, then you must enlarge the picture more to get a A4 print, so the coc will be smaller.

Dave,

I found some issues concerning my problem on the Apache forums. So I'll dig in it. I know macro's are not compatible. I don't know what he used.
I think he used an Office 97 spreadsheet, and that I've somewhere.

George

[dje]

Something I’ve been wondering. Given that, as I understand it, truly sharp focus is only at an infinitesimally thin plane, are the generally proffered limits of ‘sharpness’ in DOF tables and calculators really only ‘acceptable’ sharpness, and do they assume some standard image magnification (say 8” x 10”) and viewing distance (say 2 feet). Does a larger print (say 24” x 36”) or a closer viewing distance result in a thinner DOF at the same camera settings, since even slightly out of focus areas have been magnified to the viewer and would be more visibly OOF? If so, why don’t DOF tables or calculators (e.g. Paulo Renato’s) incorporate these variables?

Hi Greg

The CiC DOF calculator does include these variables. You can find it here. At the top rigt hand side of the calculator there is a button called "Show Advanced". If you click on this, it opens up more input fields viz Max Print Dimension, Viewing Distance and Eyesight(quality).

Dave

[Pippan]

Thanks George, that makes sense to me.

[Pippan]

Thanks Dave, I hadn't noticed the 'show advanced' button on that page. That's brilliant!

[dem]

Sooo, can I assume framing of the final image also takes into account any cropping?

I am not quite sure what you mean. Framing in this context means "field of view" - what frame you cover at focus distance. For example a 50 cm by 75 cm field of view is big enough for a head-and-shoulders portrait but is too small for a full body shot. Cropping means taking a part of the original image in post processing and enlarging it to fill the screen or the paper. This enlargement is linear - when you view a cropped image at its full size, all its imperfections will be more noticeable and its DoF will be (almost) proportionally narrower compared to the original uncropped shot. Cropping changes DoF. Would you get the same change of DoF if you were to change the framing of the original image either by zooming the lens in or by stepping closer to your subject? The answer is no.

Both the link you found (check the slope in the last figure, it is 2 on a log-log scale graph) and a recent discussion here

Help. Aside from photo stacking how do I get more in focus

show that DoF is a quadratic function of focus distance (or framing if you like). This property is true when you shoot at short camera-subject distances - for macro, birds or portraiture, but not for landscape when the focus point is close to hyperfocal distance.

At close focus distances, DoF is also a linear function of f-stop. If you want to double DOF, double your f-number. Of course there is no free lunch and you will lose 2 stops of light and might get noticeable diffraction softening. One always needs to do his own tests to decide what is acceptable image quality. The calculators will give you a good starting point but they are not a substitute for practice.

[Pippan]

I am not quite sure what you mean. Framing in this context means "field of view" - what frame you cover at focus distance. For example a 50 cm by 75 cm field of view is big enough for a head-and-shoulders portrait but is too small for a full body shot. Cropping means taking a part of the original image in post processing and enlarging it to fill the screen or the paper. This enlargement is linear - when you view a cropped image at its full size, all its imperfections will be more noticeable and its DoF will be (almost) proportionally narrower compared to the original uncropped shot. Cropping changes DoF. Would you get the same change of DoF if you were to change the framing of the original image either by zooming the lens in or by stepping closer to your subject? The answer is no.

Thanks for your explanation Dem. My question related to Bill's prior statement that "Provided the Camera Format; the Aperture and the FRAMING of the SHOT remain constant then the Depth of Field will also remain constant" and I was wondering if he meant just the in-camera framing or the final image framing, after cropping. In other words, if you take a shot with a 50 mm lens, then a shot at the same aperture with a 100 mm lens from twice as far (I think that would result in the same framing), would the DoF be the same? And further, if shots were taken from the same spot with the two lenses, and the 50 mm shot cropped in post to the same framing as the 100 mm shot, would the 'DoF' of the final images be the same? My head's spinning about now but I think you're saying no it wouldn't.

If you want to double DOF, double your f-number.

That is a really useful rule-of-thumb I was unaware of. Thanks again.

[george013]

In other words, if you take a shot with a 50 mm lens, then a shot at the same aperture with a 100 mm lens from twice as far (I think that would result in the same framing), would the DoF be the same?

Your subject might have the same magnification, but the AOV will be different.

And further, if shots were taken from the same spot with the two lenses, and the 50 mm shot cropped in post to the same framing as the 100 mm shot, would the 'DoF' of the final images be the same? My head's spinning about now but I think you're saying no it wouldn't.

No. The DoF is related to a reference print and how much the image must be enlarged to get that print. The CoC of a FF sensor is 0.02. If you crop that basic image let's say 25% lineair, than the CoC to calculate with will become 0.02/4=0.005

I'm 100% sure this subject can't be explained without drawings. Counts for me too.

George

[dem]

Bill's prior statement that "Provided the Camera Format; the Aperture and the FRAMING of the SHOT remain constant then the Depth of Field will also remain constant" and I was wondering if he meant just the in-camera framing or the final image framing, after cropping.

In-camera framing, no cropping.

Compared to the DoF change due to the change in in-camera framing, cropping changes DoF at a different rate. That is why for close up shots stepping back and then cropping image in post processing will give you deeper DoF (but less detail!) than trying to frame the shot tightly in-camera. This was discussed in the recent CiC thread I mentioned in my previous post.

[xpatUSA]

Hi guys,
I'm interested in depth of field and was thinking of how to visualise the increase and decrease of depth of field at different apertures, focal lengths and focusing distances.

For example, with each one individually, on a chart with [x and y axes], would they be linear progression, or does the effect suddenly come on?

Interesting question!

And I know that focal length is not considered to be a technical change to depth of field . . .

Reference please.

. . . rather than an enlargement of the blur, but in the real world, it clearly affects apparent depth of field, and I was wondering if this too was linear.

So in case I haven't explained myself, I'm trying to visualise three charts, one for aperture, one for focussing distance and one for focal length and trying to imagine if the line on each chart would be a straight line, or a curved one.

Thanks. Any input is much appreciated.

Geoff

Hi Geoff, if I understand what is meant by "linear", the answer is both "yes" and "no".

Assuming that we are discussing total depth of field, a conventional formula is shown here:

http://dougkerr.net/Pumpkin/articles/Depth_of_Field.pdf

See bottom of page 22 for the total DOF formula which has is a denominator with a squared term in it:

1-(nc/fm)^2.

That makes it non-linear.

However it goes on to say:

" Note that, for reasonably-large magnifications (specifically, when nc/fm is much smaller than 1), this can be well approximated by: [same formula but squared term removed] "

That makes it approximately linear.

So - your charts might be curved or might not be, depending on the magnification (which is defined in the article).

[xpatUSA]

Originally Posted by GrumpyDiver

. . . I was taught many years ago was that a good, safe approximation for minimum viewing distance is the diagonal of the print size.

Following on . . .

Noted the word "minimum" (my underlined and bold).

Cine & Television editing, tethered screens etc use a typical distance of about 1.5 ~ 2 times the screen's diagonal.

The last studio we had we edited / post produced our Still Images at viewing about 1.8 the screen's diagonal: Clients viewed the final product on big screens at a distance of 1.5 the diagonal.

Previously (in Print and Film days) we would show a 24 x 20 inch sample on an easel viewed at about 3' (36") - that's really close to the diagonal but for a good print, that viewing distance can have great impact.

WW

I was truly heartened by these two posts because I frequently infest a forum where the main obsession appears to be with 'detail' and comparisons between cameras. There, merely viewing on-screen is barely acceptable and "the print" is bandied about quite a lot - but in the context of say 40x60" viewed from a foot away, or 8x10" (A4) viewed with a 10X jeweler's loupe!

Thanks!

[Manfred M]

I was truly heartened by these two posts because I frequently infest a forum where the main obsession appears to be with 'detail' and comparisons between cameras. There, merely viewing on-screen is barely acceptable and "the print" is bandied about quite a lot - but in the context of say 40x60" viewed from a foot away, or 8x10" (A4) viewed with a 10X jeweler's loupe!

Thanks!

Ted that is one of my pet peeves as well. The photographer assembles a complete "package" for viewing and while I can understand having a closer look at content, the pixel peeping crowd is literally missing out on the "big picture".

I generally do my sharpening work while viewing at 100% and rarely need to get that close (or closer) unless I am doing some very detailed retouching work.

[William W]

Sooo, can I assume framing of the final image also takes into account any cropping?

Dem answered comprehensively in Posts #32 and #35, thank you. For expansion, I was using these technical terms, defined here:

“Frame [the scene]”; “Framing” – this is done in Camera: it is how the combination of the camera and lens frames the scene, note the Framing is not necessarily what is seen in the Viewfinder, but what is captured on the film or in the image file. “Framing [the scene]” is related to the action of choosing the Focal Length of the Lens and Dem’s response in Post #32 alluded to this by mentioning the technical term “Field of View”. The technical operations are: firstly the Camera Viewpoint is selected (this defines the Perspective of the Shot); and then the Focal Length of the Lens is chosen (this defines the Framing or the Field of View of the Shot). Nowadays, these technical operation are rarely used (or known of) in Still Photography; but are referenced and used in Cinematography; but probably not known to many Videographers unless they were taught or worked in Cinematography previously.

“Crop [the Neg], [the image]”; “Cropping” – this is done in the Darkroom or in Digital Post Production. For negatives often a Crop-mat was used, these are cardboard cut-outs made in various Aspect Ratios, each for specific negative sizes and the neg was mounted onto the mat before being placed in the enlarger.

“Full Frame Crop” – this term is a beautiful oxymoron and is used to note that the Final Image was ‘cropped’ to be the ‘full frame’ of the negative or the image file; a useful term to use if there needs to be clarity to underscore that the image was NOT cropped.

***

. . . question related to Bill's prior statement [Post #24] that "Provided the Camera Format; the Aperture and the FRAMING of the SHOT remain constant then the Depth of Field will also remain constant" and I was wondering . . . In other words, if you take a shot with a 50 mm lens, then a shot at the same aperture with a 100 mm lens from twice as far (I think that would result in the same framing), would the DoF be the same?

“I think that would result in the same framing” – Yes. (and we assume you make the same Plane of Sharp Focus)

“would the DoF be the same?” – For all practical purposes, Yes.

This is the Axiom of DoF. Very useful for Portrait Photography and all other photography which typically would use a Subject Distance relatively much shorter than the Hyperfocal Distance; the Axiom is NOT true for Micro; Macro; and Close-up Photography.

Why it is useful, is because it allows the Photographer to reckon DoF or commit DoF to rote recall by simply using the FRAMING and a set of typically used APERTURES as reference.

Hence, the production of Cheat Cards predicated on that, example here for a 135 Format Camera:



WW

Image © AJ Group Pty Ltd Aust 1996~2017 WMW 1965~1996

[xpatUSA]

Dem answered comprehensively in Posts #32 and #35, thank you. For expansion, I was using these technical terms, defined here:

“Frame [the scene]”; “Framing” – this is done in Camera: it is how the combination of the camera and lens frames the scene, note the Framing is not necessarily what is seen in the Viewfinder, but what is captured on the film or in the image file. “Framing [the scene]” is related to the action of choosing the Focal Length of the Lens and Dem’s response in Post #32 alluded to this by mentioning the technical term “Field of View”. The technical operations are: firstly the Camera Viewpoint is selected (this defines the Perspective of the Shot); and then the Focal Length of the Lens is chosen (this defines the Framing or the Field of View of the Shot). Nowadays, these technical operation are rarely used (or known of) in Still Photography; but are referenced and used in Cinematography; but probably not known to many Videographers unless they were taught or worked in Cinematography previously.

“Crop [the Neg], [the image]”; “Cropping” – this is done in the Darkroom or in Digital Post Production. For negatives often a Crop-mat was used, these are cardboard cut-outs made in various Aspect Ratios, each for specific negative sizes and the neg was mounted onto the mat before being placed in the enlarger.

“Full Frame Crop” – this term is a beautiful oxymoron and is used to note that the Final Image was ‘cropped’ to be the ‘full frame’ of the negative or the image file; a useful term to use if there needs to be clarity to underscore that the image was NOT cropped.

WW

Bill, I've also seen "100% crop" which, in 'Standard Photographic Vague', means anything Chuck Norris says it means.

But often, it means a crop done for showing on-screen with no re-sizing of the original - often for comparison by detail-phreaks.