Focusing Leica: Merklinger Method

[mike_davis]

Better late than never?

 

It's been awhile since this thread was active, but hopefully some of the original posters will reply to this...

 

If the largest circles of confusion seen in the foreground, forward of the plane of sharpest focus, are "acceptable" (i.e. "small enough") when focused at infinity, would circles of this size not be found equally acceptable if they occurred in the background as well, when focused at some point short of infinity?

 

This would extend the range of acceptable sharpness, making use of the DoF that exists just beyond the plane of sharpest focus instead of wasting it the way Merklinger and his disciples do when focusing at Infinity.

 

In other words, if the circles on the near side of the plane of sharpest focus are acceptably small, why not use those of equal size that lie beyond the plane of sharpest focus?

 

I can't imagine anyone answering this satisfactorily. Focusing at infinity when any portion of the subject space resides at some distance short of infinity is a waste of useful DoF.

 

Mike Davis

 

http://www.accessz.com

[ray_moth]

Somehow, the text in this thread became centered. I have no idea how to fix that but it's annoying.

 

I was interested to read Martin's advice when this thread started but have been unable to contribute anything useful from my own experience. I envy those who can see anything at "infinity", such as the horizon. Here, in Jakarta and surrounding countryside, there is so much atmospheric haze that the horizon is usually invisible. In the city itself, even buildings less than a kilometer away appear as ghosts.

 

There are mountains to the south of the city but they can be seen only perhaps on three or four days each year, during the rainy season when there has been heavy rain and some wind the night before. (Pollution, which is severe, tends to accumulate because this is not a windy place; most days, you could blow smoke rings out of doors!)

[WAn]

Mike,<br>

Merklinger suggest focusing at infinity as an ALTERNATIVE to hyperfocal focusing. The goal here is to get the resolution threshold=aperture hole diameter AT EVERY DISTANCE.<br><br>

 

If your scene requires only a finite DOF, i.e. there is a near and a far plane (YOU want to get a required resol.threshold only between this two planes) then you should not have to focus at infinity. According the Merklinger approach you have to focus at a plane exactly between these near & far boundaries. And the needed aperture is given by Merklinger formulas.<br><br>

 

An attempt to get rid of centering: </i><br>

 

Success?<br>

 

Or not?<br>

[MTC Photography]

Mike,

<p> There is no magic in dof, there is aways that much dof to

distribute, and there are two approaches

<ul>

<li> Front end biased, foreground objects predomenat use traditiona

DOF method

<li>Far end biased, when objects, scenes at distance are predominant

then use Merklinger way, focus at infinity<p></ul>

I think it is a matter of choice.

<p> If in a scene, the close up subject is predominant, then use

the conventional DOF method.

<p> On the other hand, if in the scene there are more scene at distance then close by, then I use Merklinger method.

<p>I do a lot of Minox picture with Merklinger's focus at infinity

method.

[MTC Photography]

For example, Minox camera at wide open f/3.5 has diameter 15/3.5=

4.3 mm. When focus at infinty, any object in the scene with diameter>4.3mm will be resolved. <p> In the sample Minox picture,

the diameter of the grids, and the spacing between the grids are all

less then 4.3mm, hence they are sharp from front to rear <p> If the Minox was focus at hyperfocal distance of 2 M

instead of at infinity, all the grids would not be solved from near

to far, because the disk of confusion grows larger and larger.<div></div>

[mike_davis]

Martin,

 

Thanks for sharing the nice photo, but using a Minox image to illustrate the acceptable DoF had when focused at infinity is like arm wrestling with Hulk Hogan to prove that bald men are stronger than those with a full head of hair.

 

The smaller the format the more DoF you get using equivalent focal lengths. Your use of f/3.5 also avoided the visible dffraction suffered by Minox cameras when stopped down.

 

You wrote, "there are two approaches" - "front end biased" and "far end biased" - stating that it's "a matter of choice." If you are knowingly "choosing" to secure softer foregrounds by focusing at infinity, fine, I will vigorously defend your choice, but if you believe your "choice" provides sharper images than can be had with the "conventional DoF method", you are mistaken.

 

You wrote, "If in a scene, the close up subject is predominant, then use the conventional DOF method."

 

Please explain what advantage focusing at Infinity offers over "the conventional DoF method" when "scenes at a distance are predominant."

 

There is no advantage.

 

You haven't answered my original question. Here it is again:

 

If the largest circles of confusion seen in the foreground, forward of the plane of sharpest focus, are "acceptable" (i.e. "small enough") when focused at infinity, would circles of this size not be found equally acceptable if they occurred in the background as well, when focused at some point short of infinity?

 

Yes or No?

 

Your "choice" to use Merlinger's method when the scene is predominantly at a distance clearly indicates that some portion of your subject falls short of infinity and you are enjoying at least some near-side DoF if you find the nearest subjects acceptably sharp. If that's the case, why are you choosing to waste the acceptably small, equal-diameter circles of confusion that lie beyond the plane of sharpest focus?

 

Merklinger's method is not an acceptable alternative to the conventional method if image clarity is our goal. Focusing at infinity and stopping down enough to get acceptable results at the nearest subject invites three sources of image degradation which you would not suffer were you to shoot at a wider aperture and focus more closely (yielding acceptably small CoC's at both the near sharp and at infinity):

 

1) Stopping down further than necessary increases the diameter of diffraction's Airy disks. (As circles of confusion shrink when stopping down, Airy disks get larger.)

 

2) Stopping down further than necessary increases your vulnerability to camera and subject motion because longer expoures must be employed.

 

3) Stopping down further than necessary may take you further away from the aperture at which your lens delivers its best resolution.

 

So, please tell me how it's advantageous to focus at infinity and then stop down enough to achieve acceptably small circles of confusion at the near sharp, when those same diameters could have been enjoyed at a wider aperture (at both the near sharp and at infinity) were you focused at the hyperfocal distance.

 

Most people are disappointed by "the conventional DoF method" because they are using DoF scales engraved on their lenses (or some other DoF reference) that were generated with too large a circle of confusion diameter for the enlargement factor and viewing distance their application requires. Many DoF calculators treat circle of confusion diameter as a constant that's unique to each format (0.03mm for 35mm, 0.06mm for 6x6cm, etc.) and the lens manufacturers' engravings are obviously inflexible - they only work for one ratio of viewing distance to enlargement factor (and are usually not very aggressive in the first place.)

 

Handling circle of confusion diameter as a constant when calculating DoF for a given format is ridiculous. It's a variable. Before you can calculate DoF you must calculate the CoC that's right for your application. Here's how:

 

Maximum Permissible CoC Diameter On-film = 1 / enlargement factor / desired print resolution

 

The average adult with healthy vision can resolve no more than 6 to 8 lp/mm at a distance of 10 inches. So, if you want a resolution of 7 lp/mm in your final print for an 8x enlargement factor, you must limit on-film circle of confusion diameters to:

 

Maximum Permissible CoC Diameter On-film = 1 / 8 / 7 = 0.0179 mm

 

(This is much smaller than the 0.03 mm value typically used to calculate DoF for the 35mm format.)

 

That's for a viewing distance of 10 inches. If, however, you intend to make a print that will be displayed such that it can be viewed no closer than 30 inches, instead of at 10 inches, you should reduce your resolution requirement by a factor of three:

 

Maximum Permissible CoC Diameter On-film = 1 / 8 / (7 / 3) = 0.0536 mm

 

(That's much larger than the 0.03mm value typically used for 35mm format.)

 

It's just that easy to come up with a value for CoC you can plug into your choice of several DoF calculators/spreadsheets/Palm programs/java scripts/whatever that actually permit you to specify the on-film maximum accpetable diameter for CoC's.

 

When coming up with the enlargment factor, don't neglect to take cropping into account. If, for example, you intend to make an 8x10 print from the 24x30mm portion of the fullframe 35mm negative, you'll want to calculate the enlargement factor using the smaller dimensions.

 

A little experimentation with tables generated using this approach will tell you by how many stops you have to offset the f-stops suggested by your lens barrel DoF scale for each lens. That way, no math is required in the field. You just have to do your homework up front.

 

At best, Merlinger's suggestion is a novel procedure that compromises image quality for the sake of convenience. It is not a reasonable alternative to the conventional method nor is there room for subjective input if image clairity is your goal. If you doubt my contention, revisit my original question and consider the pitfalls of stopping down unneccessarily.

 

Mike Davis

www.accessz.com

 

[WAn]

"If the largest circles of confusion seen in the foreground, forward of the plane of sharpest focus, are "acceptable" (i.e. "small enough") when focused at infinity, would circles of this size not be found equally acceptable if they occurred in the background as well, when focused at some point short of infinity?

 

Yes or No?"

 

 

-- Yes.

 

 

"Merklinger's method is not an acceptable alternative to the conventional method if image clarity is our goal."

 

 

If the "clarity"� is not compatible with resolution threshold at every distance = aperture diameter, then Merklinger's method is NOT an acceptable alternative. Otherwise it IS acceptable. --- It depends on your requirements.

 

"1) Stopping down further than necessary increases the diameter of diffraction's Airy disks."

 

Sure. Moreover it can be shown, that if we take into account the diffraction then there is no reason to focus beyond a certain point. But this point is much farther than usual hyperfocal distance. I don't like to post formulas into the forum, but it is not difficult.

 

 

"...when those same diameters could have been enjoyed at a wider aperture (at both the near sharp and at infinity) were you focused at the hyperfocal distance."

 

 

I beg to differ. I don't mind the "near sharp" but the resolution at infinity is much worse if we focus at hyperfocal distance compared with the case when we focus exactly at infinity (or, more precisely, at that far point I mentioned above). --- Generally speaking, it is not a surprise: anything is sharper when the focus is at it than when it is only kept within a DOF.

 

 

As an aside note.

Merklinger's approach is a special one (criterion is in terms of resolution threshold in space of objects), but it DOES relate to a classical approach (where criterion is in terms of circle of confusion in space of images). Indeed, lets generalize a bit the Merklinger's criterion: lets ask for different res.thresholds at near and at far limits of sharpness zone. The new formulae are easy to obtain. Then consider a special case when we require that the size of res.threshold is proportional to the distance. If you obtain the formulae for this case you will see that the formulas fully agree with simplified CLASSICAL DOF formulas (the simplification is for situations when focusing distance is significally greater than the focal length, -- it is ok for every situation except close ups). In other words, if we request the same ANGULAR resolution threshold at near and far limits of a scene, then Merklinger's result is totally identical with classical result.

[MTC Photography]

Mike wrote:

 

"The smaller the format the more DoF you get using equivalent focal lengths. Your use of f/3.5 also avoided the visible dffraction suffered by Minox cameras when stopped down."<p>

Stop down Minox ? <p>

There is no such thing as stopped down with Minox 8x11 cameras. For

Minox A, B,C, LX, TLX, CLX, the lens is always used at wide open f/3.5. There is no aperture control<p>

 

"Please explain what advantage focusing at Infinity offers over "the conventional DoF method" when "scenes at a distance are predominant."

 

There is no advantage. "<p>

Sure there is, great advantage<p>

 

"If the largest circles of confusion seen in the foreground, forward of the plane of sharpest focus, are "acceptable" (i.e. "small enough") when focused at infinity, would circles of this size not be found equally acceptable if they occurred in the background as well, when focused at some point short of infinity?

 

Yes or No? " <p>

The answer is No.<p>

<p>Think for yourself for a moment, " What object corresponds to 0.03mm at far zone ?"

<p> For example, at 30 meter with 50mm lens, the magnification is

approximately 600 times, a 0.03mm corresponds to 18 mm.

<p> It is far to big. I prefer not using ONE coc for both far and near object, instead, I prefer a sliding scale of coc, ie, for distance

object, use very small coc (ie, 0.005 mm at 30 meter) and at the same time, for forground object use larger coc( 0.03 at 5 meter)

<p>That is what focusing at infinity offeres: a sliding coc, smaller

coc for far object, larger coc for near objects.

[mike_davis]

Martin,

 

I stand corrected regarding use of stops below f/3.5 with a Minox. I don't know where I heard that and have never used a Minox. Thanks for catching that one. Still, the DoF enjoyed by this small format makes it a poor choice for illustrating that sufficient DoF can be had when focused at Infinity.

 

As an aside: Minox's choice to fix the aperture at f/3.5 was a great idea considering the enlargement factors this format must endure.

 

The on-film Airy disk diameter at f/3.5 is 0.0047384mm.

 

3.5 * 0.00135383 = 0.0047384 mm

 

Given that the reciprocal of 7 lp/mm is 1/7 mm, or 0.1429 mm, this means that in regards to diffraction at least, the Minox negatives can tolerate a 30x enlargement before the Airy disks produced at f/3.5 would become visible in a print viewed at a distance of 10 inches.

 

0.1429 / 0.0047384 = 30.16

 

Thus, at f/3.5, you can enlarge the 8x11mm negatives to 9.5 x 13 inches without suffering visible diffraction (7 lp/mm).

 

Regarding your preference for a "sliding CoC" - "smaller coc for far object, larger coc for near objects":

 

If we share the goal of making both the near and far CoC's too small to resolve in the final print at the anticipated viewing distance, focusing closer to infinity or at infinity (such that the far CoC will be smaller than the near CoC) will require a smaller aperture than could be used if we instead focused at the hyperfocal distance. Your preference for making the far CoC's smaller than the near CoC's means that you either permit the near CoC's to be resolvable (visibly out-of-focus) in the foreground so that you can use the same or wider apertures than you would if focused at the hyperfocal distance -OR- you don't tolerate soft foregrounds and do make the near CoC's unresolvable by stopping down to compensate for having focused beyond the hyperfocal distance. Which is it?

 

Are you allowing foreground subjects to be visibly out of focus so that you can shoot at wider apertures or are you stopping down more than you have to keep the nears visibly sharp despite having focused too long? Neither of these possibilites are pallatable to me.

 

My original question spoke in terms of the near and far both being "acceptably" sharp (not resolvable in the final print) - your response implies that Infinity subjects can not be made acceptably sharp without focusing beyond the hyperfocal distance. That's incorrect. Independant of problems with diffraction (which I discussed above), for every situation, one can find an aperture at which the CoC's will be small enough at Infinity, when focused at the hyperfocal distance, to make both the near and Infinity coc's just sharp enough to be unresolvable by the human eye. This aperture will be wider (faster) than the aperture you would have to use to make your nears visibly sharp when focusing beyond the hyperfocal distance.

 

Merklinger's suggestion compromises foreground sharpness or speed - take your pick.

 

Mike Davis

[mike_davis]

Hi Andrey,

 

In response to my having written "...when those same diameters could have been enjoyed at a wider aperture (at both the near sharp and at infinity) were you focused at the hyperfocal distance."

 

You wrote: "I beg to differ. I don't mind the "near sharp" but the resolution at infinity is much worse if we focus at hyperfocal distance compared with the case when we focus exactly at infinity (or, more precisely, at that far point I mentioned above). --- Generally speaking, it is not a surprise: anything is sharper when the focus is at it than when it is only kept within a DOF."

 

Like Martin, your words imply that it is somehow impossible to find an aperture at which CoC's will be small enough to be unresolvable for both the near and Infinity when focused at the hyperfocal distance.

 

My second posting to this thread clearly explained how to select a CoC diameter for use in DoF calculations that will be unresolvable in the final print at the anticipated viewing distance. That diameter can be achieved simultaneously at both the near sharp and at Infinity ONLY when focused at the hyperfocal distance. At the aperture calculated in "conventional" DoF calculations using the appropriate CoC value, stopping down any further in an effort to make CoC's smaller still, would be pointless - you wouldn't be able to SEE any improvement in the final print. At THIS aperture, if we were to focus closer to Infinity than the hyperfocal distance, we would only make the foreground CoC's larger (visibly defocused) and the Infinity CoC's needlessly smaller than (they were already unresolvable when focused at the hyperfocal distance.) How can you "beg to differ" when I say that Merklinger's method would require stopping down to a smaller aperture than this one to re-establish unresolvable CoC's in the foreground?

 

Also: DoF is the range of distances in which subjects are found acceptably sharp in the final print at a given viewing distance. Your statement that "anything is sharper when the focus is at it than when it is only kept within a DOF" makes no sense in the context of the definition of DoF. The only way your comment could be deemed correct would be to modify the definition of DoF to read like this: "DoF is the range of distances in which subjects MAY OR MAY NOT be found acceptably sharp in the final print at a given viewing distance." In other words, DoF defines the range of distances where everything is "acceptably" sharp, so how could anything look sharper when the focus is biased toward one end of the range, when doing so would make subjects at the other end unacceptably sharp. You used the term "DoF" in your statement without honoring its meaning.

 

Mike Davis

[nesrani]

"Sliding coc" eh? Maybe I should give the merklinger method a go...

[nesrani]

<i>smaller coc for far object, larger coc for near objects </i>

<p>

We all know what that feels like, I'm sure...

<p>

OK, enough coc jokes. I promise.

[ilkka]

There is one thing here that I do not understand, and thus do not agree with.

 

In describing Merklinger's tests, Martin Tai pointed out that focusing the lens at 9.1m, objects from 3m to 18.3 m were acceptably sharp. He then went on to say that by focusing at infinity, objects from 1m to 50 m and beyond were acceptably sharp. Now how can it be that focusing FURTHER AWAY will bring CLOSER objects to acceptable focus?

 

The only explanation I can come up with is that by focusing closer, you actually increase your focal length slightly , which should reduce your depth of field. But since the actual aperture stays the same, the relative f stop also decreases the same amount, and should compensate. (120mm lens at infinity, F8, actual aperture is 15mm. Focusing closer, say by 10mm, focal length is now 130mm, relative aperture F8.7). However, at normal shooting distances the difference is so small that it should not have an effect. And certainly not that big effect to bring objects from 1 m instead of 3m into sharp focus.

 

As I see it, there exists a depth of acceptable sharpness in front AND behind the plane of focus. Merklinger and Martin seem to agree that this field of acceptable sharpness occurs, at least in front of the plane of focus, at least when focused on infinity. Exactly how wide this field is depends on aperture and size of reproduction, as well as on how critical eye is used to judge the result, while using the same lens and film size. If we agree that this field of sharpness occurs in front of the plane of focus, I think we should also easily accept that it does also occur behind the plane of focus. In books it is commonly said that the depth of field actually extends one third in front of the plane of focus and two thirds behind it.

 

If this is true, and I really can't see any reason why it would not be, then by focusing at infinity you will lose that part of the depth of field that goes beyond infinity. And by focusing at a point closer than infinity, you would get a wider acceptable depth of field. That is, a closer object would be in acceptable focus than if you focused at infinity, while object at infinity would still be in acceptable focus.

 

Exactly where to focus and what aperture to use? Now that depends on the factors mentioned above, focal length, film size, size of reproduction and the viewing distance. Lens manufacturers use certain assumptions when engraving their depth of field scales. You must not accept them as given without testing that you agree with the assumptions! But that does not make the theory behind it invalid.

 

If your lens says hyperfocal distance focusing at 10m and aperture at f8 will bring everything from 5 m to infinity sharp, you may disagree with it and instead use 1-3 stops smaller aperture at same distance setting. To compensate for your higher requirements for what is acceptable sharpness.

[WAn]

Hi Mike,

 

 

"DoF is the range of distances in which subjects MAY OR MAY NOT be found acceptably sharp in the final print at a given viewing distance."

 

 

DoF is the range of distances in which subjects are found equally or more sharp than a minimum sharpness level required by the chosen criterion.

 

 

Your second posting to this thread is indeed very important to understand your logic. I admit I considered it only in positive sense, as a way to enhance the CoC standard, but there is also a negative sense there: any extra sharpness beyond the level of acceptable sharpness in an overkill and is either useless (invisible) or harmful (wasting of resources). Your predefined enlargement factor and viewing distance act as a cut off filter.

 

 

These restrictions are not perfect (a curious viewer can come closer, I may arrive at idea to make greater enlargement etc), but if we assume that we can guarantee that enlargement factor and viewing distance are always kept as intended, then I agree with your logic.

 

 

It can be shown that, when the focus in on hyperfocal distance, the linear resolution gets progressively worse after h.distance, and this degradation is unrestricted. On the other side, the angular resolution keeps constant. When the focus is at infinity, the linear resolution is constant, the angular resolution threshold is 0 (in realty we shouldn't ignore the diffraction in this case, but for sake of discussion we will).

 

The goal of focusing at infinity is to keep objects located at infinity as sharp as possible.

 

According your logic even if there is nothing in foreground (e.g. only distant mountains, moon and stars in the sky) it does not matter shall I focus at infinity, at hyperfocal distance or somewhere in between. You probably admit that in principle there will be difference in moon clarity, but that cut-off filter will destroy the difference and our canonical viewer will perceive all three versions as totally identical. --- Correct?

 

The same note for a scene without infinity, say from 3m to 5m. If our DoF does exactly cover this range, the extra sharpness gained at exact focus plane is useless: the canonical viewer won't appreciate it. --- Correct?

 

If I abandon the canonical viewing conditions the above considerations are no anymore true. (My objective can be different: instead of ultimate sharpness within a zone I may want to be able only to resolve a certain size i.e. perceive some adjacent things as different; I can found that the appropriate CoC implies so narrow DoF that I have to reconsider my sharpness requirements. Finally I may wish intentionally blur something.) That's why I added the words "required by the chosen criterion" in the DoF definition at the start of this message.

 

Sorry for so long message.

[mike_davis]

Andrey,

 

My answer to both of the questions you asked above is "Yes", but neither of those two scenarios address the issue being debated in this thread. Please respond to this:

 

When the subject space does NOT lie entirely at Infinity, but DOES include Infinity (1m to Infinity, for example, or even 100m to Infinity), focusing at the hyperfocal distance will allow us to achieve the desired sharpness with the widest possible aperture. Do you agree? (Previously, you said you "beg to differ", but the comments you've made more recently prove that you understand the principles involved.)

 

For THIS scenario, focusing at Infinity will require us to use a smaller aperture to achieve the same size CoC's at the nearest subject as we would at Infinity. Correct?

 

If you can answer "Yes" to both of these questions, then you must agree that Merklinger's method compromises image quality for the sake of convenience - it's just a quick and DIRTY way to select an aperture that will render near subjects sharply without having to worry about where to focus (we can just focus at Infinity.) It's "dirty" because stopping down further than one would if focused at the hyperfocal distance invites three forms of image degradation (discussed earlier).

 

Regarding our ability to accurately predict final enlargement factor and viewing distance: Completely ignoring considerations of final print size and viewing distance at the time of exposure will consistently produce poorer results than making SOME attempt to predict the final conditions. The only way to willfully ignore these issues at the time of exposure and actually come home with consistent results is to cover every possibility: You'd have to assume that every print will be viewed very closely, at a viewing distance of 10 inches (most adults with healthy vision can focus no more closely than this) -AND- you must assume it will suffer whatever enlargment factor can be delivered by the resolution of the lens and film together as a system. For the best lenses with color films, I consider this to be no more than about 9x, for the best black and films about 12x.

 

So, anyone who insists on ignoring decisions about print size and viewing distance at the time of exposure, should shoot EVERY scene with an aperture selected for a DoF that will deliver CoC's small enough to suffer 9x enlargement (or 12x for B&W) and a 10-inch viewing distance. That's excessive! Even our shortest lenses would have horrible DoF - requiring us to stay much farther away from the nearest subjects than we could were we to just make a decision about enlargement factor and viewing distance at the time of exposure.

Establishing a "cut-off filter" as you call it, by contemplating enlargement factor and viewing distance at the time of exposure, allows us to NEVER stop down more than we have to - and this translates to improved image clarity three ways. (Again: Wider aperture = less diffraction, less vulnerability to camera and subject motion, and perhaps an aperture that's closer to your lens' aperture of best resolution - usually near the middle of the aperture range.)

 

We CAN take control of ALL the variables that affect apparent sharpness instead of leaving them to chance or routinely stopping down more than necessary at the expense of diffraction, vulnerability had with longer exposures, etc. It only takes me three minutes to calculate this stuff in the field, for every exposure, using an HP48G+ (as described in the Word doc that can be found on the Tools page of my web site.) The results have been so spectacular, I can't imagine going back to guessing where to set the aperture.

 

Thank you,

 

Mike Davis

 

http://www.accessz.com

[dlegaspi]

i must admit i could hear my brain crying pain as i try to understand the technical aspects of this focusing method.

[mike_davis]

Dexter,

 

I'm dying to know which method you're referring to -

 

Merklinger's focus-at-infinity method?

 

Conventional methods using hyperfocal distance?

 

Or conventional methods enhanced with considerations of enlargement factor and viewing distance?

 

Thanks,

 

Mike Davis

 

http://www.accessz.com

[WAn]

Mike,

If one accepts your criterion and all your prerequisites then there is no choice anymore, thus my answers are "Yes" to both your questions.

 

Let me explicitly state your premises:

 

1) The goal is to find optimal way (with largest possible aperture) to achieve a minimum acceptable sharpness from a range of distances of the scene.

 

2) Definition of the minimum acceptable sharpness: a point from the scene is said to be reproduced with minimum acceptable sharpness if the circle of confusion that the point produces on print is equal to linear resolution of viewer's sight.

 

3) The enlargement factor and viewing distance are predefined and are known at the moment of shoting.

 

4) Some other obvious and undebatable assumptions like that the lenses, film and paper are perfect (the diffraction may or may not be taken into account) etc.

 

The rest is the matter of technique. No one -- neither Harold Merklinger not John Smith -- can claim that he accepts all this premises AND has found any formulae that achieve the goal (1) better than well known traditional formulae; doing so would imply that mathematics is an inexact science. Your position therefore is invulnerable.

 

People say devil is in details. I'd better say the devil is in failure to find and critically analyze any silently made assumptions. If one want to attack your position he has to attack not a result, but the premises the result is based on.

 

For example I find the condition #3 not always acceptable; I understand and respect your argumentation but there are situations where I don't want obey this rule (in certain situations I may want the LINEAR resolution in space of objects be kept constant; it depends on my objective, depends on whether the print will be on the wall and be viewed from a reasonable distance or will a portion of the negative be enlarged and I must be sure the shapes on this fragments are recognizable); inevitable consequences of such a whim is that traditional formulae may become not optimal anymore.

 

(An curious aside: there is another seemingly weak point for attack: the definition of sharpness. The statement (2) roughly speaking tells the following: the lens maps a point to a circle, but the size of this circle is the maximum size that can be resolved by viewer, therefore the viewer perceives this circle as a point. The net result is "point to point" and it probably means a sharpness. --- This reasoning could be 100% true if the viewer's eye had an ability to de-blur the circle back to a point. But it hasn't. Another observation: the above definition of sharpness is based on metamorphoses of the single point, but sharpness is rather an attribute of a shape, of a continuum of points; hence the following argument: the CoCs from neighbor points of the scene do partially overlap, the image of one point is partially mixed with image of another one and even in the best case the net result should be "point to point, minus some contrast". --- Actually this issue is worth a more thorough investigation. I tried to do some rough estimates, and good news is that the statement (2) seems to be a good criterion. My point is that this statement is not in the least obvious.)

 

To sum up: if I assume all your premises then indeed the Merklinger's method is a quick and dirty method to do the job.

 

It would be unfair if I didn't mention the assumptions the Merklinger's method is based on. The narration in his articles is clear enough but I suspect some misunderstanding still exists. So please tolerate me with my broken English some time more.

 

(continued in next message)

[mike_davis]

Andrey,

 

While we wait for the second half of your last post, I thought I'd reply to what you've stated thus far.

 

You wrote:

 

"Mike, If one accepts your criterion and all your prerequisites then there is no choice anymore, thus my answers are "Yes" to both your questions."

 

OK, that's fair. I'm glad you agree.

 

"Let me explicitly state your premises:

 

1) The goal is to find optimal way (with largest possible aperture) to achieve a minimum acceptable sharpness from a range of distances of the scene.

 

2) Definition of the minimum acceptable sharpness: a point from the scene is said to be reproduced with minimum acceptable sharpness if the circle of confusion that the point produces on print is equal to linear resolution of viewer's sight.

 

3) The enlargement factor and viewing distance are predefined and are known at the moment of shooting.

 

4) Some other obvious and undebatable assumptions like that the lenses, film and paper are perfect (the diffraction may or may not be taken into account) etc."

 

I can accept your interpretation of my premises.

 

"The rest is the matter of technique. No one -- neither Harold Merklinger not John Smith -- can claim that he accepts all this premises AND has found any formulae that achieve the goal (1) better than well known traditional formulae; doing so would imply that mathematics is an inexact science. Your position therefore is invulnerable."

 

Thank you.

 

Regarding your comments about premise #3: As I wrote higher up in this thread, if we don't make the choice prior to exposure and then adhere to that choice later, the only safe way to shoot is to assume that every image will suffer the greatest possible enlargement factor and that it will be viewed very closely (at a distance of 25 cm).

 

Again, if we refuse to consider enlargement factor and viewing distance at the time of exposure, this is the only safe alternative. Any "whim" that comes along BEFORE the exposure is made, should be addressed by adjusting the CoC diameter used for DoF calculations right then and there. And any "whim" that comes along AFTER you make the exposure had quite simply better be a "whim" to reduce enlargement factor below that which was anticipated or a "whim" to increase viewing distance. In ALL cases, the traditional formula is inescapable. There is no way to ignore it without compromising something.

 

"To sum up: if I assume all your premises then indeed the Merklinger's method is a quick and dirty method to do the job."

 

This reads as if there exists a circumstance, where one or more of my premises are not in effect, in which Merklinger's method is "clean".

 

Please answer this question: What is that circumstance where it is better to throw away all the DoF that resides beyond the plane of sharp focus by focusing at Infinity?

 

Please don't take us full circle back to discussing scenarios other than the one in question here - where some portion of the subject space falls short of Infinity.

 

You've come very close to admitting that Merklinger's method always compromises image quality for the sake of expedience, but you are still making tangential arguments that don't address the issue.

 

You wrote: "It would be unfair if I didn't mention the assumptions the Merklinger's method is based on. The narration in his articles is clear enough but I suspect some misunderstanding still exists."

 

This too is a weak rebuttal - suggesting that if only we understood Merklinger's method, we would see the value of it. I fully understand his method. It's a gimmick � a clever shortcut that compromises image quality. Here's a page written by the man himself. (Do read the whole of it at your leisure, but for right now, please consider only the photograph and its caption.):

 

http://home.fox.nstn.ca/~hmmerk/DOFR.html

 

Martin Tai stated he believes using the conventional method vs. Merklinger's method "is a matter of choice." He wrote, "when objects, scenes at distance are predominant then use Merklinger way, focus at infinity."

 

Please examine the photograph at Merklinger's page, above. Is that cannon and the ground on which it is setting "predominantly at a distance"? No it is not. Now read the caption under the picture. Read it again and again - those are Merklinger's own words. What is there to misunderstand Andrey? Merklinger is WRONG! The statement made in that caption is RIDICULOUS.

 

(Continued in next post...)

 

Mike Davis

 

http://www.accessz.com

 

[mike_davis]

(Continued from previous post...)

 

Can't you agree Andrey that were we to calculate DoF such that our maximum permissible CoC's were no smaller than those Merklinger has produced in the foreground of this image we could select an aperture that is WIDER than the one Merklinger used to shoot that picture (by focusing CLOSER than Infinity, precisely AT THE HYPERFOCAL DISTANCE, producing comparable results without WASTING the DoF that resides beyond the plane of sharp focus)?

 

Merklinger either finds the foreground CoC's to be acceptably small in that photo or he has sacrificed their sharpness to maximize that had at Infinity. Which is it? He writes in the text that follows the photo: "The foreground is admittedly not tack-sharp." Gee, sounds like a compromise to me! He goes on to write: "Had I focused at the hyperfocal distance the telephone poles in the village would have been almost erased, and windows in buildings would have been just blurs."

 

Andrey, don't you find it conspicuous that Merklinger again fails to qualify his use of the phrase "hyperfocal distance" with the CoC diameter specified for the DoF calculations? In fact, he has failed to qualify a LOT of important variables affecting perceived sharpness!

 

Even with 1/30mm CoC's at the near and far sharps, the ENTIRE image would appear "tack-sharp" if the combination of viewing distance and enlargement factor was not sufficiently demanding to allow the viewer to resolve those CoC's. Right?! Give me any combination of enlargement factor and viewing distance and I can come up with a

CoC diameter most people will be unable to resolve. DoF calculations made with this CoC diameter will produce images that ARE "tack-sharp" when focusing at the hyperfocal distance (ignoring other factors that limit the total system resolution.)

 

Speaking of ASSUMPTIONS: Please read the last two sentences of Merklinger's "Introduction" section. Merklinger says that if we focus at the hyperfocal distance, we "will have sealed in that 'minimum acceptable standard' ". What standard is he talking about? Higher up he explains that he's talking about a 1/30mm standard for maximum permissible circles of confusion at the near and far sharps.

 

Mr. Merklinger! Will you please tell me what's preventing us from ABANDONDING that standard? Will you please tell me why we can't calculate DoF tables that give us apertures and hyperfocal distances that produce SMALLER circles of confusion than this standard you find so disappointing?

 

Merklinger's entire argument for avoiding hyperfocal focusing is fallacious! Don't you see that he is ASSUMING we are stuck with 1/30mm CoC's? In my opinion, the man knows better. He's just scratching up an excuse with which to argue that his technique will yield better results when the truth is it doesn't work as well as doing things the old fashioned way.

 

Merklinger says we'll be "guaranteed" to have "mediocre" results if we focus at the hyperfocal distance. That's ONLY TRUE if we join him in pretending that 1/30mm is the ONLY diameter CoC we can use in our DoF calculations!

 

In the paragraph immediately below the picture on Merklinger's page, he writes: "The hyperfocal distance for a 90 mm lens at f/8 is 106 feet." Harold! That's the hyperfocal distance for a 90mm lens at f/8 with a maximum permissible CoC diameter of 1/30mm!!! It is not THE ONLY hyperfocal distance! It is one of MANY POSSIBLE hyperfocal distances that could result with various choices of CoC diameter.

 

If I CHOOSE to make my CoC's SMALLER than the 1/30mm "standard" that leaves you so "bothered" by "old story about maximizing DoF by focusing at the hyperfocal distance", I can achieve any degree of sharpness I desire, right up to the limits of total system resolution! And YOU can TOO!

 

It's simply amazing that a man who wrote an essay titled "The INs and OUTs of FOCUS" could pretend to miss the fact that we can CHOOSE any CoC diameter we want for our Near and Far sharps and thus produce perfectly acceptable images while focusing at the corresponding hyperfocal distance. There is NOTHING wrong with focusing at the hyperfocal distance. You simply have to calculate DoF with a CoC diameter that's aggressive enough to suit the anticipated enlargement factor and viewing distance.

 

When I challenged Martin Tai with: "Please explain what advantage focusing at Infinity offers over "the conventional DoF method" when "scenes at a distance are predominant." There is no advantage. "

 

He simply said: "Sure there is, great advantage" and then failed to explain that advantage except to throw us this bone: "That is what focusing at infinity offeres: a sliding coc, smaller coc for far object, larger coc for near objects." So what? I responded with a question he has yet to answer:

 

"Your preference for making the far CoC's smaller than the near CoC's means that you either permit the near CoC's to be resolvable (visibly out-of-focus) in the foreground so that you can use the same or wider apertures than you would if focused at the hyperfocal distance -OR- you don't tolerate soft foregrounds and do make the near CoC's unresolvable by stopping down to compensate for having focused beyond the hyperfocal distance. Which is it?"

 

Andrey, for the sake of all the people out there who may read this thread without the benefit of your ability to digest it as thoroughly as you have, can't you join me in putting this myth to death? How can you support Merklinger's contention that hyperfocal focusing will always produce mediocre results, when that contention is based on the assumption that hyperfocal distances are somehow permanently fixed to a 1/30mm standard for CoC's?

 

Mike Davis

 

http://www.accessz.com

 

[MTC Photography]

Mike wrote"

"It's simply amazing that a man who wrote an essay titled "The INs and OUTs of FOCUS" could pretend to miss the fact that we can CHOOSE any CoC diameter we want for our Near and Far sharps and thus produce perfectly acceptable images while focusing at the corresponding hyperfocal distance. There is NOTHING wrong with focusing at the hyperfocal distance. You simply have to calculate DoF with a CoC diameter that's aggressive enough to suit the anticipated enlargement factor and viewing distance.

"<p>

I just don't like to use same coc for foreground/background.<p>

 

I prefer use larger coc for larger object and smaller coc for background object.<p>

[mike_davis]

Martin,

 

You wrote: "I just don't like to use same coc for foreground/background. I prefer use larger coc for larger object and smaller coc for background object."

 

You've taken us back to where we were a few days ago! You still haven't answered a question I asked then. Instead, you have only reiterated the statement which prompted me to ask the question in the first place! Your avoidance of this question may be seen by others as evidence that you are unable to defend your choice to use Merklinger's method:

 

"Your preference for making the far CoC's smaller than the near CoC's means that you either permit the near CoC's to be resolvable (visibly out-of-focus) in the foreground so that you can use the same or wider apertures than you would if focused at the hyperfocal distance -OR- you don't tolerate soft foregrounds and do make the near CoC's unresolvable by stopping down to compensate for having focused beyond the hyperfocal distance. Which is it?"

 

You have several times contended that Merklinger's method is a valid alternative that's a matter of personal choice, but you refuse to respond to my contention that his method compromises foreground sharpness -OR- it compromises shutter speed (by forcing the use of a smaller aperture than would be necessary if you just focused at a hyperfocal distance calculated to yield acceptable CoC's for the anticipated enlargement factor and viewing distance.)

 

Again I ask: Which compromise do you prefer?

 

No matter how you answer that question, we're left with this one: Why do you tolerate either compromise? Why are you "choosing" to shoot yourself in the foot? You said there was a "great advantage" to using Merklinger's method. Let us hear your explanation of this advantage!

 

Lastly: How long will you continue to be evasive? Surely you must realize that silence is as convicting as your refusal to directly address the arguments I've presented. Are you interested in communicating the truth, for the benefit of those less knowledgeable than you, or is your mission simply to save face? Do you want people to conclude that all your technical articles are suspect?

 

There are only two ways to preserve the integrity of your reputation as a resource for photographic knowledge: Agree with my arguments or give us a logical, defensible rebuttal to the challenges I've made.

 

Mike Davis

 

http://www.accessz.com

[WAn]

Gentlemen, I apologize for the delay, the cause wasn't a whim :)

 

 

TWO points are said to be RESOLVED if their images can be discerned, i.e. if we can recognize them as two different spots, not as a single blob. The magnification is not restricted.

 

 

Please note:

 

 

1) Resolvability is an objective property, it can be measured. Sharpness is a subjective impression.

 

 

2) Resolvability in this definition (a commonly used definition, by the way) does not imply sharpness; the images of two points can be arbitrary blurred, the two CoCs can be arbitrary large, only one thing is needed: the distance between them should be large enough. Very roughly speaking, the resolvability is more relaxed requirement than sharpness.

 

 

Harold Merklinger (further H.M., I hope he won't mind) speaks not about two points; he speaks about a disk in space of objects. The lens equation makes it equivalent to the definition above if we chose two ends of any diameter of the disk as those two points from the definition and if we say that two CoCs are resolved if they are tangent to each other (see the graph).

 

 

(Mike, please note: you used the term "resolution" in opposite sense. You spoke about resolvability of CoC by viewer�s eye: if CoC is resolved then it is visible, then the point is not sharp; resolvability=unsharpness (bad thing). And vice versa: if CoC isn't resolved the picture is sharp (good thing). H.M. (and I in all my posts above) have used the term resolvability in the sense of definition above: if an object of a given size is resolved then it is visible clearly enough (net necessary sharp but already cannot be confused with other objects); if the object is not resolved then it is too blurred to be perceived separately from other objects. To emphasize the fact that the object whose size we test is located in space of objects, H.M. suggested a special term "DISK of confusion"; it is by definition the minimum size in space of objects that can be resolved (or maximum size that cannot be resolved, -- it is the same), it is the threshold for linear resolution in space of objects. An example what resolved and unresolved objects look like can be seen at http://www.photo.net/photodb/folder?folder_id=75190 except very first photo. Yellow and orange blobs are leaves in foreground that were either smaller than DoC (unresolved, transparent even in the center) or greater (resolved, opaque in the center) but still not sharp. Please don't take it as self-promotion.

 

 

I still assert that the most technically correct way to describe the difference between the traditional and the HM's approaches is to say that the latter requires equal LINEAR resolution in space of objects at near and far DoF ends, and the former requires equal ANGULAR resolution. (If you're not sure please let me know, I'll post the proof, it is short). In particular, the threshold of angular resolution of traditional method is always = c/f (where c is acceptable CoC, f=focal length; and therefore the linear resolution in space of objects according _traditional_ approach is =s*c/f, where s is the distance to near or to far DoF end). Since resolution of human eye by its nature is an angular one, the traditional approach wins in most "normal" situations, but these are not all the possible situations.

[WAn]

"This reads as if there exists a circumstance, where one or more of my premises are not in effect, in which Merklinger's method is "clean".

Please answer this question: What is that circumstance where it is better to throw away all the DoF that resides beyond the plane of sharp focus by focusing at Infinity?"

 

 

It follows from my boring prelude above that the H.M.'s method is clear in situation when we have a strange requirement: linear resolution in space of objects must remains constant everywhere inside a given range of distances. How can one arrive at such a perverse objective? --- Consider the example: I'm standing on a balcony, there is a huge crowd below me, from my feet till 200, 300m... I�m a detective. I want to photograph all the people and my objective is: I must recognize on the photograph every face from 1m to 300m. In darkroom I won't mind to use any needed enlargement. I take the size of Disk of confusion ca 3mm, focus at infinity and push the button. --- I readily agree Mike, that you can calculate such a CoC that every face in this zone will be sharp when viewed from certain viewing distance. But I don't need any face to be razor sharp, IT IS NOT MY OBJECTIVE, not my mission! I want them to be recognizable only! And by the way, I suspect that making every face "sharp" requires narrower aperture that I needed for my dirty purpose ("sharp" means sharp for any enlargement I may want to use to recognize a face 300m away from me). --- I hope Mike you see what I call the "criterion" and why accepting this or that criterion can change the method. The example is rather unusual; this sort of mission is more suitable for technical (crime?) photography rather that for general landscape photography, but such situations are possible and legal.

 

 

Personally I see 3 virtues of H.M.'s method:

 

 

1) A convenient tool in unsharpness/sharpness control. Maximization of sharpness in a zone is only a part of this general task, even though a very important part.

 

 

2) Appropriate and natural tool when the mission is specified in terms of linear resolvability in space of objects.

 

 

3) The method does not depends on format, lens focal length, all I need to know is the diameter of aperture hole (and distances to subjects and their size, of course, at least approximately). This issue is rather an academic one, but it makes academic life easier.

 

 

I don't agree to consider this discussion as a process "Mike Davis vs. Harold Merklinger". I consider the two methods as two different tools that are intended for different tasks. Apples are as good as oranges.

 

 

About my notorious "beg to differ". I misunderstood your questions. You wrote: "So, please tell me how it's advantageous to focus at infinity and then stop down enough to achieve acceptably small circles of confusion at the near sharp, when those same DIAMETERS could have been enjoyed at a wider aperture (at both the near sharp and at infinity) were you focused at the hyperfocal distance." -- The word "diameters" I interpreted as DISKS of Confusion. Surely, the DISKS do grow unrestricted when focus is everywhere except infinity, hence the objection. --- It was my error, I sincerely apologize.

[MTC Photography]

Mike wrote:

"It's simply amazing that a man who wrote an essay titled "The INs and OUTs of FOCUS" could pretend to miss the fact that we can CHOOSE any CoC diameter we want for our Near and Far sharps and thus produce perfectly acceptable images while focusing at the corresponding hyperfocal distance."<p>

That is absolute lie !

<p>Merklinger specifically discussed about selecting smaller coc, in fact 1/150mm coc in his book<p>

 

 

"There is NOTHING wrong with focusing at the hyperfocal distance. You simply have to calculate DoF with a CoC diameter that's aggressive enough to suit the anticipated enlargement factor and viewing..."<p>

 

Ah, so convienient !?

For my requirement, I need a coc of 0.01mm instead of 0.033mm.<p>

According to St. Mike, I "have to" and "simply" "calculate" DOF<p>

Ok suppose I followed St. Mike's advice, and start to calculate

I usually use f8 on my Carl Zeiss Planer 50/1.4. <p> Opps, I don;t

have a calculator with me <p> Go home and find a calculator<p>

Yep, I got the hyperocal for f8, it reads 31.1 meter<p>

Unfortunatly, the dumb Carl Zeiss engineer did not anticipate

that some one like St Mike wanted 31.1 M marking on the dial.<p>

On my Planer, from 10 M to infinity, there is no marking in between<p>

St. Mike, sir, how do you put 31.1 M ?<p>

The next day, I happen to use my Carl Zeiss 35-135 zoom, I want to use a setting of 120mm @ f11. Oh, my god, according to St. Mike

I again "have to" "simply" calculate another hyperfocal distance...I diligently followed St Mike's instruction and plug the number in

my calculator, (which from now on, I must carry with me, to calculate

the hyperfocal distance of my zoom lens, which has infinite choice

of focal length) and lo, it is 131 meter<p>

I tried straining my eye to figure out where to put the 131 M hyperfocal distance on my lens ..."<p>

 

Next, I use the zoom at about focal length = 95 m, f8<p>

 

I "have to" "simply" calculate another DOF again ???<p>

 

<P> Forget it. If I follow St. Mike, I "have to" spent a lot of

time plugging on my calculator instead of taking pictures.<p>

 

<p> What the heck ! It is by far simpler to follow Merklinger put the lens at infinity and shoot.<p>

<p> It turns out St. Mike's that You "simply" and " have to" calculate DoF

is the most stupid method<p>