Locus of Focus

[Stasch]

It has been difficult to find out what constitutes a proper lens for photographing 2D artwork, say a painting about 16" x 20". One of the factors I would want is a lens that offers a sharp focus over the entire painting.

Of the opinions I have read so far, on various websites, a couple of statements stand out beyond the obvious requirements of good lighting and having the painting perpendicular to the axis of the lens. In particular:

(1) The artwork should fill the field of view.
(2) Primary macro lenses are best because they have sharpness and a "flat field".

A written definition of a "flat field" has so far escaped my search, but it probably means that the locus of the points that are in sharp focus form a plane perpendicular to the axis of the lens. (I would like to call such a lens a "planar" lens, but that seems to be a trademark name used by Zeiss instead of a technical term.) A natural guess for me would be that the locus of sharpest focus across the field of view would be a curved surface, perhaps the surface of a sphere about some point in the lens. That would be bad news for photographing a painting if the lens had a very small depth of field.

My impression is that depth of field numbers apply to points in focus near the axis of the lens. Surely it is important to know the 2D locus of (sharpest) focus as well as the 3D region of focus across the entire field of view. I would very much appreciate a reference to a good publication that deals with this.

Stasch

[Manfred M]

You mean like these:





I shot those two with my f/2 105mm portrait lens. Camera was on a tripod and set perpendicular to the paintings.

A fixed lens is better than a zoom because there will be little or no pincushion or barrel distortion. You will want to use a short telephoto to ensure you are far enough away to not cast any shadlows on the work you are photographing.

The most important thing is not the lens, but appropriate lighting that is very even and does not cause any reflections. I use two softboxes, as shown in the following diagram. The watercolour I was photographing was behind glass.




I wouldn't get too wrapped up on exotic lenses, they aren't needed. I would not bother with a macro lens. unless you are planning to do macro photography.


M

[Stasch]

Great photos of fascinating paintings. I was surprised that you would photograph an artwork behind glass. I gather that with the 105 mm f/2 lens you did not need to do geometric adjustments (other than cropping) after taking the pictures.

[Stasch]

Thanks for the link to the Kodak publication "Copying and Duplicating: ...". It would be good to find a book that is devoted to the digital perspective since I no longer use film.

[Mike Buckley]

It would be good to find a book that is devoted to the digital perspective since I no longer use film.

Light: Science & Magic explains why Manfred's lighting setup works and how to achieve the same results even if you don't have the luxury of space that his setup requires. The physics of light is the same whether using film or digital sensor, so I don't understand your comment about a need for a "digital perspective."

[Stasch]

My objection to older texts that include film was just a gripe ... I don't like having to thumb through pages of detail about how to choose the right film, etc. It would be more interesting to know what the requirements would be for a digital system that could produce a file that could be used for a first-class reproduction, even a Giclee quality print. Even though such a print is not my immediate goal, I would find it satisfying just to know the capabilities and limitations of digital systems, and their lenses.

Indeed the physics of light is constant, but what is that physics? My original question, about what is the locus of focus of a lens throughout the field of view, is part of the quest for such knowledge. Everywhere I look there is information about "depth of field", but nowhere do I find clarity about the locus of focus throughout the field of view. Does the depth of field change as one moves away from the center axis, etc.? I would very much appreciate a reference to the physics of light for camera lenses that answers such questions.

Stasch

[Mike Buckley]

Sorry, Stasch. I misunderstood. The book I recommended is fabulous for explaining how to photograph the various kinds of reflected light in different situations but it does not explain the stuff you are considering.

[Stasch]

Mike, I appreciate references to both the "best known practices" and to the "theoretical" side ... I would like to know as much as possible about both!

Stasch

[ajohnw]

A macro lens should have very low inherent distortion and high resolution. Planar probably just means flat field - sharp focus across the entire field. Optics generally have a problem called Petzval curvature which will explain why flat field is important for an even resolution across the field.

A macro lens is probably your best bet these days but they are computed for working of relatively short distances. At these distances the Petzval curvature needs to be low to meet the other needs. I'm sure when people like Nikon made a lens for everything they also did one for copying at certain reproduction ratios. I suspect one of these would be hard to find. The 200mm Nikon macro lens also had specified reproduction ratios but I don't think it will cover your needs. You might find out which one it was on this page under special purpose lenses part way down - it's late in the uk so off to bed. They may even make a modern version of it.

http://www.mir.com.my/rb/photography...rces/index.htm

John

[Stasch]

John,
Thanks for the pointer to reproduction lenses ... I am not familiar with them. It may well be that a high quality planar field of focus is beyond my modest means... but nonetheless I will be happier if I know that I have done a decent job of finding out what works best for me, and how it compares to a first-rate system. I will sit down later this evening and write up a summary of my findings, and what concerns me about them, and post them here ... I have recently purchased a grand old DSLR, the 8 Megapixel Canon 1D, Mark II, and I want to make reasonably sure that I choose a nearly best possible lens (modulo my budget) for photographing artwork with it. Perhaps I will need to use my venerable 4 Megapixel CoolPix 4500 for small pieces ... it seems to have a respectable macro capability.

Stasch

[Manfred M]

Great photos of fascinating paintings. I was surprised that you would photograph an artwork behind glass. I gather that with the 105 mm f/2 lens you did not need to do geometric adjustments (other than cropping) after taking the pictures.

Photographing behind glass makes a lot of sense. Removing the picture from the frame is a lot of work and risks damaging a valuable piece of art. It is far better to use a technique to shoot through the glass and save time and reduce the risk.

The only adjustments in the first two paintings was cropping out the picture frames. I had set up the camera on a tripod quite carefully and aligned things very well (lens perpendicular to the painting). I think I shot around f/5.6 or f/8. so the depth of field certainly took care of any minor "errors"; the images are sharp. Stopping down also remove the vignetting that any lens has before it is stopped down a bit.

As I said in my previous posting, you don't need a speciality lens. If you look at classic reprographic lenses, these were typically wide angle lenses as the material to be copied was placed on a copy stand, with the camera shooting straight down. Banks of lights illuminated the subject from the sides. The last time I used one, the lens to subject distance was around 30cm / 1 ft away from the subject. This is certainly overkill, unless you are taking pictures of something very small like coins or stamps; in a bit of a production environment.

Your old Coolpix 4500 would be terrible for this type of work; those lenses had pretty significant distortion. The 1D Mark II has a 1.3 crop factor sensor, which means the common (relatively inexpensive) APS-C lenses are not going to be an option for you. You might want to look at the used market for a suitable lens.

Just as a point of interest; the house in the picture in my first post was located in your part of the country; in Fergus, but was torn down decades ago.

[Stasch]

Let me add a bit more detail about what I see as one of the challenges to find a good lens for photographing artwork. The picture



shows two scenarios for the "field of focus", that is, the points in the field of view which are in focus. The first would be ideal for photographing 2D artwork ... a flat field of focus. The second one I suspect may be closer to the truth, that lenses tend to have a curved field of focus. If so, by focusing on the center of an artwork, there is the possibility that the photo slowly goes out of focus as one proceeds away from the center. It seems important to know how much curvature one has, if the second case holds, and how large is the depth of the focus at various points in the field of focus.

Recently I acquired a used 1D Mark II, and I would like to find the best lens (among those I can afford) to use for photographing artwork. So my first idea has been to see which lenses will actually allow one to keep the artwork in focus while pretty much filling up the field of view. Unfortunately, without knowing information about the field of focus throughout the field of view, I do not see any way to come to a good conclusion on theoretical grounds.

To take an example from data I gathered on the internet, suppose we have a painting with a 1 ft diagonal diameter and we are using a Canon 100 mm f/2 lens on my 1D Mark II (which has a crop factor of 1.3). To have the painting fill the field of view one would need to place the camera 3 ft from the painting since the angle of the field of view is 24 degrees. (3 ft just happens to be the minimum focus distance for this lens.) At that distance one of the online charts says that the depth of focus is 0.2 inches. If the field of focus was perfectly flat, and the depth of focus held for the entire field of focus (at 0.2 inches), it might be a challenge to get the camera perfectly aligned, perpendicular to the painting, to have the entire surface within the field of focus ... but this is theoretically possible. On the other hand, if the field of focus is curved, then almost certainly only a central portion of the painting will be in focus if one sets the focus to be perfectly sharp on the center of the painting.

For a painting with a 2 ft diagonal, using the same lens and camera, the camera needs to be at a distance of 6 ft, and the depth of focus is now 1.2 inches (according to the online chart). This sounds like a feasible situation if the field of focus is perfectly flat... I have my doubts about the feasibility if the field of focus is curved. In such case it seems the optimal setting for the lens would be to have it focus in the center on a point just a bit behind the surface of the painting, in such a manner that the corners of the painting are somewhat in front of the lens focus distance at their locations.

I suspect that experienced artwork photographers have found lenses that work for them by trial and error since there seems to be a lack of published theory to guide them. If that is the way things must be done, then of course I will accept it and depend on the advice of those who are experienced. But it would be nice to know for sure that I have not overlooked informative publications on these matters.


Stasch

[Manfred M]

I strongly suspect the reason you are not finding anything in your searches is because you are trying to solve a non-existent problem. Lens focus is not paper thin; depth of field will ensure that everything is sharp if you stop down to a reasonable level.

https://www.cambridgeincolour.com/tu...calculator.htm

Shooting a 100mm lens at 72" from the subject of f/8 means you will have a DoF of around 5".

[Stasch]

Indeed an excellent point ... for a distance of 72", I calculate that the distance from the lens to the corner of a 2 ft painting would be about 1.5" more than to the center of the painting, so a 5" depth of field would certainly take care of the possibility that the center of focus lies on a spherical surface. If going to f/8 (or higher) is the right strategy, then it seems like a f/2 lens does not offer an advantage over say an f/4 lens (which Canon does not make in the 100 mm size), at least for this kind of photography.

There seems to be possibly one assumption that you are making that I do not see how to justify, namely that the 5" depth of field applies across the entire field of focus. So far all I can find out about determining the depth of field is that the calculations assume that one is very close to the axis of the lens. Here is a quote from the Wikipedia article Depth of Field:

"Limitations
Most DOF formulas, including those discussed in this article, employ several simplifications:

1 Paraxial (Gaussian) optics is assumed, and technically, the formulas are valid only for rays that are infinitesimally close to the lens axis. However, Gaussian optics usually is more than adequate for determining DOF, and non-paraxial formulas are sufficiently complex that requiring their use would make determination of DOF impractical in most cases.

2. Lens aberrations are ignored. Including the effects of aberrations is nearly impossible, because doing so requires knowledge of the specific lens design. Moreover, in well-designed lenses, most aberrations are well corrected, and at least near the optical axis, often are almost negligible when the lens is stopped down 2–3 steps from maximum aperture. Because lenses usually are stopped down at least to this point when DOF is of interest, ignoring aberrations usually is reasonable. Not all aberrations are reduced by stopping down, however, so actual sharpness may be slightly less than predicted by DOF formulas."

This item is about half way down a rather lengthy article ... if the author means that the central DOF is usually a good estimate for the DOF across the entire field of view, that would be good to know. The author goes on to say:

"The lens designer cannot restrict analysis to Gaussian optics and cannot ignore lens aberrations. However, the requirements of practical photography are less demanding than those of lens design, and despite the simplifications employed in development of most DOF formulas, these formulas have proven useful in determining camera settings that result in acceptably sharp pictures. "

Does the author mean that the formulas are always useful in practical photography...I suspect not since the word "usually" keeps cropping up; and if not, it would have been helpful to know when they might fail to perform...

Stasch

[Stasch]

Thanks for the insights on photographing objects under glass. Sorry to hear about the shortcomings of my old CoolPix 4500 --- I had the idea they would be good for close up photos because of stories that they were very popular at one time for both photography with microscopes as well as with various long range scopes. (I also read that the addition of converter lenses was not so successful.)

I don't mind parting with $500 or so for a suitable lens for the Mark II... but I would want that to be the only lens I needed at that price... your positive experiences and fine samples of your work are very encouraging! However it seems to be curious that there is a clear split between those who prefer macro lenses and those who prefer non-macro lenses, and I would like to get to the bottom of what is behind this ...

Stasch

[Manfred M]

Stan: The reason the fuzzy words like "usually" come up is that in theory, all of the points do make sense, but they do not, really make any real world difference. The diagrams do demonstate real world issue, but what is throwing you off is the scale. I'm certainly no lens designers and the last time I had a real good look at any optical theory was during my first year university physics courses; too many years have passed for me to remember anything other than the must rudimentary details. I think all of the measurements were alway taken with respect to the optical axis of the lens, but with a complex lens, I just don't know where the other measurements would be taken from.

First of all, let's assume instead of a large picture, you are trying to take a picture of a postage stamp. Even with your 1:1 macro lens (this means the size of the image hitting your sensor is the same size as the real world object. Having a lens that varies by even one mm corner to corner would have an impact at that distance as the DoF will be quite shallow, certainlyly less than this, and even if your camera is perfectly aligned, you will get out of focus areas at the edges. This is why macro photographers are looking for lenses that have an extremely field and will shoot at very high f-stops, etc.

As you move further back from your subject, the DoF increases and the angle of light from the subject to the lens gets closer and closer to being "nearly parallel", so the theoretical limitations become less meaningful.

You seem to be quite concerned about the optical performance of the lens. I think you are forgetting something far more important, and a factor that is actually relevent in the type of photography you are looking at; camera alignment. Even if you have a perfect lens; you still have to line up your camera so that it is centred and parallel to the subject. The usual tool is a tripod, and frankly getting that "close enough" is important. All that being said, the last image I posted was hand-held, so I know that the camera wasn't perfectly aligned, yet the image had excellent edge to edge sharpness.

Another "real world" issue that you will likely come across is chromatic aberations from the lens. If you pixel-peep, you will find that pretty well any telephoto lens will display this characteristic, as the light from the different parts of the spectrum focus at slightly different places, and a bit of purple and green fringing will be seen if you look closely enough. Nikon has been doing in-camera correction of this for years, but only the very latest Canon cameras have this feature.

The reason for the various opinions of macro versus non-macro are really related to the question you posed; you wanted to know which lens would give you the flattest field. My point was that for the type of work you were looking at, the question is irrelevent to what you are looking at doing.

You could certainly use a macro lens to do this work, but my point was simply that you do not need to restrict yourself to a macro. I believe that Canon makes a f/2 100mm lens that would be around $500 and the 100mm macro can be picked up for around $600, so you might want to look at the used lens market. I find that Aden camera in Toronto usually has the best prices for lenses in your area and I've bought most of my glass from them.

[Stasch]

Manfred, many thanks for your detailed comments on my focus concerns. One more question about lighting, using two softboxes. I was planning on using continuous daylight fluorescent photography bulbs ... if you stop down to f/8 and stand 6 feet away, what wattage would you be using with two soft boxes?

[pnodrog]

Wait until you get on to the hard bit - reproducing the colours accurately. A few years ago I used to take photographs for a gallery for their catalogue. Getting the images evenly lighted, in focus and corrected for any obvious distortion was the easy bit. One of the gallery owners was a highly qualified art restorer and we had plenty of rather warm discussions. You could get the painting and the photo to match reasonably under one set of lighting then when comparing them under different light the colour shift between the two was significant. We ended up concentrating more on ensuring that the tonal contrast matched.

[Stasch]

Manfred, it just occurred to me that if one takes the maxim "fill the view with the painting", presumably to get the most out of the available pixels in the camera, then with a 100mm lens I would always be looking at having the picture occupy the full 24 degrees of the field of view. Moving back will get the rays from the object of interest closer to the axis of the lens, but only at the price of sacrificing the full use of the available pixels in the camera.

[Manfred M]

That solution is not going to give you very good results. Fluorescent tubes, especially the commodity ones that you pick up at the Home Depots of the world have are going to give you terrible colour reproduction. Daylight is really more of a marketing term than anything else and the spectral curves vary from manufacturer to manufacturer. You are very likely to get some spike in the green and blue parts of the spectrum and the light will be deficient in reds and yellows. I, like most photographers, try to avoid fluorescents for that reason.

I do know that companies like KinoFlo put out specialty bulbs that are reasonably good for photography or videography, but I understand that these bulbs are very expensive. I have no personal experience with them.