Locus of Focus

[Stasch]

Lovely photo, Ted. I have zero knowledge about LED lighting ... I seem to recall reading that it was used with some microscopes, and perhaps from this I had the impression that it was only for small objects. My range of sizes will be about 6" to 4 ft.

Best,
Stan

[Manfred M]

Lovely photo, Ted. I have zero knowledge about LED lighting ... I seem to recall reading that it was used with some microscopes, and perhaps from this I had the impression that it was only for small objects. My range of sizes will be about 6" to 4 ft.

Best,
Stan

Stan - Ted deals with photographs of watches, so his LED lighting requirements (size and cost) are going to be a tiny fraction of what would be needed for the work you are planning to do. The LED panels used for pro video production are more along the size you would required and they are certainly out of most people's price ranges.

http://www.vistek.ca/store/ProVideoL...-daylight.aspx

As for a "cheap" full frame camera on the used market; that is still going to be pretty costly. It will be interesting to see what happens with the Canon 5D Mk I prices when the 6D hits the market in a few months. There don't appear to be any bargains in the Nikon line right now.

[Stasch]

John, sorting out the best choices for lighting will surely be a challenging project ... if halogen desk lamps work, it would be a very economical solution. I was playing a bit this afternoon with my current setup, taking a photo under 5200K fluorescent lighting, viewing it on my monitor, and then printing off a copy. It brings back memories of a few years ago, with the math poster project and all the trouble colour matching caused. I expect it to be every bit as difficult this time around, and I will almost surely settle for a compromise ... hopefully and agreeable one within an agreeable timeframe. At least I want the pictures to be nicely in focus.

[Stasch]

Manfred, Indeed spending over $2,000 for an LED light array is not what I had in mind... I have finally decided that it is time to stop asking probing questions (for the time being) and start making some decisions on how to proceed. It has been wonderful having such an attentive and well-informed audience to discuss these matters with ... one of these days I may even have enough experience under the belt to be on the other side of the fence ... Of course I do look forward to reading other threads ... there is no joy quite like learning new things!

[Stasch]

John, many thanks for your input ... with all the excellent responses I have enough material to keep me busy for quite a while, so I am going to stop probing for good ideas and start sorting things out and making decisions ... But probably later on, after I have run into actual problems, there will be a reason to return to the forum with more questions... It seems that there is a lot of the "art" of photography in the practice of photography, and I have so much to learn...

Stan

[xpatUSA]

Stan - Ted deals with photographs of watches, so his LED lighting requirements (size and cost) are going to be a tiny fraction of what would be needed for the work you are planning to do.

As for a "cheap" full frame camera on the used market; that is still going to be pretty costly. It will be interesting to see what happens with the Canon 5D Mk I prices when the 6D hits the market in a few months. There don't appear to be any bargains in the Nikon line right now.

Just to be clear, I was talking about "cheap" APS-C size sensors, not full-frame. I bought a re-furbished Nikon D90 from B&H a few months ago for about $600 U.S. I quess Canon APS-C would be quite a bit more than that, but if Stan already has Canon lenses . .

A D90 just this minute sold on eBay for $479 (2K shutter count) and there's some up New for ~ $650.

Perhaps Stan just needs a good white or gray card for his purposes, especially if he intends to shoot RAW?

[xpatUSA]

I have zero knowledge about LED lighting ... I seem to recall reading that it was used with some microscopes, and perhaps from this I had the impression that it was only for small objects. My range of sizes will be about 6" to 4 ft.

4 ft, eh? So, there could be quite a distance from the lighting to the subject. 8 feet or so (~ 2m)? Maybe four of my KolorOne S8816 LED 40 deg floods costing about $65 U.S. would do it but, as already said in this thread, at that distance there is now a trade between lamp lumens and camera exposure involving a bit of calculation.

I just went and measured one of my LED floods with a lux meter which, as you know, measures light falling on a subject. At about 8 ft away it read 100 lux in the middle and about 80 lux at the edges of a 5-6 ft circle. 4 lamps would give 400 lux on the subject which is about office lighting level or sunrise/sunset on a clear day (Wikipedia - lux table).

The ones I bought were size PAR30 floods with a 40deg lighting angle. There's a report on them here:

http://www.greenelectricalsupply.com...M79_Report.pdf

I also wrote a comparison with Halogen lamps here:

http://kronometric.org/article/lampComp/mas.html

I chose the S8816 because it has a good built-in diffuser.

There's a bigger version, 3500K color temp, here:

http://www.greenelectricalsupply.com...lbs-3500k.aspx

Report (with spectrum) here:

http://www.greenelectricalsupply.com...M79_Report.pdf
.

[ajohnw]

Some would agree with that Stan, big pixels too. There is some mileage in it actually as well. Small pixels tend to need higher minimum iso for shorter exposure times due to noise build up in the sensor. Are the newer pixel counts better? Pass. At the other end of the scale on comes diffraction disks - the ultimate limit on resolution. The spot size in the disc goes up as F numbers increase. Some feel that the latest 5D has reached the limit already. As photo optics aren't generally diffraction limited and fall well short they might be correct but there are other factors as well such as bayer masks and anti alias filters. One things for sure. It's possible to pay for more pixels than are ever likely to be really used especially on compacts. Dlsr's show signs of going the same way.

John

[Stasch]

Ted, the D90 is a camera one of my relatives has, and I have long admired their photography with it. The issue of which format is best is very interesting. When looking for a good used camera, the "bonus" that sold me on the 1D mark ii was the flexibility in the bracketing options. I have little experience in shooting RAW, but the thought occurred to me that it might have the advantage of creating files that can be better adjusted later on, when I do have more experience in color matching....

I am winding down my participation in this thread ... there have been a lot of fine suggestions for me to think about, and thanks for participating.

[Stasch]

Ted, thanks for the detailed pointers on LED and halogen lighting ...

[Stasch]

John, the details of the physics of sensors are certainly items that I want to know more about ... I can see many days of learning to look forward to.

[Manfred M]

Stan - the images that I posted were taken with a D90, so it does work well, and frankly I find that bracketing options are somewhat over-rated as shooting on manual will certainly allow just about any exposure bracketing you want to use. My D800 has all kinds of bracketing options that look interesting, but most of the ones that are "speciality" ones will affect jpeg images only.

I'm in the school that a larger sensor is probably going to be more important that the megapixel count, as the image will need less magnification from a larger sensor, and the less you magnify, the lower the impact of any lens flaws.

Your Canon 1D Mk I APS-H has a 28.7mm x 19.1 mm = 548 sq mm 8.3 MP sensor. If you went to a D90 APS-C, with its 23.6mm x 15.8mm = 373 sq mm at 12.2MP. This means that your camera's sensor has 1.47 times more surface area than the D90 sensor, and 1.47 x 8.3 = 12.2, so in effect the two sensors going to give you similar resolutions, but with your current camera, you will have to enlarge the image less, meaning less sensitivty to enlarging any effects from lens aberation effects.

[Stasch]

The idea of using (number of pixels) x (area of sensor) as a measure of the capability of a camera body is quite interesting ... I haven't seen that before.

I meant to get back to you with a comment on measuring the deviation of the corners of a painting from the plane of focus when the axis of the lens is not properly lined up perpendicular to the painting. Actually the calculation is simpler than I thought. Suppose the camera is at point P and focused along the lens axis on the point C on the painting. Suppose the line CP (along the axis of the lens) makes an angle q with a perpendicular to the painting. Let A be any point on the painting ... we ask, how far is A from the (ideal) focal plane? The answer is: at most

(length of AC) times sin(q).

For a given angle q and length of AC, this value is actually attained when the line AC is perpendicular to the line of intersection of the painting and the focal plane. Of course if A is on the line of intersection, it is in the focal plane, and the distance to the focal plane is zero.

Perhaps I am daydreaming, but it seems that, for a nice flat painting, there should be some simple, quick and inexpensive way to line the camera up within a degree or two of the perpendicular. If the angle q is 2 degrees, then sin(q) is 0.035. For a painting with a 2 ft diagonal, if the point of focus P is the center of the painting, then the corner of the painting will be at most 0.42 inches from the focal plane.

[Manfred M]

Without a custom built setup, you'd be lucky to keep things within double that angle, in three different axes. I have yet to see a tripod that has anything better than a bubble level built is to set things up, Certainly I've never see anything resembling a protractor for alignment on any axis. At least better video cameras have a hook so that one can manually pull focus. The best a still camera does is indicate where the focal plane is.

[xpatUSA]

Perhaps I am daydreaming, but it seems that, for a nice flat painting, there should be some simple, quick and inexpensive way to line the camera up within a degree or two of the perpendicular. If the angle q is 2 degrees, then sin(q) is 0.035. For a painting with a 2 ft diagonal, if the point of focus P is the center of the painting, then the corner of the painting will be at most 0.42 inches from the focal plane.

In photography, it can be self-misleading to study a single factor to the exclusion of others. Don't ask how I know that.

In this case, an overriding factor is the good old depth of field (DOF) - there is a good tutorial on this site and, better yet, a DOF calculator - more of which in a sec.

So let's say your 2ft diagonal painting is in the golden ratio then the vertical might be 520mm and say you decide to hold an APS-C camera in the portrait position (not the same aspect ratio, I know) 1300mm away. It has a 60mm lens. A bit of trig. and the top of the painting shows up as 26mm further away than the center. Horrors! you say but you heave a sigh, set the lens to f/5.6 (it's sweet spot) and take the shot anyway. To your surprise, the image looks quite good even with the camera misaligned by quite a bit more than 2 degrees.

Depth of field - the photographer's friend: https://www.cambridgeincolour.com/tu...calculator.htm

The calculator says that the DOF is a whopping 100mm (4") for a 60mm lens set to f/5.6 at a distance of 1.3m. So your 0.42" would not be observable, depth-wise :-)

Here's a post I did elsewhere about shooting a watch at an oblique angle, FWIW:

http://forum.tz-uk.com/showthread.ph...Depth-of-Field
.

[ajohnw]

The general suggestion for aligning set ups like this is a spirit level. That can get the painting vertical and the lens horizontal. The other angular error is more difficult but within reason could be corrected with software. If I was feeling manic I would have say a straight 2m bar set square to the camera lens on the tripod and use a tape measure. Do check an spirit level you buy by reversing it both horizontally and vertically. The bubble should show the same error either way or better still if the surfaces are accurate exactly the same reading.

Personally I would use circles of confusion based on pixel sizes for true sharp depth of field within the maximum capability of the camera used. In practice it's probably possible to use a 1 maybe 2 stop slower aperture to gain more depth of field but a lot depends on how good the lens is and detail in the subject.

-

[Stasch]

Manfred, It still seems feasible to me that one can get the camera very close to the perpendicular line through the centre of the picture (I will call this the "axis of the picture", unless there is a better name for it). As for getting the correct focus on the centre of the picture, that is another matter altogether, depending on the quality of the autofocus or the ability of the photographer to work with manual focus. I have no clear idea about the errors on that.

As for the axis of the picture, perhaps you have different constraints in mind than what I will be dealing with. For me, the object is to photograph new artwork that is not in a frame. Here is a possibility. I can set up an easel with a nice flat board on it on which to place the artwork. First I find out how far back to set the camera so the artwork fills the field of view. Since I probably don't want to touch the artwork (especially if it is done with pastels), I mark the centre of the artwork on the flat board. As an example, to get within two degrees of the axis of the artwork with the camera at 3 ft distance means I have to locate the centre of the lens within 1.25 inches of the ideal point. If I put a mirror on the flat board, over the centre point for the artwork, and cover the mirror except for a 1inch hole over the centre point, then, by standing back and looking at the flat board with one eye open, if I move my head to a position to where I can see my eye in the mirror, then my eye is within 1/2 inch of the axis of the artwork, no matter how far back I stand. If I can put my eye so close to the axis of the artwork then surely I can put the center of the lens just about as close...

[Stasch]

Ted, indeed there are a lot of factors to master for good photography. I had run up against colour in another context (a poster project), so I know that one is going to be tough-sledding. Since I am not in a rush to take some photos of artwork, I wanted to enjoy the luxury of tackling one item at a time, starting with what seemed the easiest, getting the picture into focus. Without that, the ballgame is over.

Thanks for the links ... your oblique shot of the watch came out very clean. When photographing 2D artwork, there seems to be a golden rule that one is supposed to photograph it as nearly perpendicular to the lens axis as possible ... perhaps that is so one can make prints that fairly accurately reproduce the artwork. So I understand why Manfred talks about how he carefully sets up so as to be square with the artwork. Manfred did point out in one of the early messages in this thread that stopping down solves the DOF problem ... then I wondered why anyone would want a fast lens for photographing artwork if the first thing one does is to stop down to f/5.6 or f/8.... Also I wondered how likely it was that a nice sharp DOF of say 4 inches in the centre of the artwork would still hold near the edges given that the photographer will likely not be so lucky as to put his camera lens exactly on the axis of the picture (so the "plane of focus" will be at an angle to the artwork), and add to that the likelihood that the field of focus is not a nice planar slab, but has some curvature (which is not documented by the manufacturers evidently). I suspect that the focus will not be uniformly sharp across the picture ... it may look sharp on the camera monitor, on a computer monitor or a letter-size printout, but will it be sharp printed at full size... There seem to be a number of photographers who like to use macro lenses to shoot artwork, even when it is not close up. Do macro lenses give a better quality photo, say one that can be used to turn out bigger quality prints than a normal prime lens? If so, why not shoot photos with a macro lens? If not, why not.... So many questions....

Best,
Stan

[xpatUSA]

Thanks for the links ... your oblique shot of the watch came out very clean. When photographing 2D artwork, there seems to be a golden rule that one is supposed to photograph it as nearly perpendicular to the lens axis as possible ... perhaps that is so one can make prints that fairly accurately reproduce the artwork. So I understand why Manfred talks about how he carefully sets up so as to be square with the artwork.

I imagine that artwork shots should be taken as square as possible to minimize distortion of the outline from an "unwanted perspective" point of view. I was just talking about depth of field and it's effect on the apparent focus of the image.

Manfred did point out in one of the early messages in this thread that stepping down solves the DOF problem ... then I wondered why anyone would want a fast lens for photographing artwork if the first thing one does is to step it down to f/5.6 or f/8....

If you read up on lenses a bit you will learn that a wide-open lens shows aberrations at the edges of the glass. This effect is reduced by stopping down (not "stepping") the lens. Most lenses have a "sweet spot" where the image quality is "best" (a compromise between aberration and diffraction). It can be anywhere between f/4 and f/11 depending on the particular lens.

Also I wondered how likely it was that a nice sharp DOF of say 4 inches in the centre of the artwork would still hold near the edges given that the photographer will likely not be so lucky as to put his camera lens exactly on the axis of the picture (so the "plane of focus" will be at an angle to the artwork), and add to that the likelihood that the field of focus is not a nice planar slab, but has some curvature (which is not documented by the manufacturers evidently). I suspect that the focus will not be uniformly sharp across the picture ... it may look sharp on the camera monitor, on a computer monitor or a letter-size printout, but will it be sharp printed at full size...

Firstly "nice sharp" and "DOF" are somewhat exclusive terms, IMHO. In my watch example, I was trading sharpness for DOF in the first two pics. My guess is that any DOF you get will hold to the edges of your pic for all intents and purposes. The DOF part is where you obtain an axial range of focus that will appear to be constant as far as the image plane (sensor) is concerned.

Do macro lenses give a better quality photo, say one that can be used to turn out bigger quality prints than a normal prime lens?

My feeling is no. Unfortunately "quality" is not a well-defined word. Neither is the term "quality prints". The advantages of macro lenses for other than close-up photography are best addressed by folks other than myself.

Good luck in your endeavor!

[Manfred M]

Stan - I think you missed my point here. The answer is really very much in line with having to deal with the slight curvature of the focal plane of the lens; DoF of the lens solves both problems as long as the misalignment is small. Any minor misalignment, for instance keystoning, can be corrected in PP.

PP software will nicely do pixel based interpolation, and you the photographer can tweak until you get a result that you are satisfied with; understanding of course that no copy will be perfect. You get to define "good enough".