PDA

View Full Version : Curious Minds – Image Circle, Circle of Illumination, ULF, and Contact Prints



Len Middleton
24-Mar-2013, 07:39
I did have a good look at the Large Format Home Page and a look at some of the other information available, so looking for some additional insight from the collective knowledge here.

Actually in my mind all the above items are related. I have an 8x20 camera (nominal diagonal of 21.5” / 550mm, and when looking at suitable lenses continue to be confronted by data sheet specifications that say you cannot use that lens for 8x20 e.g. 450 Nikkor M IC= 440mm, 480 Apo-Ronar IC=396mm, 355 G-Claron IC=442mm, yet are commonly used by other ULF photographers.

There is some confusion in the information available, versus my understanding. There seems to be information out there that implies the image circle and circle of illumination are one and the same e.g. Wikipedia and Ebony web sites.

• It is my understanding that the image circle at infinity is the diameter of the circle of acceptable image quality as defined by the lens manufacturer e.g. high enough MTF value (Modulus Transfer Function) and is defined at a specific aperture (usually f16 or f22).

• It seems to me that circle of illumination is a vignetting issue, where one can even use the optical equivalent to a “Coke bottle bottom” could illuminate, yet not provide suitable image quality.

Given that ULF are contact printed, then the lack of enlargement will reduce the relative image quality requirements, then image circle would seem to be less of an issue and it impact will be in the corners and one can burn those in…

So the questions are:

1. Is my understanding of the difference in image circle and circle of illumination correct, or is there something I am missing?

2. How much difference is there in definition of lens quality e.g. different MTF values, between manufacturers i.e. Nikon versus Rodenstock?

3. Given the difference in application i.e. reproduction / graphic industry versus general photography, is there difference in the definition of image circle e.g. MTF value for Apo Ronar versus Sironar? Bob throwing that one your way…

4. How rapid typically is the fall off in MTF at defined image circle and beyond?

I strongly suspect that I am not the only one lacking knowledge in this area.

Thank you for your insights,

Len

Dan Fromm
24-Mar-2013, 08:38
4. How rapid typically is the fall off in MTF at defined image circle and beyond?

Len, this depends on the lens and the manufacturer so there's no really solid answer.

Example one. Rodenstock 75.4 Apo-Rodagon D. The MTF plots here http://www.rodenstock-photo.com/mediabase/original/e_Rodenstock_Printing_CCD_43-62__8230.pdf (p. 61) are flat to the edge. Coverage has to be limited by mechanical vignetting.

Example two. Rodenstock 150/9 Apo-Ronar. MTF at 16 lp/mm is near zero at the edge. See https://skydrive.live.com/#cid=8D71BC33C77D1008&id=8D71BC33C77D1008!182 p.3-6

Example three. Schneider 150/5.6 Apo-Symmar XL. In the MTF charts here https://www.schneideroptics.com/pdfs/photo/datasheets/ApoSymmarL/ApoSymmarL_56_150_2.pdf MTF is falling fast at the edge of the field. Schneider no longer posts MTF charts for the 210/6.1 Xenar, but as I remember them MTF was near zero at the edge.

David Lobato
24-Mar-2013, 09:03
Image circle criteria is also dependent on the f-stop, why f22 typically accompanies an image circle spec. Stopping down further will gain you a little more coverage. It's because the edge of the circle goes from a blurred zone to a more sharply defined edge as the lens is stopped down. The useable image circle thus gets larger.

This partly explains why some photographers get away with using a lens considered too limiting for the larger format. If the lens is focused closer than infinity you also get a bigger image circle, and sometimes just a little bit more is all you need.

Nathan Potter
24-Mar-2013, 09:05
Len, here's my answer to some of the questions you pose. One needs to generalize a bit though.

1. Yes I've always taken the circle of illumination to be the inside edge of vignetting at a define aperture whereas the image circle at infinity is that radial distance from lens center where the resolution approaches a minimum acceptable by the manufacturer. Different manufacturers have different standards and AFAIK no one posts those standards.

2. For the top quality lens designs there is, on average not a large difference in resolution/contrast between the big four, Nikon, Schneider, Fuji and Rodenstock for similar focal length lenses. There seems to be a bigger variation within the same model than between different manufacturers for the same focal lengths. Most all quality normal focal lengths (90 to 210 mm) show resolution/contrast numbers from 40 to 60 lp/mm across the field of view.

3.As far as I know the definition for image circle between a general purpose LF lens and a reproduction lens is about the same. But I would defer to Bob for his comments. The MTF values are determined in a standard way so there should be little difference between MTF values as long as they are determined at the optimum magnification for that particular lens be it 1:1 or ∞.

4. With lenses I've looked at the edge of the image circle already has a drastic MTF falloff so as a percentage of the lens image circle radius I tend to utilize about 90% of the radius. I can tell you that near the point of vignetting some lenses show as low as 5 LP/mm for MTF50 criteria where at the listed image circle one could see only 15 to 20 LP/mm at MTF50. But the variation is huge at the peripheral region. My sense for this is based entirely with testing a number of lenses quite a few years ago for industrial photography. Those were aerial image tests on an optical bench using a scanning densitometer. For precise macro work on IC chips I tried to stay within the center 4/5 of the image circle.

Nate Potter, Austin TX.

Jac@stafford.net
24-Mar-2013, 09:27
Image circle criteria is also dependent on the f-stop, why f22 typically accompanies an image circle spec.

Is that true for later Biogons?

Oren Grad
24-Mar-2013, 10:34
Image circle criteria is also dependent on the f-stop, why f22 typically accompanies an image circle spec. Stopping down further will gain you a little more coverage. It's because the edge of the circle goes from a blurred zone to a more sharply defined edge as the lens is stopped down. The useable image circle thus gets larger.

This is the nub of the matter for ULF applications. You can't tell the whole story from the specification for f/22, because some lenses gain much more usable coverage than others as you stop down. Thus the apparent anomaly of lenses like the G-Clarons or the 450 Nikkor M, which have unremarkable specifications for coverage at f/22 but in practice are mainstays for ULF photographers.

Mark Stahlke
24-Mar-2013, 12:37
You can't tell the whole story from the specification for f/22, because some lenses gain much more usable coverage than others as you stop down.Is there a way to calculate the image circle as one stops down to smaller apertures? Just curious.

Oren Grad
24-Mar-2013, 13:41
Is there a way to calculate the image circle as one stops down to smaller apertures? Just curious.

No. It depends on the lens and on one's criteria for acceptable quality.

John Jarosz
24-Mar-2013, 15:32
Some process lens' coverage is quoted for a specific resolution (L/mm). They do that so a purchaser who will use it for process work will know exactly what to expect, and the numbers they quote are guaranteed. I use the 420mm Nikkor on 8x20 and the coverage allows movements. You'd never know that by looking at the specs. Contact printing 'art' photography has no relation to process camera work. Try the lens out, but you're right that it's hard to make a purchase without knowing if it will cover.

John

Jac@stafford.net
24-Mar-2013, 15:44
Is that true for later Biogons?

Bumb

Dan Fromm
24-Mar-2013, 16:30
Oh, come on, John, you have enough of the lenses to know the answer.

Jac@stafford.net
25-Mar-2013, 07:40
Oh, come on, John, you have enough of the lenses to know the answer.

I'm Jac. :) John is my evil twin, and an only child.

The biogons I use gain no coverage when stopped down, but I remain ignorant of variants of the so-called Biogon.

E. von Hoegh
25-Mar-2013, 08:13
I did have a good look at the Large Format Home Page and a look at some of the other information available, so looking for some additional insight from the collective knowledge here.

Actually in my mind all the above items are related. I have an 8x20 camera (nominal diagonal of 21.5” / 550mm, and when looking at suitable lenses continue to be confronted by data sheet specifications that say you cannot use that lens for 8x20 e.g. 450 Nikkor M IC= 440mm, 480 Apo-Ronar IC=396mm, 355 G-Claron IC=442mm, yet are commonly used by other ULF photographers.

There is some confusion in the information available, versus my understanding. There seems to be information out there that implies the image circle and circle of illumination are one and the same e.g. Wikipedia and Ebony web sites.

• It is my understanding that the image circle at infinity is the diameter of the circle of acceptable image quality as defined by the lens manufacturer e.g. high enough MTF value (Modulus Transfer Function) and is defined at a specific aperture (usually f16 or f22).

• It seems to me that circle of illumination is a vignetting issue, where one can even use the optical equivalent to a “Coke bottle bottom” could illuminate, yet not provide suitable image quality.

Given that ULF are contact printed, then the lack of enlargement will reduce the relative image quality requirements, then image circle would seem to be less of an issue and it impact will be in the corners and one can burn those in…

So the questions are:

1. Is my understanding of the difference in image circle and circle of illumination correct, or is there something I am missing?

2. How much difference is there in definition of lens quality e.g. different MTF values, between manufacturers i.e. Nikon versus Rodenstock?

3. Given the difference in application i.e. reproduction / graphic industry versus general photography, is there difference in the definition of image circle e.g. MTF value for Apo Ronar versus Sironar? Bob throwing that one your way…

4. How rapid typically is the fall off in MTF at defined image circle and beyond?

I strongly suspect that I am not the only one lacking knowledge in this area.

Thank you for your insights,

Len

As for image circle/circle of illumination, they are two different things entirely. Some lenses do have internal baffles so that the two circles coincide, but generally most lenses will illuminate far more they will cover sharply.
I use (on 8x10) an 8 1/4" Dagor type lens which at f:32 presents an aperture that viewed from the cut corner of the GG looks like a cat's pupil. The corners need to be dodged, but they do retain contact printable detail.
Certain manufacturers have very conservative coverage figures, quite a few lenses will give contact printable detail well beyond their rated coverage.

Len Middleton
25-Mar-2013, 08:29
Gentlemen, thank you for the insights.

Correction to my original note, in that MTF is Modulation Transfer Function, as modulus is an engineering materials issue from my distance past.

Dan, so from the examples you provide you are saying YMMV… Given the different common lens types used for ULF, which of the different designs i.e. plasmat (G-Claron), dialyte (Apo Ronar) and tessar (Nikkor M), falls off more rapid than another?

David and Oren, certainly aware of the impact of stopping down on coverage and using the more of the centre of the lens and less of the edges, but like Mark I am uncertain as to the extent, nor do I expect any means of determining, other than of course the ground glass and the print. And if we go too far down that road, then we run into diffraction impacting image quality.

Nathan, interesting regarding your comment on variation in MTF in focal length; then does it logically follow that the longer lens focal lengths have a less stringent MTF requirement? Given your requirements in the industrial photography environment, you have taken the opposite tact of what we are talking about here for ULF, in that you are essentially de-rating the image circle to address your stringent requirements for that application.

John, yes trying is the most certain way, but of course is time consuming and expensive i.e. Ilford HP5 in 8x20 cost over C$12 per sheet when I bought it last year. Hence the questions to try to understand and reduce the number of “dead ends” I need to travel down.

I will likely follow the common trail traveled by others, but as indicated in my thread title, I do have a curious (and analytic) mind.

FYI, for the 8x20 I do have a 355 G-Claron in a shutter, a 420 Apo Ronar in a shutter, and a 480 and a 600 Apo Ronar in barrels. The 420 Apo Ronar (IC=346mm?) will cover 8x20 at around 10 feet no problem, but need to try it for infinity, yet.

Dan Fromm
25-Mar-2013, 09:45
Dan, so from the examples you provide you are saying YMMV… Given the different common lens types used for ULF, which of the different designs i.e. plasmat (G-Claron), dialyte (Apo Ronar) and tessar (Nikkor M), falls off more rapid than another?

Len, if the lens doesn't have a field stop to limit coverage -- I think my little 80/6.3 WF Ektar does, that's why its claimed coverage is much less than that of other very similar 4/4 double Gauss type wide angle lenses -- MTF plummets towards the edge of the field covered. Whether that happens inside or outside of the coverage the manufacturer claims depends on the claim. Ain't no rule.

Apropos of "ain't no rule," for contact printing and for many situations where the negative is enlarged too discussions like this one are somewhat irrelevant. Contact printing doesn't demand as much image quality as enlargement and sharpness in the frame's corners isn't always important for the image. This last point -- who needs good image in the corners? -- has been made many times and bears repeating.

Well, come to think of it, there is a rule. People selling used lenses often claim greater coverage than the manufacturers did.

What this means for you is that you'll have to do your own testing. There's no other way. You can't trust anyone to know your requirements and standards.

If you can't afford to test, assume that the manufacturer got coverage right and and to be safe try not to set up shots with significant detail in the corners. You can't be sure that what looks sharp on the ground glass will look sharp on film or paper, you can be sure that what doesn't look sharp on the gg won't look sharp on film or paper.

Michael Graves
25-Mar-2013, 09:51
...If the lens is focused closer than infinity you also get a bigger image circle, and sometimes just a little bit more is all you need.

This is a trick I have used with success with my 120mm Fujinon. It has an image circle of 288mm, according to the specs. However, when I focused on an object that was only about 20 feet away, the image circle adequately covered an 8x10 negative. Then, when stopped down to f22, the depth of field was sufficient that the clouds on the horizon were quite sharp.

Just an added comment...different lenses do "circle of illumination" different. Some lenses seem to drop off gradually, and while the circle of adequate sharpness is (for example) 300mm, the circle of illumination might be significantly larger. Other lenses, such as my 7.5" Ilex Paragon, stay sharp until the illumination drops off like a rock. Therefore, there is little difference between the two.