My Theory on Comparing Lenses of Different Focal Lengths...Am I Crazy or Does This Make Sense?

[j._mose]

I have always felt that the lines per mm test is an unfair comparison

when judging lenses designed for different formats. Here is how I look

at it (whether it is right or wrong)....let's take three formats:

35mm; 2 1/4 square; and 4 X 5. They are matched with a 50mm, 80mm and

150mm respectively. Additionally, all three lenses yield an image

circle that matches the format size. When shooting with the three

formats: If all three shots are taken of the same subject, from the

same distance, what occurs? The 50mm lens is projecting the subject on

a small image circle (lets say 50mm) while the 150mm lens projects THE

SAME SUBJECT on a 150mm image circle. If the 50mm lens yields 150

lines per mm on a 50mm image circle; AND the 150mm lens yields 50

lines per mm (for example only) on a 150mm image circle; WOULDN'T IT

BE FAIR TO COMPARE

RESULTS BY ENLARGING THE SMALL IMAGE CIRCLE TO THE SAME SIZE AS THE

LARGE IMAGE CIRCLE? OR VICE VERSA...COMPRESS THE LARGER IMAGE CIRCLE

TO MATCH THE SMALLER IMAGE CIRCLE...THEN MEASURE LINES PER MM????

 

Doesn't this make sense or am I missing something?

 

Plus, a large format lens typically has a larger image circle than the

format size it is intended for...so now do the math!

 

If this point of view has any merit then it would make perfect sense

why test results, using lines per mm as the criteria, would be better

for the 35mm lens than a large format lens.

 

Any thoughts or comments?

[Troll]

The best way to compare lenses of different focal lengths, is to move the camera to different distances, so the image size remains the same on the negative. (IMHO)

[leonard_evens]

What you say makes sense, but I believe it is just the common understanding.

 

Here is the way I think about it, and again I believe it is also the common understanding. What is important is the final print (or other final image). You have to make some assumption about how large it will be and how it will be viewed. One common assumption is that a typical viewer looking at an 8 x 10 print from closeup---about 10-12 inches---can resolve about 5 lp/mm in the print. Now multiply that by the the degree of enlargement needed to produce that print from the given format. There is some fiddling necessary because of differences in aspect ratios, but if we take that to be 4:5 as a standard after cropping, then the enlargements are as follows: 35 mm - 8.5, 6 x 4.5 - 4.5, 6 x 7 - 3.6, 4 x 5 - 2.12. Thus one would need to start with the following resolutions in the source: 35 mm - 42.5 lp/mm, 6 x 4.5 - 22.5 lp/mm, 6 x 7 - 18 lp/mm, 10.6 lp/mm.

 

Of course, some people have better vision than others and might find the 5 lp/mm in the print too low. Some peole claim to be able to resolve 8 or even 10 lp/mm. That changes the arithmetic.

 

If you want to make a larger print, there are two ways to think about it. It is often assumed that viewers will stay about the diagonal of a print distance away from it. That means they will stay proportionately further away from a larger print than from an 8 x 10 print. If we make that assumption, the 8 x 10 figures still apply. On the other hand, some people insist on putting their noses as close to the print as they can get it---we call them grain sniffers. If you want to please such people, you have to change the enlargement factors. For example, for a 16 x 20 print made from a 4 x 5 negative, you would take an enlargement factor of 4.24, not 2.12 and seek a resolution in the source of 21.2, not 10.6 lp/mm.

 

It is important also to realize that the resolution in the source depends on more than the resolution of the lens. The most important other factor is the resolution of the film. Also, when you combine resolutions, you generally get a result LOWER than the resolution of any component. For example, if you use a lens with a resolution of 100 lp/mm with film with resolution of 60 lp/mm, you are going to get less than 60 lp/mm in the negative or slide. (The best way to estimate the net effect is to use MTFs, but there are a couple of common rules of thumb for doing it.)

 

Beyond this, there are other factors, such as grain, which may effect the result. But the upshot of it all is that large format lenses need not have the high resolution needed for 35 mm lenses to produce high quality results. In fact, modern large format lenses do quite well considering they are designed to accomodate camera movements and so have to have large circles of good defintion.

[leonard_evens]

P.S.

 

The title of the thread is a bit misleading. It is the format that is relevant, not the focal length of the lens. For any given format, you would use the same analysis for every focal length. Of course, if you keep the angle of view the same, you get different focal lengths for different formats, but all you need is the format size and some assumptions---as I indicated above---to make any calculations about resolution you may need.

[jim_galli4]

Line pairs per millimeter has very limited use for realistic comparison. The driving factor is the finished products requirements. It's been a long standing "rule" that you need about 7 lppm in the finished product to fool the human eye into perceiving the picture as sharp. So we do the math WITH the end enlargement in mind to see if our "system" can get us there. Most 35mm lenses are resolving 75 or 80 lines and then most users aren't even getting that because of sloppiness. We banty around these over 100 numbers but the fact is that those are flagship type lenses used on a tripod with perfect atmosphere. So it's not that hard to get to 8X10 with a 35mm neg and still look sharp. But what if you need 20X30? The math gets more difficult. You need a 120 lppm lens corner to corner to get 6lppm in that 20". On the other hand, there are many LF lenses that are in the 60lppm world and some that go over 70. So with a good 4X5 neg at 60 it's only a 5X enlargement to get to my 20" and I'll have over 11 lppm to do it with. Plus smaller grain. Plus greater and smoother tonality. LPPM doesn't mean a lot until you start to think about the math to get to an end result. What good is a 60 lppm 4X5 that goes into a magazine at 4X5". 10 times more info than required. But if a 40X50 print is desired, you'd still have 6 lppm which can seem extremely sharp at a normal viewing distance for a 40X50 print. LPPM in itself doesn't mean much until considering end results. My lawn mower engine only burns .5 gallon per hour but it won't get my pick-up down the road.

[donald_miller1]

What Jim is saying is true...In engines there is no substitute for cubic inches of displacement...in photography there is no substitute for square inches of film. That is why I shoot 12X20.

[robert_davis2]

But isn't there more then sharpness? The example of a 4x5 being over kill for a 4x5 print would imply contact prints are always over kill. Yet many of us think a contact print looks better then an enlargement.

[kelly_flanigan1]

Smaller formats tend to be enlarged alot more.<BR><BR>I have the same lenses that I can use on radically different formats. When the lens is used on the smaller format; NONE of the edge performance matters; that really doesnt hit the actual film.<BR><BR>Since smaller formats are usually enlarged more; they usually have tighter definitions for what is "in focus"..... MAYBE THAT IS WHAT YOU ARE GETTING AT?<BR><BR><BR>The circle of confusion for 8mm cine film is 0.0005 inch; for 16mm 0.0010 inch; and 35mm movie 0.0020 inch........Nikon F slr is 1/30mm .....4x5 press lens snf many large format; usually 2 arc minutes....<BR><BR>In the last thread; the poster was using 35mm film; and wanted to know is MF lens were OK it use sometimes. In thus situation; the format IS defined; thus the edge performace of an MF lens; that is NOT used when on the 35mm camera; doesnt matter.....Once the film format is defined; the user should care about the lens performance on THAT format; ie line pairs /mm.<BR><BR>Here I have a 178mm F2.5 Aero Ektar that fits on a 4x5 camera; when 8x10 is printed; the enlargement is only about 2X+.. I have the SAME lens on an Exakta 35mm slr mount; here an 8x10" print requires an 8X+ enlargement........Once we placed the lens on an 8mm cine camera; using C mount adapters; D-mount adapters.....; here the 8mm film gate is only 0.192 inch; thus a projected image that is 40 inches wide is magnified over 200 times!<BR><BR>In the three extreme examples above; the line pairs/mm is the only thing that matters. The SAME lens has been used on 4x5" film; 35mm film; and 8mm cine film........If the lens resolved say 20 line pairs/mm wide open; the resulting 4x5 enlarged to 8x10 would have a resolution of 10 line pairs/mm (GREAT) ; the 35mm negative enlarged to 8x10 would have a print resolution of 2.5 line pairs/mm (OK; but NOT great) ; the 8mm cine would be enlarged by 200X; so the resolution on the screen would be a 20/200 = 0.1 line pairs/mm ; ie 1 line pair/10 mm; horrible........Typical decent 8mm cine lenses resolve 60 to 80 line pairs/mm on the film><BR><BR>Using line pairs/mm is the proper way to test lenses. diffierent formats are enlarged by different amounts; with smaller formats ususally enlarged more.......<BR><BR>The DOF scales also will change; when a lens is used on a smaller format; which has stricter standards. The standards for circles of confusion for each format were standardized in Kodak literature in the World War 2 era.........smaller formatshave tighter criteria.............for the Kodak Ektra; the same critera was used for sharpness; when the foacl length changed; this is because the film format is the same; but the focal lengths varied............<BR><BR>Kodak literature shows stricter sharpness criteria; when a 150mm lens is used on press cameras; 35mm; or 16mm cine.......the smaller formats have tighter criteria; they are enlarged more..

[kelly_flanigan1]

<img src="http://www.photo.net/photodb/image-display?photo_id=1387989&size=lg">

[kelly_flanigan1]

Very few 35mm slr lens will resolve 100 line pairs/mm on film......I have tested alot of good glass for 3 + decades; using USAF targets; cameras on tripods; perfect timed lighting (no shutter bounce); focus verified; and Panatomic-X film (now gone)...........<BR><BR>Probably only one out of 2 dozen 50mm lenses even reached 85 line pairs/mm; most peak at about 60 to 75; on the better lenses..........Some off named cameras (not Leica Nikon or canon )only reached 40 to 50 line pairs/mm........<BR><BR>My old Kodak 127mm 4x5 press camer lens resolved 80 line pairs/mm at the center; on film....<BR><BR>Many quote 100 line pairs/mm on film like it is an everyday event; which really requires lab test conditions; B&W sharp film; and testing several Summicrons; and varying the exposure......<BR><BR>Most name brand lense today really only resolve about 50 line pairs/mm ; with todays 400 color films; on a good day...when used in the field...

[pvp]

Jim Galli got it right, and I think his response is what J. Mose was trying to say: what's important is the end result, and enough resolution is .... enough.

[jim_galli4]

"My old Kodak 127mm 4x5 press camer lens resolved 80 line pairs/mm at the center; on film"

 

Kelly lives in the real world. I've got a cupboard full of old single coated tessars that we use on our 35mm movie cams. That's .75X1.00 inch image. We know that a good tessar can put GREAT pictures on a piece of film that size. So with an 80mm tessar I'm just using a little bit out of the sweet spot in the middle. We also use a lot of Pentax 6X7 lenses on the movie cams. Even better. Would I use a 127 f4.7 on 4X5? No because down in the extremes you'll be at like 15 lppm.

[ralph_barker]

I agree that it is the final print size and the anticipated viewing distance, and thus the degree of enlargement, that should be the controlling factor in such considerations. How one goes about doing the math depends on how one approaches the problem. I believe, however, that there are a couple of factors that complicate the process of trying to reduce the question to a mathematical analysis.

 

First is the apparent difference between what the average person's eyesight can "resolve" at the anticipated viewing distance, and what they can "percieve" beyond that limit. This gets to the heart of the motivation to use large format, I think. While we commonly use the range of 5-8 lines/mm as the measure of what the average person can "resolve", it seems we can "percieve" differences well beyond that. Hence, a contact print from an 8x10 neg generally "looks better" than even a 2x enlargement from a 4x5 neg, even though the 4x5 neg's resolution typically exceeds the mathematical criteria for an 8x10 print.

 

Secondly, it is the aesthetics of the final print that really matters, I believe. This is the territory of "artistic vision" that tends to obviate the usefulness of the math. For example, some lenses that may not meet the resolution requirements that one might calculate, still consistently produce images that have knock-your-socks-off aesthetics. We often refer to this as the "signature" or the "finger print" of the lens - a quality that seems to defy mathematics.

[dick roadnight cotswolds]

A fair way to compare lenses would be to calculate the total resolution - lp/mm * Image circle area.

 

When what we now call large format was all there was, they assessed all formats on the assumption that the end result would be a 10 * 8 print - now it is not logical to use LF unless you intend to make big enlargements.

 

It is difficult to make a lens with a 500mm image circle that can give you as many lp/mm as a 35mm lens, (70-100) but that should be the ideal to which they strive.

 

If you photograph a group of hundreds of people, you would like to be able to produce a sharp 10 * 8 print of each face in the picture.

[pvp]

<I>If you photograph a group of hundreds of people, you would like to be able to produce a sharp 10 * 8 print of each face in the picture.</I><P>

What? No enlargements? :o)

[robert_a._zeichner1]

A couple of thoughts:

 

First, apparent sharpness is the product of resolution and contrast. Lenses with higher resolving power, but that lack good contrast often appear to be less sharp than average resolution lenses with good contrast. That kind of blows the theory of using the resolving power of lenses alone as a yardstick for sharpness.

 

Second, in order to achieve a given level of resolution on a negative, it is necessary to have film and glass with individual resolutions that are much higher. There are a couple of ways to calculate this, of which I offer one: The net resolution of any lens and film is equal to the product of the two divided by the sum of the two. That would mean that if you had a lens and film, each with resloving power of 100 lppm, the net resolution of the two would be no greater than 50 lppm. (100 x 100 = 10,000, which divided by 100 + 100 [200] = 50. Leonard, I believe you know the other formula which is a bit more generous in its results, but the point is the same: Film and glass with equal resolution results in something that is significantly lower in resolution.

 

When Kodak developed the orginal Technical Pan for aerial surveilance, Perkin Elmer was developing a lens capable of making the film useful! The 36" f3.5 9-element weighing 400 lbs. in its mount was the result. Resolved golf balls at an altitude of 40,000' from a U-2, railroad ties at 80,000'

[kelly_flanigan1]

Robert ; the USAF target is of a high contrast; when several lenses are compared to the same target; the test does give one data point for lens A,B & C ; which can be compared............Other tests can be done with standard targets of a lower contrast; again for lenses A, B, & C.......Many different tests; done with targets of different contrast; and many different resolution group/pairs; involves alot of work...This is why MTC curves are done; whioch gives the whole story................<BR><BR>The film does influence the resolution pf the lens film combination..For decades; panatomic X was used for 120 and 35mm films; now almost none exists.....I have some rolls in my freezer............<BR><BR><b>The resolving power of Panatomic-X was quoted as 60 line pairs/mm ; during World War 2; in the kodak literature; with D-76...Plus-X was 55; Super-XX 45; and Kodak Micro File a LARGE 160 lie pairs/mm!..All these figures are with a subject contrast of 1 to 30...</b>

[wieslaw1]

J. Mose - the answer to your question is the following: yes, you need to compare identical, or at least similar angle of view (focal length) of one lens vs. another. Introducing different film formats as a variable means that you need, say, a 70° lens in 35mm to compare with a 70° lens in a large format. Then for fair analysis you use the same magnification for both formats. You enlarged 10x the 35mm negative, do the same 10x magnification for a 4x5. Assuming we have top quality lenses, the large format lens will always be inferior, in terms of sharpness or resolution, as compared with the small format lens. In simple terms try to enlarge a portion of your 4x5 negative with a 35mm enlarging lens and compare it with a 35mm/35mm print.

 

 

Donald �In engines there is no substitute for cubic inches of displacement...in photography there is no substitute for square inches of film. That is why I shoot 12X20�

 

- You ARE WRONG ON BOTH ACCOUNTS.

 

In photography it is the mind, eye and talent of the photographer that counts! NOT the negative size.

 

And the talent of a designing engineer resulted in improved small capacity, effective engines. (Unless you are in the military, the taxpayers foot the bill, and you need a tank to win oil fields)

[kelly_flanigan1]

With a 35mm focal length lens on a 4x5" camera; the lens is darn close to the ground glass. The PC Nikkor has a bigger than average image circle; it is a shift lens......Th image appears as a big circle; on the 4x5 negative.........With a 50mm enlarging lens; it makes a macro lens; when focused close; and is very sharp...<IMG SRC=http://www.ezshots.com/members/tripods/images/tripods-298.jpg><IMG SRC=http://www.ezshots.com/members/tripods/images/tripods-296.jpg>

[ellis_jordan]

It seems to me that macrophotography is not the subject of this discussion.

[kelly_flanigan1]

The 35mm PC Nikkor focuses at infinitiy; of the 4x5 above; thus it not a micro lens..........

[scottelly]

I wondered myself about the validity of arguments raging about the quality of 35mm lenses vs. large format lenses. I say that it doesn't matter so much how many lines per mm the lenses make on the film. I think it matters more - how many lines the lenses are capable of making. How many total lines is a 150 mm 4x5 lens capable of making vs. a 50 mm lens for a 35 mm camera?

 

Here's the math:

 

150 mm lens for 4x5 camera

50 lines per mm times 170 mm (minimum image circle) = 8,500

 

50 mm lens for 35 mm camera

100 lines per mm times 50 mm? (minimum image circle?) = 5,000

 

my numbers may be a little off, because I don't know where to find the image circle numbers at the moment, but I think you get my drift.

 

I think this explains why lenses for 8x10 cameras are so expensive. They have to be precise enough, and probably more importantly - large enough, to create a much larger image with aproximately the same resolution (lines per mm). Ultimately, I think the wavelengths of light are our main limitations in photography of all types, and large format will ALWAYS win when it comes to those physical limitations. At least for our lifetimes.

 

This is just a newbie's perspective. I welcome challengers.