What do you think of my Lens Test?

[s_p]

Months ago I promised other forum contributors that I would come

up with my own lens test. I have been critical of lens testing

in my postings in the past; I still think a lot of these tests

really skew the results because of the way in which they are

done.<p>

I think at this point I have come up with a test that works

pretty well and doesn't require a lot of fancy equipment like

lasers or $3000.00 collimators. I have written a description

located in <a href="http://www.icon-stl.net/~stefan/

sworn_enemy.html">http://www.icon-stl.net/~stefan/

sworn_enemy.html</a>.<p>

If you just want to see my numbers, go to <a href="http://

www.icon-stl.net/~stefan/results.html">http://www.icon-stl.net/

~stefan/results.html</a>.<p>

Some of the results suprised me. Conventional wisdom states

that MF lenses are less capable of rendering extreme detail than

35mm; in other words, a MF camera depends heavily upon its

larger film size for the better prints you can expect from it.

In my test, my 80 f2.8 MF tested about the same as 50mm primes

for 35mm at smaller f stops.<p>

I would be interested in others views of both my test and the

results.<p>

stefan

[mike dixon]

I thought your tests were interesting, and the results made sense.

[My personal bias is that the maximum resolution of camera lenses has

far more importance placed on it than is really warranted.]

 

<p>

 

The possible flaw that came to my mind is that perhaps the resolution

on your good enlarging lens might be the limiting factor in evaluating

the results from your best lenses. Was the enlarging lens used at its

optimum aperture?

[tim_brown]

I did a series of lens tests on the Mamiya TLR lenses:

http://db.photo.net/bboard/q-and-a-fetch-msg.tcl?msg_id=0005l3

 

<p>

 

There are earlier tests posted showing the refinement of my technique.

 

<p>

 

In order to compensate for focus errors (I found that I couldn't

discern exact focus in the viewfinder) I made 3 staircase targets out

of cardboard and rearranged the USAF stripes into strips (crash course

in postscript). Each of 7 strips is one inch closer than the last. I

focus on the center strip then look at the negative for the sharpest

strip. One target is in the center, the other two in opposite corners.

I can tell that the target plane and camera are parallel if the same

strip (say, one closer than the center strip) is in focus at both

corners. I can see field curvature by comparing center vs corner

strips.

 

<p>

 

I found that Tech Pan gave a peak resolution of 90 lp/mm vs 71 for

Tmax 100. I used multiple strobes set at low power (short duration) to

safeguard against vibration even with a tripod. I used a nice Zeiss

microscope (at work). I don't think enlarger alignment is a factor if

you're focusing on each target but still I could not read over 64

lp/mm with my Componon-S 50/2.8 and Rodagon 80/4.

 

<p>

 

In conclusion, I had to be far more careful than in normal shooting to

reach the full potential of my lenses.

[scott_eaton]

I think the biggest reason I support lens testing such as the ones

Stefan did is there is always something you will learn other than the

true optical performance of the lens. Issues such as enlarger lens

quality, halation effects in dense areas of negatives and so forth

came to light (pardon the pun) even though they were not the main

things tested for.

 

<p>

 

Many points raised here are things that have greatly concerned me

about photographic reproduction, but many not many are taken

seriously by novice or even professional phoographers. For instance,

I've given up on the threads where half a dozen posters claim they

get better quality out of their neg films by over-exposing their film

two stops, or not considering the quality of enlarging lenses when

evaluating camera lens sharpness.

 

<p>

 

I liked this test because unlike point-n-drool mags like Pop-Photo,

Stefan used relative testing as oppsed to unrealistic statistic

testing. Experienced eyes and familiarity with equipment is a better

judgent tool than "anechoic room" numbers. Shutter-bug and Pop-Photo

will rarely do side by side lens testing because it might infuriate

the manufacturers that advertise in their magazines.

 

<p>

 

Some comments on Stefan's tests:

 

<p>

 

-I'm still in the crowd that believes that brand name, fixed 35mm

lenses are better than MF lenses. Not so much because they are better

over-all, but because 35mm optics don't vary as extreme as MF lenses

do. A fixed Canon, Leica, or Nikon 105mm is going to be a "killer"

lens regardless of which one you use. A 90mm Hassie, Mamiya, or

Bronica is going to be a shot in the dark depending on age, series,

what number/letter is on the lens, etc. Also, the most common

problems I see with MF lenses are geometric distortion, lateral

chromatic aberation (color fringing) and lack of contrast. These are

very difficult to test for. Perhaps a test in which off-set lights

were pointed into the lens during testing would be interesting.

 

<p>

 

-Shutter vibration is less of an issue with 35mm cameras than mirror

based MF cameras, because, obviously the mechanism is much smaller

so you have less vibration. The now defunct Modern Photography did

some testing on this several years ago and the results were

impressive. I think for their testing they set the cameras on top of

a thick sponge and pointed them at the test target. Also, you really

have to get beyond 50mm to start seeing shutter vibration problems.

The FA, being a great performer for vibration dampening should

seriously out-class the Pentak K-1000 - the latter sounds and behaves

like a shot duck hitting a lake.

 

<p>

 

-Most enlarging lenses suck and 9 out of 10 that I've used in

commercial darkroom situations were so bad I refused to use them on

my own work. This is seriously a case of hype being greater than

reality and generally it's a case of you get what you pay for. The

mentioned 80mm Nikkor is one of my favorites and a top-notch

performer - better than the 50mm Rodenstock or Schneiders I've used.

Also, the problems that minilabs face is not the quality of their

lenses, but the fact that they get banged in and out of the print

housing and get "out-of-focus" real easy. Very, very few minilab

managers are smart enough to recognize the problem or how to fix it.

A commercial machine package printer at a pro-lab may contain 100

lenses or more, presenting a logistical nightmare to keep all those

lenses in focus. BY FAR, and by quite a marging the best enlarger and

lens I've used for custom printing was the Leica Focomat V35

 

<p>

 

-As density increases in a negative film you lose sharpness - we can

thank halation for that. Conventional B/W and color neg all behave

differently as density increases and different types of neg film will

behave differently. My favorite Fuji NHG color neg film has very

limited over exposure lattitude for emulsion sharpness. Kodak Ektar

emulsions are better and B/W chromogenic films like TCN or XP-2 are

better yet.This would be difficult to test with slide film because

the high-density areas are shadows that don't contain much detail.

Having a different emulsion structure all together than E-6 or C-41

Kodak Kodachrome excelled in this category.

 

<p>

 

good tests---

 

<p>

 

//scott

 

[gene_crumpler6]

Stefan:

 

<p>

 

Good work! Welcome to the world of the sharpness nut. I think(as

you concluded) that the open flash technique is the best at getting at

the performance of the lens, without any other influences. I was a

little suprised at the performance of the hassie lens, some what less

than I expected considering all the "Hassie Hype". In my tests, a

1970's 80mm Mamiya Sekor C series lens performed a little better than

your 80mm, specifically at F8 (mine was 72lp/mm). Maybe I ought to

give that old lens a little more respect!

 

<p>

 

The sharpest lens that I own, the 55mm micro nikkor receives a 72++

rating at f11 (my chart only gives me 72lp/mm). I have a 16x20 B&W

print of the Effiel(sp?)Tower at night that is pretty darn sharp( I

was looking at it last night). It would have been sharper but I

forgot a cable release and was stuck with 1 second on my FM at f2.8.

If I could have used the 55mm at f11 for about 30 seconds, it probably

would have been even sharper.

 

<p>

 

I also agree that the 80mm El Nikkor is a super lens. Since using it,

I'm thinking about upgrading my old 50mm f4 El-Nikkor to a new f2.8. I

also sold my 75mm f4 El-Nikkor after getting the 80mm.

 

<p>

 

As to mirror and shutter vibration, I've not done any objective tests

of my p67, but I have concluded that I need to use 1/250 as a minimum

speed when hand holding my 55 and 105 lenses. I've seen slight

degradation at 1/125 with or with out MLU. Same with the mamiya C33.

 

Keep up the good work!

[s_p]

In my description of my test I may have been a little hasty in

condemning my other enlarger lenses; does anyone know of a good

way to test enlarger lenses? I tried the 135 Schneider again; it

seems pretty good but I suspect an alignment problem. One way to

test this, I guess, is to photograph a flat paper covered in

stripes or graph paper, develop the film and see how sharp and

straight the lines are from edge to center. Anyone have any

better ideas?<p>

To answer the question regarding whether I used the best aperture

of my enlarger lens; I put the negs back in and looked again at

all apertures; it made the image darker but I would not change my

judgement of the results. Maybe I would if I had that 100x

microscope.

Part of my problem is that I really don't have the time or

inclination to invest in darkroom stuff right now; I am currently

interested in making lightjet prints instead of tray process

prints but would just like to know how well this stuff works for

when I move into a new place where I can build a darkroom.<p>

The hasselblad suprised me, not because I thought it did badly in

the test, but because I thought in fact it did very well. It

tested right around the better 35mm at smaller f stops and fell

off as I opened up. I had been told that MF lenses were awful

compared to 35mm but this seems to indicate otherwise. For the

record, my Planar is the older 80 f 2.8 C lens, not the CF. I am

told the CF version has 1 or more extra elements to improve

performance at larger apertures. Sadly, I sold my Rollieflex to

buy the Hasselblad. I wish I had been able to keep them both but

needed the interchangeable backs and Polaroid capabilities of the

Hasselblad and couldn't afford 2 MF cameras.

[scott_eaton]

>>Anyone have any better ideas?<<

 

<p>

 

The simpliest (and very effective) way to align an enlarger that I've

found is to not worry about what's in the center of the frame - just

the edges. You need a 35mm carrier that's been filed out and a 11x14

or 16x20 sheet of photographic paper that was cut from the factory

(flip it upside down and use the white side). Get a 35mm piece of

exposed film that is fairly dark around the edges and shot with a

high quality body like a nikon, canon or leica.

 

<p>

 

Simply project the image onto the the sheet of paper, focus it, and

center it until the image is just a few mm inside the the borders of

the photographic paper.

 

<p>

 

I think from here it's pretty obvious what's going to happen. Simple

geometry; You know the film is rectangular, you know the piece of

paper is perfectly rectangular, there-for the projected image should

be perfectly parallel with the photographic paper from both vertical

and horizontal edges. If the projected image isn't perfectly square

you know the film board isn't parallel with the base board. If the

lens wide open projects an image that is sharper on one side more

than the other you know the lens isn't in axis with the film or base

board.

 

<p>

 

This is just as accurate as playing with spirit levels, lasers,

gravity detectors and the other voodo stuff I see listed. This

technique is just as good. Remeber to try it with the enlarger head

at different heights. A razor blade in the lens carrier can also help

to focus on.

 

<p>

 

Oh yeah, glad to see you're playing with the light-jet. Cymbolic

Sciences is a serious player in digital and they wrote the back on

absolute high end digital film reproduction.

 

<p>

 

//scott

 

<p>

 

Remember

[roger_urban3]

Color me simple and color me practable.

 

<p>

 

Personally, I would find lens tests much more useful IF next to the rsults, the tester included visual results. Say, for example, an 8x10 or larger picture, along with a portion of the picture blown up 100x or something large so we can all see the grain and detail and relate it to what 72 lines per millimeter means tangibly. Then, we'd all have something to relate the numbers to, rather than abstract numbers.

 

<p>

 

If anything, it would either strengthen or weaken the machismo of lens resolution arguments. One might determine that for all intents and purposes, it doesn'tt really matter if one expensive lens produces 72lines per mm while another one does 10% less. And, cheaper medium format lenses might visually out perform very expensive 35mm lenses, and arguments that 35mm optics are superior looses force in the practical world.

[gene_crumpler6]

Rodger;

 

<p>

 

I've been told that 6 lp/mm on a print should look sharp at normal

viewing distances. If that is the standard, and one assumes that the

enlarging lens does not introduce any significant degradation, then a

72 lp/mm lens on a 35mm negative can produce a 12 x18 sharp print.

From my own experience, I would tend to believe that is a good

approximation.

[roger_urban3]

Gene,

 

<p>

 

That's sort of my point. See, you wrote "I've been told". If you were shown what it looked like you'd have more certainty over what was acceptable for yourself, rather than comparing abstract numbers.

 

<p>

 

Perhaps it's testing is inherently flawed because there are so many variables involved: Lens MTF, film used, developer used, enlarging lens used, etc.

[s_p]

Roger;<p>

The illustrations would mean nothing in comparison to your own

lenses unless you did the test yourself exactly as I did mine.

They would also be a great deal of trouble to provide since in

order to show such fine detail I would have to enlarge the tiny

(about %5) portion of the 35mm neg that contains the test chart

(most of the picture is blank wall) hundreds of times. To give

you an idea of how tiny the chart is, pin a 6x6 card to your

wall. Put any focal legnth lens you care to on your 35mm.

Multiply the focal legnth X45 and place your camera that far from

the wall in mm. No matter what focal legtn you use, the card will

always be the same size(that is the best thing about this test).

It will also be very small in the frame. In order to give you a

jpeg on which you could read the grain, I would have to create a

HUGE file. I don't know if jpegs (even huge ones) could show you

the photographic grain very well. In order to see it, I had to

raise my enlarger head all the way up and view the test target

with a grain magnifier (which has -- I dunno 10x magnification as

well? more?). Even then, making out the really fine lines was

difficult.<p>

If you simply must see the film, drop me a line and I'll snip a frame

a frame and send it to you.<p>

The point of this test was not to prove that my lenses were

better than yours. The point was to see how well my stuff works.

I still get pictures that I am not technically happy with; I am

trying to narrow the causes. I discovered that of the 6 35mm

lenses I tested (4 that I use and 2 as controls), the 4 I use

come out about the same. My test showed me that even a little

bit of overexposure causes a greater loss of resolution than

switching from an A+ lens to a D- lens.<p>

For the many people who have been sending emails asking, the film I

used was

Tmax 100 in tmax developer 1:4 at 68 for 10 mins. I'm sure there are

finer grain films that would extend the test further; I don't think I

need to go there. The test chart is the one on page 54 of Ed

Romney's book that I mentioned in the text. Your mileage may vary.<

p>

[gene_crumpler6]

Rodger;

 

<p>

 

As I said, my experience confirms that math. I've actaully printed some of the lens charts, to see how they look on a test print as large as I can blow up(16x20 plus a little bit). A 35mm lens that tests out at 72 will produce a very sharp 11x14 print and if the conditions are all perfect, a 16x20 will be acceptable sharp. A lens that tests at 40, will produce a 11x14 that is noticable unsharp. Both of the examples have. I took a 105mm f2.5 nikkor (the lens that produces 40 wide open)( was trying to save weight)to England in January and I ended up shooting it wide open most of the time and all of the 11x14's I printed were not very sharp. I made one great shoot on that trip of the London Bridge with my 28mm (72+) at f11 (the sun was out that day) and I have a sharp 16x20 with depth of field from about 8 ft to infinity.

 

<p>

 

My field experience bears out my test results.

[roger_urban3]

Stephan,

 

<p>

 

First off, let me say I applaud your efforts. It's something that I have been meaning to do with my own lenses.

 

<p>

 

Second, I didn't intend any of my postings to sound like a "your lens is better than mine" or vice-versa, and apologize if they were ambiguous enough so that someone intrepreted them that way.

 

<p>

 

Third, I see your point on the size of the 16x20 JPEG. What about a small 2x3 inch section then?

 

<p>

 

Fourth, I didn't realize some of the logistics you are encountering, such as lens to subject distance, although I am familiar with enlarger height limitations. When and if I ever get to do my own tests, I'll probably look at using a Schneider 40mm APO Componon HM or perhaps a 28mm Rodagon to help get byond the enlarger height problem....if I can't do wall projections.

 

<p>

 

 

Gene,

 

<p>

 

Can you possibly post a small sample of your results so the difference between 72 and 40 lines/mm can be seen?

 

<p>

 

Thanks Fellas,

Roger

[harald_gaunitz]

Hi,

 

<p>

 

I think also, the resolution is only one but a very important point of

a lens. But what is happend when harsh lighting and how are unsharpe

areas handled if it's combined with spot lights. Many zoom lenses will

add a picture of the aperture blades someware at the film. Many

so-called

high quality lenses are very poor in that point of view. They can add

some light corona around all point lights especially if those are at

the

edge of the frame. So-called mirror spots can be created also if the

light

is placed out of the frame. (Can be eliminated by a compendium)

My experience is that the best lenses today are made by Fuji. They

have

found out an overall compromise together with an extremely high

resolution.

Mamiya lenses I think are also very good. Many lenses to 35 mm cameras

are very good but the most outstanding I have found the Hexanon at the

compact camera Konica Hexar is.

 

<p>

 

The old classic brands from e.g. Germany maybe were the best 30 years

ago

but I don't think it's useful to believe one can live on old success

for

ever. But Rodenstock is my friend in the darkroom. Rodagon 80mm/4.0

serie N

and G I think is in the same class as the Fuji lenses.

 

<p>

 

Harald

[ryszard_stasinski]

Stephan,

 

<p>

 

Nice to hear that you don't negate lens testing anymore, but I admit

that I liked many of your former statements, too. Namely, IMHO results

of lens tests should be taken with a grain of salt.

 

<p>

 

My main complaint is that in practical situations, except for

shooting of an object having very small DOF from a tripod loaded

with a sandbag at 1/250s (I exagerate a bit), lens contrast is more

important for subjective 'feel' of lens performance than lens

sharpness. Simply, contrast does not degrade with minor shake, and

is clearly visible even in the smallest pictures, including the

image on a camera screen.

 

<p>

 

On the other hand, what we are doing (me too) is measuring the tiniest

details for which lens can exhibit any sharpness, and hoping that

this tells us usefull things about lens performance for reasonable

detail sizes (nobody is looking through a loupe on 20x25 prints).

Clearly, we hope that if lens A has still some resolution at

72 lp/mm, while lens B at 64 lp/mm, then lens A is better than B

for 0-20 lp/mm, i.e. in the range that is the most important for

subjective lens evaluation. Fortunately, MTF lens curves are usually

very regular, especially for simple lens, hence, our guess is

usually correct.

 

<p>

 

BTW, there is a quirk with lens testing that undermines all my belief

in MTF curves and the like. Namely, MTF plots for some complicated

lens (e.g. zooms) are on a par with those for simple lens (see

photodo.com). At the same time many people, including me, insist that

prime lenses are better performers than zooms, my opinion is that

the greatest difference is in contrast. I have done comparison

(purely subjective) for small apertures (5.6 or so), and infinite

distance (distance of MTF tests). Could anybody competent comment

on this failure of MTF tests?

 

<p>

 

Ryszard

[brian_c._miller]

I found what claims to be a very close approximation of the USAF 1951

lens test chart at the <a

href="http://www.ans.com.au/~chrisb/photo/technical/index.html">Techni

cal Photography Home Page</a>. If you bring the file up in a text

editor, it has some documentation in it for modifying it for your

printer output variances. (Could someone tell me what the correct

length should be for the first line set in the bottom right? Or any

of the line sets?)

 

<p>

 

What I would like to know is how to use this chart. There doesn't

seem to be any reference on how to read the spiral of lines and

numbers.

 

<p>

 

From what I've gathered from reading this thread, the camera is

placed at a distance, measured in millimeters, computed from the lens

focal length multiplied by 45. So the proper distance for a 50mm

lens = 2250mm = 7ft, 7-/4in.

 

<p>

 

Then what?

[gene_crumpler6]

Rodger;

 

<p>

 

I have no scanner or access to a scanner, so I can't give you anything.

 

<p>

 

Sorry

 

<p>

 

Gene

[s_p]

A follow up on my enlarger story.<p>

Recently I found a like-new used Beseler 23C11 enlarger with 6x6

and 35mm carriers for a bargain price to replace my worn out

Omega. I do not yet have the lens board for 50mm lens, but the

80mm works great. The difference is night and day. The problem

wit the Omega was that the gears/wheels/etc., that adjust focus

were so worn that they would not stay tight and replacement parts

were unavailible at any price.<p>

My initial tests confirm my suspicions that a lot of my

dissatisfaction with print quality came from enlargement. I have

only tested the 80mm enlarging lens since that is the only lens I

have a lens board for at this time. Previously I stated that I

was less satisfied with my 50mm lens; I will test it again after

I get a board for it and may change that opinion. I suspect that

with the 50mm any flaws from misalignment of the enlarger would

appear worse than with the 80mm since the 50mm projects a larger

image.<p>

[michael_buchmeier]

Your links are no longer operating, are the tests still available?

 

"The requested URL /~stefan/sworn_enemy.html was not found on this server."