Lens testing - anything else to consider?

[ebogaerts]

I plan on running a test of three 50mm Nikkors of different maximum

apertures. I have come up with a list of factors involved in this

test. First of all, I plan on testing all three of the lenses

focused at 15 feet. For focusing purposes, I plan on using a tape

measure to ensure good focus. Where should I start the tape measure

at - at or close to the film plane? At the front of the lens?

 

All lenses will be tested with a Nikon F4, firmly fixed to a tripod

with the mirror locked up. This test will take place at or around 2

pm and naturally will involve daylight. The comparison between each

lens at each aperture will involve the same shutter speed

(i.e. "correct" exposure, or same EV). I don't imagine it will take

me more than 5 minutes to do all of the exposures involved with all

three lenses, so there won't be much variation if any with the light.

 

The subject will be a portion of the brick on my house. I plan on

taping up a section of newspaper that will encompass the upper left

quarter of the frame, so I can check the sharpness of the lettering

on that newsprint from one corner of the lens gradually to the center

of the frame.

 

I plan on using Tri-X film, developed in Microdol-X, and doing all of

the exposures on one roll of film. I use Tri-X more than any other

black and white film, so this is more of a "personal" variable, just

like the focus distance of 15 feet, which is typical for me.

 

However, are there any strong recommendations *against* using Tri-X

for this sort of test?

 

The scanning of each shot will take place on a Canon flatbed, which I

realize isn't the benchmark of quality, but once again - all of the

shots will be compared on the same scanner with the same settings.

Sections of each shot will be isolated and critiqued, such as the

corners, center, lettering on the newsprint, ect.

 

Is there anything else that I am forgetting, not taking into account,

or any other variable that I am forgetting that I could reasonably

control, given the average resources available to me?

 

Thanks for all help/opinions!

[kelly_flanigan1]

A 50mm Nikkor will resolve on film greater than my Canon FS4000US 4000dpi/ppi film scanner or Canon FS2710 2720dpi/ppi film scanner; or any of my Epson Flatbed scanners; at its best apertures. With a flatbed; you are placing a rough 30 line pair/mm filter over your lens test data. A sharp 80 line pair/mm negative of mine; shot with a Panatomic-X and Microdol; resolves only about 25 to 28 line pairs/mm; with my Epson scanners. This makes all perfect lenses appear like a poor lens test; for a 35mm camera.

[kelly_flanigan1]

Sorry for being a little negative; I have been shooting lens tests since the 1960's. <BR><BR>Most 50mm Nikkors peak about F4 to F8; in the central core; edges sometimes require more<BR><BR>A slower film will give better numbers; but one should aways do some tests with the emulsions YOU actually use too..LOng ago; Pamnatomic-X was used alot.<BR><BR>One dirty secret is that actual resolution varies a decent amount with exposure. Too much exposure can radically drop measured resolution.<BR><BR>

[dan_fromm1]

What do you want to learn from this experiment?

 

What do you have against direct examination of the negatives, ideally through a low-power microscope? Scans are trash. They're one of the reasons so many people conclude in error that film is no better than digital.

 

Tri-X isn't as sharp as your lenses. If what you want to learn is which lens to use with Tri-X at the distance and apertures you used in your trial, fine.

 

If you want to learn which lens produces better image quality under your test conditions you'd do well to use at least TMX, ideally TP. And if you want to learn how they'll do at other distances, well, you'll have to test at those distances too.

 

On the whole, what you propose to do seems like an uninformative waste of effort.

[ebogaerts]

O.k., then perhaps I won't make the attempt.

[stormfront]

Eric, it sounds like good fun to me. Why not just have a bash and see what the results are. As you say all three will be tested in the same manner at the same time and evaluated in the same way, so even if 15 feet is not the normal testing distance or your newspaper is the NY Times and not the LA Times and it is not the 1960s anymore... well who cares, the test should yield some value.

 

I am pitting my Canon EF 28-90mm f/1:4-5.6 Mk II against my new Canon EF 24-70mm f/2.8L USM just for the hell of it. Get a kick out of doing your own test, don't stress about "following tradition".

[david_h._hartman]

<em>"I plan on using a tape measure to ensure good focus.

Where should I start the tape measure at - at or close to the

film plane? At the front of the lens?" --Eric Bogaerts<br>

</em><br>

You have an F4. Forget the tape measure and use a DW-21, 6x waist

level finder. Test the camera for body focus. Then you will have

the most accurate focus you could hope for. The DW-4, DW-21, DW-31

are some of the best reasons to own the F3, F4 and F5. I would

not buy this finder for tests. Its amazing finder for macro

photography and sometimes use it with telephotos. <br>

<br>

<em>"All lenses will be tested with a Nikon F4, firmly fixed

to a tripod with the mirror locked up." --Eric Bogaerts<br>

</em><br>

Rather than using mirror lockup use speedlights, preferably a

pair, and use TTL flash in dim light. Focus with a large high

power quarts flood then turn it off for the actual exposure.<br>

<br>

I recommend using charts first then later practical subjects for

comparison at key apertures.<br>

<br>

<em>"I plan on using Tri-X film, developed in Microdol-X,

and doing all of the exposures on one roll of film." --Eric

Bogaerts<br>

</em><br>

For charts I recommend Tech Pan in Technidol. This is the closest

you can get to a true test of the lens not the film. For

practical tests Id use your favorite films.<br>

<br>

<em>"The scanning of each shot will take place on a Canon

flatbed, which I realize isn't the benchmark of quality, but once

again - all of the shots will be compared on the same scanner

with the same settings." --Eric Bogaerts<br>

</em><br>

This will reduce the test to one of the scanner. Borrow or rent a

low power microscope or a high power loupe.<br>

<br>

<em>"Is there anything else that I am forgetting, not taking

into account, or any other variable that I am forgetting that I

could reasonably control, given the average resources available

to me?" --Eric Bogaerts<br>

</em><br>

<em>"A slower film will give better numbers; but one should

aways do some tests with the emulsions YOU actually use too..LOng

ago; Pamnatomic-X was used alot." --Kelly Flanigan<br>

<br>

"One dirty secret is that actual resolution varies a decent

amount with exposure. Too much exposure can radically drop

measured resolution." --Kelly Flanigan<br>

</em><br>

Both of these are true. I prefer a test of the lens with the

highest resolution film I can use then practical tests with films

I will use. If you normally shoot Tri-X but sometimes shoot

Velvia a test on Tri-X will be valid only for Tri-X only.<br>

<br>

I dont much care about lines per millimeter. Ive only

calculated this once. I added an insurance photo of a 135/5.6

Componon-S on the end of a lens test. I was so surprised at how

sharp it was that I enlarged a small area and with the aid of a

needle counted the lines in the tooling on the aluminum barrel

then calculated 110+ lines/mm. The taking lens was a 55/2.8 AIS

Micro-Nikkor at f/5.6 with flash, probably an SB-15. This shot

was hand held.<br>

<br>

As I recall Tech Pan is able to resolve ??~200 lines/mm. Velvia

60~160 lens/mm. <br>

<br>

Most prime Nikkors will give the best results at f/4.8~5.6 on

high resolution films. When I test I want as objective

information as I can get. When I shoot I factor in the limitation

of the film. Some lenses may gain contrast while loosing

resolution while stopping down a little more. This may extended

the sweet spot towards f/6.3 or f/7.1 with lower resolution films.

You can sometimes see center resolution fall as edge gains with

test on very high resolution films.<br>

<br>

Remember sharpness can be important but its not everything.<br>

<br>

Regards,<br>

<br>

Dave.

[steve_levine]

anything else to consider? stamp collecting?golf?fishing?

[ebogaerts]

Steve - Those are good things to consider as well. :)

 

Thanks everyone for your advice so far. I went ahead and bought a roll of Tech Pan, and I will go ahead with the test. David's suggestion of testing both films is a valid one.

 

My only real motivation in doing this test is to see for myself what my samples of my 50mm lenses can do. I know these things have been done before, by people more qualified than I am. I'm looking to evaluate center sharpness, and I'm also planning on testing light falloff against a while wall, as well as bokeh tests wide open.

 

And yes, I do have a loupe. :) I was planning on using the scanner if I was going to "publish" the results on a webpage. If the scanner cannot replicate or even approximate what I am able to see in the loupe, then I won't even bother with it.

[imaginator]

Forget using the flatbed scanner. I would use slide film! Also, bracket exposures just to be sure. At least you understand the idea of keeping things consistent (controlled) and reducing variables. Make sure you shade the lens (no direct sunlight should hit the lens) Use manual focus (not distance scale and a tape measure) Check out Bob Atkins website for more info on lens tests.

[bacsa]

http://194.100.88.243/petteri/pont/How_to/ha_Testing_lenses/a_How_to_test_a_lens.html.

 

And i would agree about using a microscope (around 20x should do) instead of scanning.

[ebogaerts]

^ Actually, on my last post just above, I *was* going to say that I am going to evaluate center sharpness and the sharpness progressively towards the edges.

[vivek iyer]

In that event (centre/corner sharpness), you should rest the film for at least 30 secs after each forward before shooting. This will minimize the film curl as it comes out of the 35mm film spool. If you really want to be very sure that your film is flat, you will need a pin registry camera! If you need a "standard"lens for testing, it is the Micro-Nikkor 55mm f/3.5 at 1/10X and at f/5.6 (100 lp/mm).

[ebogaerts]

Thank you, Vivek! That's a variable that I was completely unaware of.

 

Is 30 seconds good enough? Would waiting a full minute be better?

[david_h._hartman]

Eric,<br>

<br>

Here is a low resolution sample of the chart I use.<br>

<br>

<img

src="http://www.photo.net/bboard/image?bboard_upload_id=16867684"><br>

<br>

You can download a 600 dip, Photoshop 7.01, Zip compressed TIF

file at the bottom of this thread...<br>

<br>

<a

href="http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=005mHf"

target="_new"><u>http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=005mHf</u></a>

<br>

<br>

Hope this helps,<br>

<br>

Dave Hartman.

[vivek iyer]

On film flatness: at least 30 secs- 1 minute is better.

 

Also, the "flat-field lens", Micro-Nikkor 55mm f/3.5 lens has a distortion of 0.5% (no vignetting at f/5.6).