Lens Testing

[Editor's Note: I am preserving this post and related .TIF materials in this archive...]

[My understanding is that the USAF chart etc. is in the public domain as it is a government produced chart and related documentation...]
From: David Chien [email protected]
Newsgroups: rec.photo.equipment.35mm,rec.photo.digital
Subject: Brief Lens Resolution Tests - Film Cameras
Date: Wed, 08 Dec 1999

(cross posted to rec.photo.digital - Of note to those digitizing film, see PhotoCD resolution and slide scanner resolving capabilties below.)

Ricoh R1
Minolta 600si w/28-200mm Tamron Super
Canon AE-1P with 1.8 50mm lens
Heavy Bogen tripod
Pictures set off by 10 second timer delay.
Auto AE/AF (around 5.6f-8f for most images)

Fuji SuperG 100 speed film of 12 frames, expiration date of 10/2000. Rated to have HIGH resolution in magazine, corresponding to maximum capability of 100-150 lines of resolution.

Modern Photography lens resolution test using USAF chart from article in magazine from years ago. Download at:

http://users.erols.com/johnchap/lenstest/lenstest.htm

[Editor's Note: see below for text at that site, and archive links...]
Backyard setup with direct sunlight hitting light gray brick wall at

30x Radio Shack pocket microscope used to view negatives.

============================================
Notes: A review of the Olympus Stylus Epic by one of the photography magazines earlier said the lens rated Excellent to Outstanding, with a tested lens resolution of just under 90 lines/mm. This is around 2x the resolution of any lens below, and closer to those expensive Leicas, etc. of high praise.

An interesting point is the resolution loss of the 28-200 is not as great at the far and near zooms. One would think then that because the center resolution is somewhat better than a pocket P/S (besides the excellent Epic above), subjects are usually centered, and nothing of note is usually to the side of them, the lens is 'good' enough for most people because it has a sharp center, and the edges at min/max are close enough to the center resolutions of a regular P/S that most people will not notice any significant resolution problems with the lens (ie. looks just like my P/S picture or better -- thus, it's a great lens! says the average user).

Also, for pictures of people, couples, etc. - eg. for weddings, you will get a center image close to what you'd expect from a 50mm fixed lens in image resolution, so unless you've worked with far better lenses, the 28-200 will seem similar to your old equipment, albeit with greater edge resolution falloff.

I'd love to hear others doing this same test and see if they can come up with a list of results, esp. for the 28-300 tamron and the Ricoh GR1.

My initial concern was the sharpness of my Ricoh R1. Clearly, it is 1/2 as sharp as the Olympus Stylus Epic, and even more so than the sharper rated Ricoh GR1 lens (guessing from Ricoh's GR1 MTF chart - which looks more like a Leica lens test than a P/S lens test). Close to an average SLR zoom lens or 50mm, but with significant edge resolution falloff (also noted in the magazines tests of the R1).

Of note, another site tested the resolving capabilities of PhotoCD by submitting a slide of the resolution chart. Test resolution was around 50 line pairs/mm. The images scanned onto PhotoCD from film was also noted to be of higher quality, sharpness, resolution than any currently available as well.

Note that for a 2800 dpi slide scanner, this corresponds to about 110 lines per mm, or about 55 line pairs/mm for maximum resolving capability. 4000 dpi slide scanners will do around 157 lines per mm, or about 78 line pairs/mm -- still far lower than the resolving capability of Fuji SuperG 100 speed film. (ie. that's why the pros drum-scan negatives at 10,000 dpi+ - around 200 lp/mm.)

============================================
M  = Center of frame
MM = Midway between corner and center of frame
C  = Corner of frame

============================================
Ricoh R1
Readings off a couple frames show:
30mm lens mode
M  = ~50 line pairs/mm +/- 5
MM = ~45 +/- 5
C  = ~30 +/- 5

24mm lens mode (one frame reading)
M  = 56 lp/mm
MM = 45
C  = 25

============================================
Canon AE-1P w/50mm f1.8 (two frames)
M  = 63, 71
MM = 56, 56
C  = 56, 56

============================================
Minolta 600si w/28-200 Super Tamron
@ 28mm
M  = 64
MM = 50

@ 50mm
M  = 50
MM = 45
C  = 35

@ 200mm
M  = 63
MM = 50
C  = 50

============================================

[Ed. note: here is the text from the above URL on lens testing archived here:]

Because I am probably changing ISPs, this site will likely terminate in a few days. Look in the rec.photo.xxxx newsgroups for an announcement for its next location.
You are looking at this file because you have some interest in checking the resolution of your lenses. The ability of a lens to resolve lines per millimeter (lpm) is not the only factor in lens quality. High resolution is only a necessary, but not sufficient, condition for lens quality. Contrast, distortion, flare, physical construction, etc. are all important factors. However, a lens that cannot produce decent LPM figures can never be a really good lens.
Comments and/or questions should be emailed to me at "[email protected]". I will try to respond to them individually unless the answers would be of wide interest, in which case I will post them as well.
The files listed below, which are available on this site, are scanned copies of the resolution chart and instructions published by Modern Photography, which stopped publishing nearly 10 years ago. In those years it was a competitor of Pop Photography. The lens testing kit contained original charts, instructions, and a loupe of about 26X power. I trust making copies of the instructions available is not deemed a copyright violation since the "owner magazine" no longer exits, and this is a labor of love and not profit.
FYI, in one of the last few issues of Pop Photography an article referenced a test chart currently being marketed, if you have any interest. I think the same chart might have also been mentioned in Shutterbug as well. At any rate...
To obtain actual lines per millemeter figures for your lenses you should follow the instructions exactly. While I base my test procedures on the instructions, I have modified and simplified them to facilitate my testing.
First, I include a known lens of high resolution in every test roll to serve as a benchmark. If the results for the benchmark lens are consistent across test rolls of film, then the balance of the tests on those rolls are probably valid. For me this lens is the 50/1.4AI which I shoot at 1.4 and 2. Note that in certain lighting conditions a very high shutter speed (1/8000?) might be necessary to test at 1.4 . Secondly, I use TMAX 100 developed in Dektol. Slow high resolution slide film is easier to use and read, but it is much more costly and time consuming than developing TMAX in my darkroom. I understand that Pop Photography is converting over to using Tech Pan film for its resolution tests. I have an old piece of masonite that I hang on my garage door. On it I have mounted the center chart and the two upper corner charts set the correct distances apart. I don't bother with the lower corner charts. All I do is hang the panel on the door, back up far enough so that the two corner charts appear in the corners of the viewfinder and then shoot. This places the lens sufficiently close to the exact distance from the chart that the instructions specify. My own experience is that a 20X loupe is barely sufficient for examining the charts. I normally use a 30X. I consider any lens that can resolve at the center the 0,1 group or smaller to be a sharp lens. Corners need to be at perhaps the -1,4 or smaller to be sharp there. I generally only test at max aperture and max+1 aperture since that is where nearly all my shooting is done.
Never test using AF, unless you are simply testing the camera's AF capability. Manually focus only. Use a MF camera if possible, since the bright focusing screens of a modern AF camera are not as precise. Third, use a focussing magnifier if your camera allows its use. By doing all these things you are eliminating variables/factors that could distort the result. Even doing all this, I still end up with frames where the corners are resolving more LPM than the center, which means I have slightly misfocused the center and the curvature of field is making the corners appear at their sharpest.
Because the camera to target distance is rather short, I also set up two additional sets of corner charts and test at two other longer distances. The center chart remains in the center. Naturally, the groups that can be discerned with the loupe change the farther the test distance.
I scanned the resolution chart at 300 dpi. There are some flawed places in the scan, but I believe the result is of sufficient quality for the task at hand. It is set up so that if the image file is printed to an 8.5x11 paper size, the chart part of the page should be the correct size for the tests. The actual chart should be 4"x5". The instructions are available in either TIFF, GIF or JPEG format. The TIFF format produces a cleaner and clearer output, and is a much smaller file than the JPEG format, and as such should download quicker.
To attempt to maintain control over how the relevant files are printed -- particularly the chart -- they are all in TIF or GIF format. They should be downloaded, opened in an appropriate graphics editor and printed. By taking the following hotlink you should get the directory of relevant files available on the web site.
http://www.erols.com/johnchap/lenstest
You should download and print out at 300 dpi or higher RESCHRT3.TIF (print at 4.5"x5.5") or REWSCHRT4.GIF (print at 8"x10") and files TXTP1 - TXTP15 in either JPG or TIF format.