Roger Clark's Sharpness Tutorial Questions....

[Arthur Morris]

As one who has never performed a single sharpness or other photographic test in the 28 years that I have been doing this (I'm a seat of the pants type of guy :)), I was intrigued when I visited Roger's "Relative Lens Sharpness" tutorial here (http://www.clarkvision.com/articles/relative-lens-sharpness/).

But as I read through it I had many questions. Rather than just bug Roger privately I asked him if it would be OK for me to bug him publicly and he kindly agreed. (Much better to share what I learn with the great folks here at BPN.)

I will keep things simple. I will not be asking Roger about Dawes limits or about pixel pitch. But I do hope to learn how to do accurate tests to compare the sharpness of my 400 DO with the sharpness of the 800 f/5.6 and with the sharpness of the 70-200 f/2.8L IS II with the 2X II TC in place.

Roger begins:

How good are lenses? Are prime lenses really better than zooms? Do teleconverters degrade sharpness? On this page I'll show experiments that illustrate relative lens sharpness that can answer these questions. Because there are many lenses, only a few are shown to give examples.

Below is the test chart, Figure 1. It was printed at 300 ppi and imaged with several lenses but normalizing the distance to the target so that it appears the same size in the frame. This allows "relative image sharpness" to be evaluated. For my tests, the distance to the target in feet = focal length in mm /10. For example, for a 500 mm focal length, the distance is set to 50 feet; for 100 mm focal length, it is set to 10 feet.

Figure 1. Roger's resolution test chart. You have my permission to download the above test chart and test your personal lenses (non-commercial applications only). Set the ppi to 300 and print on high quality photo paper, then the lines at 1 pixel (labeled 1 px on the figure) are separated by 1/150 inch.

Note: you need to visit Roger's site to download the file.

[Arthur Morris]

First question: We have two types of photographic paper here, a glossy paper and a nice matte paper. Which should I use? Is there an ideal paper for creating a test chart? (I am not looking to view the results with a microscope; I will just be eyeballing things.)

[Roger Clark]

Artie,

It really doesn't matter what paper you use. It matters only that for the lenses you test you use the same method with each lens so you don't introduce new variables. So simply make some prints, tape them to a board and image with your different lenses, all using the same camera. Then the only variable is the lens.

Roger

[Arthur Morris]

Thanks Roger. Do yo normally focus calibrate each lens with the camera before doing the test?

[Arthur Morris]

I right-clicked on the file and saved it as is. It shows as only 212 kb... That seems tiny for a resolution chart.... Did I screw up?

Also, in the same vein, how large should the chart be printed?

BTW, all are invited to chip in with their own questions as we proceed. This is not a private thread :)

[David Stephens]

I right-clicked on the file and saved it as is. It shows as only 212 kb... That seems tiny for a resolution chart.... Did I screw up?

Also, in the same vein, how large should the chart be printed?

BTW, all are invited to chip in with their own questions as we proceed. This is not a private thread :)

Great thread Art. I'll watch for now and, perhaps, start following along by doing it with my other lenses.

[Arthur Morris]

Thank you kind sir.

[Roger Clark]

Thanks Roger. Do yo normally focus calibrate each lens with the camera before doing the test?

Artie,

(Another long day of meetings-tomorrow too.)

The tests on that page were done a few years ago before cameras had microadjustments. So those tests were done as the camera auto-focused with no additional adjustments. Today with a 1DIV, and 5D2, I do check focus. A quick and simple way to check focus and whether or not one needs microadjustment settings is to first focus on a target (camera on a tripod), then start live view and zoom in with maximum zoom on the live view, then half press the shutter button (using a cable release so you don't shake the camera) and note if the focus changes (getting better). If you see no change, you don't need any AF microadjustment. If you do, then you could use some microadjustment.

On my 1DIV and 5D2, I've never seen a need for any microadjustment. But on my 300 f/4 L IS (which also focuses perfectly on the 1DIV), it needs microadjustment on a friend's 7D. And my 500 f/4 and 300 f2.8 focus fine on that same 7D. Strange.

So to I would recommend checking AF for needing microadjustment. Although one could do the tests using contrast focusing with live view.

Another test of a lens would be to test AF. For example, use the lens as you would in the field, e.g. hand hold for BIF. Move the lens off of focus, then quickly point the lens at the test target and take a picture. Check for sharpness. This test (once one understood the how sharp the lens is on a tripod) will show weaknesses in ones method for getting a sharp image in the situation tested (e.g. hand held BIF).

Roger

[Roger Clark]

I right-clicked on the file and saved it as is. It shows as only 212 kb... That seems tiny for a resolution chart.... Did I screw up?

Also, in the same vein, how large should the chart be printed?

BTW, all are invited to chip in with their own questions as we proceed. This is not a private thread :)

Artie,

Yes, the test pattern image on BPN does not look right. The image on the web site is a gif image. Gif is lossless compression and works well for this kind of subject, so is quite small. The version you posted on BPN is a jpeg, which is lossy. I suggest everyone juts pull the image off the web page your referenced.

Print the target at 300 ppi (it comes out really small, about 1.5 inches wide). I print out several and tape them to a flat board so I have the target in the center and edges of the field in the camera, and at an intermediate position. You can copy a bunch of the charts on an 8x10 canvas and make one print with many targets that you can cut out and tape to a board.

The next trick is to set up the board so it is perpendicular to the lens at the chosen distance. I find a simple way to do this is to use anything with a right angle (e.g. the edges of a pad of paper works as a guide. Set up the board so it looks perpendiculat to the camera. Hold the paper pad so one edge rests on the board and another edge points in the direction of the camera. Does the edge of the pad point toward the camera? Our eyes+brain are very good at estimating this. If not adjust the board until the right-angle edge points to the camera. A double check of this would be to tape a small mirror to the board near the center. Then see where the reflection is located. You should see the camera in the reflection. Or use a laser point held next to the lens. The laser should reflect back to the camera. (Be careful not to reflect the laser into your eye!). I fashioned 1/4-20 screw threads to my board so I could attach the board to a tripod for easy adjustment.

Roger

[Tom Graham]

Thanks Roger. I like how you have sized it so easy to know distance to target in feet. It does print small, 530/300 is 1.767 inches, and by golly that is its real life size on paper :) . And obviously can print several targets on same piece of photo paper. Any advantage to having it in colors? I'm going to set up targets, center and corners, and compare zoom lens focal length settings like you did with the 100-400mm. Also check my hand holding and VR/IS capabilities.
Thanks again - Tom

[Roger Clark]

Thanks Roger. I like how you have sized it so easy to know distance to target in feet. It does print small, 530/300 is 1.767 inches, and by golly that is its real life size on paper :) . And obviously can print several targets on same piece of photo paper. Any advantage to having it in colors? I'm going to set up targets, center and corners, and compare zoom lens focal length settings like you did with the 100-400mm. Also check my hand holding and VR/IS capabilities.
Thanks again - Tom

Tom,
My original idea of using color bars was to see the different resolutions of in the different colors with the Bayer sensor. What I found was that there is no difference (shows how effective Bayer sensors really are). Also note the mixed colors target. That pattern degrades on that pattern faster than the color bars on the white background.

Roger

[Tom Graham]

Thanks Roger,
The FOV as a rectangle (from a chart) for a 1.5 crop factor sensor, 100mm lens, at 10 feet is 27x18 inches.
So, I plan to make a target 27x18 inches with several, maybe 9, of your targets arranged in a "Y" within the 27x18. And a border around it all. Then with the camera viewfinder framing the 27x18, I can easily compute lens focal length by using distance to target . (Distance to target in feet times 10 gives focal length). Why so "convoluted" technique, you may ask. Because a zoom lens I have does not have focal length markings on it. Ahh, but it is in the EXIF data. Well, whatever turns you on ! :)

Tom

[Arthur Morris]

Thanks Tom. After reading your comment I re-read Roger's comment in Pane #9 and now I get it....

Roger. I knew that I had downsized your GIF and converted it to a JPEG (for illustrative purposes).

[Arthur Morris]

Roger, To simplify the distance to chart thing, if I print a chart with multiple targets (all at 300 dpi) at 12 X 18 inches, would it be fine to simply place the rig (lens alone or lens with a TC) so that the target pretty much fills the frame?

I am assuming that if I do so that I will be able to at least compare the results from various lenses fairly accurately.

I am gonna do this testing :) I especially want to compare the 70-200 2.8 II to the 400 DO to the 800 alone.

[Arthur Morris]

Help!!!

OK, I saved the resolution chart as a GIF by right clicking on the image on your web page. Then I re-sized it to tiny at 300 dpi. Then I hit CTRL A CTRL J. It selected the image but would not put the selection on its own layer..... So I re-saved it as a TIFF. Same thing. I tried it on another computer. Same thing. I did the same thing with other images and everything worked just fine.

As the late Vince Lombardi would shout, "What the heck is going on here?"

[Roger Clark]

Help!!!

OK, I saved the resolution chart as a GIF by right clicking on the image on your web page. Then I re-sized it to tiny at 300 dpi. Then I hit CTRL A CTRL J. It selected the image but would not put the selection on its own layer..... So I re-saved it as a TIFF. Same thing. I tried it on another computer. Same thing. I did the same thing with other images and everything worked just fine.

As the late Vince Lombardi would shout, "What the heck is going on here?"

Artie, You need to convert the image mode to rgb color from indexed color, then you will be able to make layers. The gif format and indexed color do not support layers.

In photoshop click on "imge" then "mode" then "RGB color"

Roger

[Roger Clark]

Roger, To simplify the distance to chart thing, if I print a chart with multiple targets (all at 300 dpi) at 12 X 18 inches, would it be fine to simply place the rig (lens alone or lens with a TC) so that the target pretty much fills the frame?

I am assuming that if I do so that I will be able to at least compare the results from various lenses fairly accurately.

I am gonna do this testing :) I especially want to compare the 70-200 2.8 II to the 400 DO to the 800 alone.

Yes, you can do that. What you will be testing will be the lenses at the distance of the target. So instead I would suggest finding the distance that you like to image birds at and print for that distance. You are very good at getting up close and lens performance might be different up close than further away.

(Side note--I was offline yesterday afternoon traveling home, and got (mild) food poisoning at an airport restaurant so, I'm a little slow today. That's now 4 times getting food poisoning while traveling in the last 35 years--the last 3 were from airline food.)

Roger

[Arthur Morris]

Artie, You need to convert the image mode to rgb color from indexed color, then you will be able to make layers. The gif format and indexed color do not support layers.

In photoshop click on "imge" then "mode" then "RGB color"

Roger

Thanks Roger. That works. I had never heard of indexed colors before :)

Next question: I have designed my chart with lots of the little buggers. I will be printing one at 8 X 12" on matte paper. Should I sharpen the file for printing as I would with any other image?

[Arthur Morris]

Yes, you can do that. What you will be testing will be the lenses at the distance of the target. So instead I would suggest finding the distance that you like to image birds at and print for that distance. You are very good at getting up close and lens performance might be different up close than further away.

(Side note--I was offline yesterday afternoon traveling home, and got (mild) food poisoning at an airport restaurant so, I'm a little slow today. That's now 4 times getting food poisoning while traveling in the last 35 years--the last 3 were from airline food.)

Roger

Thanks for all the help. I will print one at 8 X 12 and then decide on a larger size and compare.

Here's a great alternative medicine trick for food poisoning: travel with a bottle of activated charcoal tablets. Take 12 at the first sign of food poisoning. It absorbs the bad stuff. I have seen it cure folks who were near death in 45 minutes. And I once had two lawyers on a San Diego IPT who refused to try the capsules when I offered them; they laid in their room for 2 1/2 days sicker than dogs....

Even your MD will be OK with the activated charcoal; many recommend the same.

[Arthur Morris]

PS: Hope that you are feeling better.

[Roger Clark]

Thanks Roger. That works. I had never heard of indexed colors before :)

Next question: I have designed my chart with lots of the little buggers. I will be printing one at 8 X 12" on matte paper. Should I sharpen the file for printing as I would with any other image?

Artie,
That is a tough question. It depends on the printer. I suggest printing without sharpening and using a magnifying glass to see if the smallest bars are well separated on the print. If so, then you are fine. If not, then try sharpening, but it might not help (meaning it is a limitation of the printer).

If the printer can't print the finest lines real sharp, then it might be better to move the chart far enough away so the lenses are not resolving details in the last bars. It will be trial and error here.

Roger

[Arthur Morris]

Thanks. I was concerned because of the matte paper that we will be using.... I will not likely get to test till I get back from San Diego on the 26th:)

[Roger Clark]

Thanks for all the help. I will print one at 8 X 12 and then decide on a larger size and compare.

Here's a great alternative medicine trick for food poisoning: travel with a bottle of activated charcoal tablets. Take 12 at the first sign of food poisoning. It absorbs the bad stuff. I have seen it cure folks who were near death in 45 minutes. And I once had two lawyers on a San Diego IPT who refused to try the capsules when I offered them; they laid in their room for 2 1/2 days sicker than dogs....

Even your MD will be OK with the activated charcoal; many recommend the same.

Artie,
Thanks for the info. I've known about charcoal, but never thought about getting tablets to carry with me in such a situation. I am feeling better, almost back to normal.

Roger

[Chris Poole]

Roger another remedy is Apple Cider Vinegar. My first wife was a trauma doctor at Chatlos Trauma Center at Orlando Regional. When we traveled to Patagonia I got food poisoning during a trout fishing trip and between the Activated Charcoal tablets and Apple Cider Vinegar and some bed rest I was cured.

[Roger Clark]

Roger another remedy is Apple Cider Vinegar. My first wife was a trauma doctor at Chatlos Trauma Center at Orlando Regional. When we traveled to Patagonia I got food poisoning during a trout fishing trip and between the Activated Charcoal tablets and Apple Cider Vinegar and some bed rest I was cured.

The breadth of things we learn on BPN is amazing..

Thanks,
Roger

[Tom Graham]

Been playing with Roger's test chart. Basically to see if the two zoom lenses I use a lot appear to be performing acceptably (for a zoom lens). Also to test my hand holding ability at various shutter speeds. I'd be curious for any comments about this test, it appears to be acceptable to me. But there are so many variables involved that I'm not sure if any comparisons can be made except for the lenses on my camera under my shooting conditions.
Here's the specs on the shot-
Nikon D200 body, Nikon 70-300 zoom lens, VR/IS on, hand held, crop factor of D200 is 1.5
target 30 feet away, target actual size 1.77x1.37 inches, target is center of lens, (the FOV at 30 feet is 27x18 inches)
lens at 300mm (max), wide open at f5.6 (max at 300mm), shutter at 1/1500, ISO 180, target in sunlight
D200 set for fine jpg out, 0EV, contrast-saturation-sharpness set normal
no additional PP except to resize image for here
http://www.mediafire.com/imgbnc.php/ae923e25d2d67442c2530e2bf12e7953001eb6a679275ca571 2c8ed8761852546g.jpg

Perhaps in addition-
at f8 the lens shows better contrast
my -printer- resolves the target 2px lines but don't think it does the 1px.

Regards - Tom