Pages: [1] 2   Go Down

Author Topic: Testing a Lens  (Read 4827 times)

jonathan.lipkin

  • Full Member
  • ***
  • Offline Offline
  • Posts: 158
Testing a Lens
« on: June 14, 2011, 09:45:05 pm »

Ok, perhaps a basic question, but how do you test a lens? I've seen many lens reviews that use targets to test the resolving power of lenses, both at SLR gear and dp review.

I've just bought a Hasselblad 50mm HC and want to test for sharpness. So, I've pointed it at a bulletin board that has a woven fabric, and so some texture, and taken photographs at different apertures. I don't have a test target. After taking the photos, how should I evaluate them? Is on-screen at 100% sufficient - I've heard this referred to derisively as pixel-peeping. Should I make prints at my expected output size? Should I sharpen the images?

My workflow with DSLR was to import to DNG, use Photokit sharpener to capture sharpen, adjust then output sharpen before sending to inkjet printer.

Hope I'm not opening a can of worms, and please let me know if this has been covered in another thread that I've missed.

Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Testing a Lens
« Reply #1 on: June 15, 2011, 03:25:54 am »

Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Testing a Lens
« Reply #2 on: June 15, 2011, 06:26:00 am »

Ok, perhaps a basic question, but how do you test a lens? I've seen many lens reviews that use targets to test the resolving power of lenses, both at SLR gear and dp review.

Hi Jonathan,

How, depends on what you want to test. Do you want to compare it relative to a known reference, or do you want absolute numbers? Do you want to know if corner performance is helped by stopping down, or do you want to know what the best area of the image circle is, or if it is centered?

Also important is, do you want to know how it performs on a flat surface, which involves the influence of curvature of field, or is it more important to know how the corner performance is even if it is a bit in front/back of the center plane of focus?

Using a test target will remove lots of variables from the test, so you can arrive at certain conclusions faster. However, such tests do not say much about how the bokeh looks, both in front and in the rear defocused areas (can be quite different in non-symmetric lens designs). I also doesn't tell you how the vignetting behavior is, or the sensitivity to veiling glare.

For absolute numbers, it's hard to beat the Imatest software mentioned above because it also encourages to work methodically. For a visual and relative test you can use targets that are available for download on various sites, e.g. here.

Cheers,
Bart
« Last Edit: June 15, 2011, 06:28:41 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Testing a Lens
« Reply #3 on: June 15, 2011, 06:37:45 am »

I think the first major technical question is to see whether the (auto)focus is on-plane.
This is the one point you should really determine.

Edmund


Ok, perhaps a basic question, but how do you test a lens? I've seen many lens reviews that use targets to test the resolving power of lenses, both at SLR gear and dp review.

I've just bought a Hasselblad 50mm HC and want to test for sharpness. So, I've pointed it at a bulletin board that has a woven fabric, and so some texture, and taken photographs at different apertures. I don't have a test target. After taking the photos, how should I evaluate them? Is on-screen at 100% sufficient - I've heard this referred to derisively as pixel-peeping. Should I make prints at my expected output size? Should I sharpen the images?

My workflow with DSLR was to import to DNG, use Photokit sharpener to capture sharpen, adjust then output sharpen before sending to inkjet printer.

Hope I'm not opening a can of worms, and please let me know if this has been covered in another thread that I've missed.


Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

MarkoRepse

  • Full Member
  • ***
  • Offline Offline
  • Posts: 234
    • www.markorepse.com
Re: Testing a Lens
« Reply #4 on: June 15, 2011, 06:52:07 am »

It is also worth mentioning that lens performance can vary at different focus distances.
Logged

ced

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 287
Re: Testing a Lens
« Reply #5 on: June 15, 2011, 09:54:36 am »

The best I think that is for free is the ISO Chart that is on the site recommended above and you can have my 2 cents worth here:
https://picasaweb.google.com/cedmus/RandomImages#5496369040804408306
Logged

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Testing a Lens
« Reply #6 on: June 15, 2011, 10:21:39 am »

The most sensitive test of lens quality is when you image point sources.

Stars are best, but you need a tracking equatorial mount for best results, as stars are dim and the Earth turns noticeably in exposures over a few seconds.

I often substitute distant urban lights for stars. Tilt your tripod-mounted camera so that the urban skyline runs diagonally from corner to corner of the frame. Then shoot at different apertures. Maybe 1 sec at f2.8, 2 sec at f4, ...etc. Use everything you can to reduce vibration - mirror lockup, remote release, self timer...

When you examine the images you may be shocked by how bad the point-lights are in the corners, expecially wide open; but also encouraged at how much better they get 1 or 2 stops closed down. As I said, this is a very sensitive test.

What I also like about this is that it is charge-free and fuss-free - no printing out and mounting of test charts at special distances and positions, or any of that chicanery. Just pick your spot, arrive in the evening and shoot. Society's addiction to wasteful light pollution has given you a gigantic, free "test chart".

Here's an example from my film days - same lens, first f2.8 (bad coma and oblique spherical aberrations) then f4 (very well cleaned up).

Ray
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Testing a Lens
« Reply #7 on: June 15, 2011, 10:50:27 am »

What I also like about this is that it is charge-free and fuss-free - no printing out and mounting of test charts at special distances and positions, or any of that chicanery. Just pick your spot, arrive in the evening and shoot. Society's addiction to wasteful light pollution has given you a gigantic, free "test chart".

Hi Ray,

While distant lights are useful, one should try and avoid atmospheric turbulence creeping into the test result (another issue that also applies to shooting stars, even directly overhead). Therefore, mornings are better than evenings if one wants to test this way. Also, exposure times of 1/15th to 1 second are most vulnerable to camera shake/mirror slap. Testing is not simple ..., if solid conclusions need to be drawn.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

jonathan.lipkin

  • Full Member
  • ***
  • Offline Offline
  • Posts: 158
Re: Testing a Lens
« Reply #8 on: June 15, 2011, 11:16:38 am »

Thanks for the advice everyone. Should have been more specific in my initial question.

I've heard that the 50 is one of the weaker HC lenses, and am considering purchasing the 50 II. I've made large (40x60) prints from the 35 and am happy with those prints. However, I have not shot with the 50 in real world applications, and don't have access to the 44" printer now. So, I have the 35 as a point of reference and want to compare the 50 to it.

If I photograph a known target at the same aperture (say f/11, which I'm told is a good one for that lens) what is the best way to compare two images. I may have a chance to test the 50II, but only for a few minutes in the store. What would be the best way to do that. If I were to take two pictures and then look at them side by side on screen, would that suffice?? Do I use any input sharpening (Photokit sharpener does not seem to produce any noticable effects, so I suppose I'll use Phocus to sharpen)?
Logged

Dustbak

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2442
    • Pepperanddust
Re: Testing a Lens
« Reply #9 on: June 15, 2011, 03:44:57 pm »

I have owned the HC50vsI and enjoyed using it. I think that is the only thing that is important to me. Both my zooms were sharper than the HC50 however the 50 on a body fits in a much smaller bag and is much nicer to use because of that. The HC50 has some weaknesses. You will learn these pretty quickly and stay away from them if that is important to you.

Eventually I did not use it anymore and sold it planning to buy the vsII eventually. My point is, you have it. Just use it, if you feel you would like to have the vsII buy it eventually and enjoy it just as much. Unless you have a complete dud you should be able to put it to good use.

« Last Edit: June 16, 2011, 10:58:23 am by Dustbak »
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Testing a Lens
« Reply #10 on: June 15, 2011, 06:32:18 pm »

The most sensitive test of lens quality is when you image point sources.

Stars are best, but you need a tracking equatorial mount for best results, as stars are dim and the Earth turns noticeably in exposures over a few seconds.

I often substitute distant urban lights for stars. Tilt your tripod-mounted camera so that the urban skyline runs diagonally from corner to corner of the frame. Then shoot at different apertures. Maybe 1 sec at f2.8, 2 sec at f4, ...etc. Use everything you can to reduce vibration - mirror lockup, remote release, self timer...

When you examine the images you may be shocked by how bad the point-lights are in the corners, expecially wide open; but also encouraged at how much better they get 1 or 2 stops closed down. As I said, this is a very sensitive test.

What I also like about this is that it is charge-free and fuss-free - no printing out and mounting of test charts at special distances and positions, or any of that chicanery. Just pick your spot, arrive in the evening and shoot. Society's addiction to wasteful light pollution has given you a gigantic, free "test chart".

Here's an example from my film days - same lens, first f2.8 (bad coma and oblique spherical aberrations) then f4 (very well cleaned up).

Ray

The point spread function is not the best way to test a lens. See page 3 and 4 of this Zeiss article. The PSP of a digital camera with a low pass filter is rather frightening.

Regards,

Bill
Logged

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Testing a Lens
« Reply #11 on: June 15, 2011, 08:01:47 pm »

The point spread function is not the best way to test a lens. See page 3 and 4 of this Zeiss article. The PSP of a digital camera with a low pass filter is rather frightening.

Regards,

Bill

Aye Bill, but none of the digital backs used on a H-series or other MF body would have a low-pass filter. If Jonathan wants to test his HC 50mm lens, he need not worry about that issue.

Besides, the Zeiss article dismisses the PSF and waxes lyrical about the MTF - but fails to mention that the two are mathematically interchangeable and convey exactly the same information: the PSF is the Fourier Transform of the OTF, and the MTF is the magnitude of the OTF.

I know several optical designers and they don't plot MTFs to illustrate their lens' performance: they primarily plot spot diagrams (PSFs), as a function of either field position or wavelength.

Ray
Logged

David Eichler

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 823
    • San Francisco Architectural and Interior Photographer
Re: Testing a Lens
« Reply #12 on: June 15, 2011, 09:44:12 pm »

A good, traditional quick-and-dirty test is to shoot a brick wall in evenly lit open shade, at various distances and at all apertures. Of course, make sure to use a solid tripod and lock the mirror up, and use a remote release. Then, try shooting into various bright light sources, also at all apertures, with the light source at various positions in the frame. Not an exhaustive test, by any means. But if the lens is mediocre or has problems, this should show them quickly.

Oh yes, when shooting the wall, make sure the camera is level and that the wall and sensor are parallel to one another.
Logged

mediumcool

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 770
Re: Testing a Lens
« Reply #13 on: June 16, 2011, 05:25:24 am »

Take pictures with it, and make prints, or get them made for you, if that is your practice.

If you normally shoot brick walls, then by all means keep shooting brick walls. Otherwise shoot what you want to shoot, and keep records if that floats your boat.

There is a lot more to photography than sheer resolution and distortion or lack thereof.
Logged
FaceBook facebook.com/ian.goss.39   www.mlkshk.com/user/mediumcool

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Testing a Lens
« Reply #14 on: June 16, 2011, 06:50:01 am »

Aye Bill, but none of the digital backs used on a H-series or other MF body would have a low-pass filter. If Jonathan wants to test his HC 50mm lens, he need not worry about that issue.

Besides, the Zeiss article dismisses the PSF and waxes lyrical about the MTF - but fails to mention that the two are mathematically interchangeable and convey exactly the same information: the PSF is the Fourier Transform of the OTF, and the MTF is the magnitude of the OTF.

I know several optical designers and they don't plot MTFs to illustrate their lens' performance: they primarily plot spot diagrams (PSFs), as a function of either field position or wavelength.

Ray

...actually I was being a bit unfair. To give the Zeiss author his due, near the end, on page 29, I now notice that he does acknowledge that the MTF, representing only the magnitude of the OTF, is less informative of image quality than the PSF:
"The MTF does not tell us anything about this difference, because it does not yet completely describe the characteristics of the point spread function...".

The phase of the OTF is what's absent in the MTF, and present in the PSF. This is why the MTF has to be separately computed and plotted for different position angles (tangential, sagittal); and even those 2 representative orthogonal angles fall far short of providing the complete 360-degree radial phase information of the PSF, including any asymmetry.

Which means, in a nutshell, when you look at a well-sampled PSF, you can see at a glance not only how good the optics are, but also what's wrong with the optics (which gross aberrations are present). MTF plots don't give you that useful diagnostic feedback.

Ray

Logged

Christoph C. Feldhaim

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2509
  • There is no rule! No - wait ...
Re: Testing a Lens
« Reply #15 on: June 16, 2011, 08:36:00 am »

Take pictures with it, and make prints, or get them made for you, if that is your practice.

If you normally shoot brick walls, then by all means keep shooting brick walls. Otherwise shoot what you want to shoot, and keep records if that floats your boat.

There is a lot more to photography than sheer resolution and distortion or lack thereof.
+1

Two must-read links (pdf files):
Zeiss Camera Lens News - a two part article about measuring lenses and MTF
CLN 30
CLN 31

Once you have read and understood these two you'll have a theoretical background, why the cited text above is right.

Cheers
~Chris


Disclaimer: I have zero affilitation with Zeiss, don't own Zeiss lenses (unfortunately) or receive money from them.

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Testing a Lens
« Reply #16 on: June 16, 2011, 09:04:59 am »

+1

Two must-read links (pdf files):
Zeiss Camera Lens News - a two part article about measuring lenses and MTF
CLN 30
CLN 31

Once you have read and understood these two you'll have a theoretical background, why the cited text above is right.

Cheers
~Chris

Disclaimer: I have zero affilitation with Zeiss, don't own Zeiss lenses (unfortunately) or receive money from them.


Chris,

You seem to be "+1"ing a viewpoint which is actually opposite to your own. You obviously believe in the merits of optical testing, and Mediumcool doesn't. So I don't follow why you "+1"ed him?  ???

Also the 2 links you gave are only to "stubs" for the Zeiss articles.  The full articles cannot be retrieved, it seems. But I imagine they have much the same content as the one that Bill linked below.

Ray
Logged

Christoph C. Feldhaim

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2509
  • There is no rule! No - wait ...
Re: Testing a Lens
« Reply #17 on: June 16, 2011, 09:13:56 am »


Chris,

You seem to be "+1"ing a viewpoint which is actually opposite to your own. You obviously believe in the merits of optical testing, and Mediumcool doesn't. So I don't follow why you "+1"ed him?  ???

Also the 2 links you gave are only to "stubs" for the Zeiss articles.  The full articles cannot be retrieved, it seems. But I imagine they have much the same content as the one that Bill linked below.

Ray

Did you read the Zeiss articles?
Not only can one learn a lot about MTFs, but also on the limitations of MTFs and why self testing of a lens is important and not everything is in the numbers and curves.

For me the links work and get me PDF files of the "Camera Lens News" issues with the respective articles in them. No idea why it doesn't work for you.

Cheers
~Chris

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Testing a Lens
« Reply #18 on: June 16, 2011, 11:31:08 am »

Did you read the Zeiss articles?
Not only can one learn a lot about MTFs, but also on the limitations of MTFs and why self testing of a lens is important and not everything is in the numbers and curves.

For me the links work and get me PDF files of the "Camera Lens News" issues with the respective articles in them. No idea why it doesn't work for you.

Cheers
~Chris

I'd love to read them! But....the CLN 30 pdf comes to a dead stop on page 9, right after this:
"Dr. Hubert Nasse, Senior Scientist
with Carl Zeiss AG, has exhaustively
investigated this subject. Read on if
you want to become an expert too."

And CLN 31 is similarly teasing - ending like this on page 8:
"You can find the complete article
online at www.zeiss.com/cln in issue
31 of the Camera Lens Newsletter."

That www.zeiss.com/cln link redirects you automatically to http://blogs.zeiss.com/photo/en/
Within the new blog site, there is an Archive section at http://blogs.zeiss.com/photo/en/?page_id=18

Clicking on the Archive version of CLN 30 or CLN 31 takes you right back to ... the same PDFs that end abruptly on pages 9 and 8 respectively, and send you to www.zeiss.com/cln ...with its Archive section...

It's a perfect circle of hell  :D
Welcome to Carl Zeiss Optics - you'll never leave!   :o

Ray
Logged

Christoph C. Feldhaim

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2509
  • There is no rule! No - wait ...
Re: Testing a Lens
« Reply #19 on: June 16, 2011, 11:41:50 am »

I'd love to read them! But....the CLN 30 pdf comes to a dead stop on page 9, right after this:
"Dr. Hubert Nasse, Senior Scientist
with Carl Zeiss AG, has exhaustively
investigated this subject. Read on if
you want to become an expert too."

And CLN 31 is similarly teasing - ending like this on page 8:
"You can find the complete article
online at www.zeiss.com/cln in issue
31 of the Camera Lens Newsletter."

That www.zeiss.com/cln link redirects you automatically to http://blogs.zeiss.com/photo/en/
Within the new blog site, there is an Archive section at http://blogs.zeiss.com/photo/en/?page_id=18

Clicking on the Archive version of CLN 30 or CLN 31 takes you right back to ... the same PDFs that end abruptly on pages 9 and 8 respectively, and send you to www.zeiss.com/cln ...with its Archive section...

It's a perfect circle of hell  :D
Welcome to Carl Zeiss Optics - you'll never leave!  :o

Ray


Yes.
They have changed something on their website which is awful.
This Link leads to the english CLN main site.
Issues 30 and 31 are the ones with the articles:
http://blogs.zeiss.com/photo/en/?page_id=18


EDIT: Crap! now I see - the PDF is crippled ... time to mail Zeiss for that ....
However - the german version is online ..
« Last Edit: June 16, 2011, 11:53:20 am by Christoph C. Feldhaim »
Logged
Pages: [1] 2   Go Up