Size of aperture as it affects sharpness

[Nodda Duma]

Correct. to be convinced, use 2 telescope with different mirror diameters. High chance the bigger might give you the best image. If it were not true, professional observatories would all be equipped with 4in. refractors...

Not true!

Telescope mirror diameters get larger to increase contrast, not image quality. Larger mirrors are harder to fabricate and statistically have worse figure than smaller mirrors for a given f/#. Aberrations scale with f/# (faster optic, bigger aberrations) for *all* optics... the application does not make this fact magically go away. Larger optics for an equivalent focal length will be more complex and/or difficult to make.


The "best image" in the context of this thread for astronomical telescopes is a function of atmospherics and fabrication quality. Hubble's primary is f/24 and was precision figured for image quality aka "best image" reasons. Hubble's primary is 2.4m diameter to increase contrast and see faint stuff. Another way to look at it is that the Hubble primary's huge diameter didn't correct the spherical aberration resulting from incorrect figuring.

 

[Nodda Duma]

All of these "rules of thumb" don't really give an adequate answer. There are millions of lenses out there, and they're all different.

If you want to know, do a test. Get a test chart or pin up some printed pages and do a close test, then go outside and shoot a detailed building or something. Use a tripod, get every aperture, take notes.

It's an unanswerable question (my .02 anyway) unless you test the lens itself, and then you have the answer.

I agree with this response (except I would measure MTF on an MTF bench).

-Jason

 

[RalphLambrecht]

When shooting landscapes with MF, I put the DOF i want to achieve first and ignore diffraction and lens sharpness. In fact, I'll stop down one additional stop to make sure the DOF is covered. I'm using Mamiya RB67. Am I wrong?

this technique will always leave infinity focus at the threshold of sharpness and will never be in true focus.

 

[Dali]

Not true!

Telescope mirror diameters get larger to increase contrast, not image quality. Larger mirrors are harder to fabricate and statistically have worse figure than smaller mirrors for a given f/#. Aberrations scale with f/# (faster optic, bigger aberrations) for *all* optics... the application does not make this fact magically go away. Larger optics for an equivalent focal length will be more complex and/or difficult to make.


The "best image" in the context of this thread for astronomical telescopes is a function of atmospherics and fabrication quality. Hubble's primary is f/24 and was precision figured for image quality aka "best image" reasons. Hubble's primary is 2.4m diameter to increase contrast and see faint stuff. Another way to look at it is that the Hubble primary's huge diameter didn't correct the spherical aberration resulting from incorrect figuring.

Strange to read that contrast is not contributing to image quality... Also, I always thought that angular resolution was function of the optical diameter. Not true anymore? Last, who said that large mirrors have to be badly shaped and big scope badly aligned? Since John Dobson's big guns in the '70s, amateurs got more critical on optical and mechanical qualities of their scopes. Several test procedures were developed for surface and shape quality control and alignment tools got more sophisticated, hence the possibility now to get short f/D ratio for convenience. Again, if it were not true, all amateurs would use 4" scopes.

 

[Nodda Duma]

No offense intended at all, but you have the fundamental relationships confused. Dobson would agree with me. Recommend reading Hecht Optics for starters. There's some better astronomical optic specific references but I can't recall off-hand. Don't want to derail the thread.

 

[Alan Edward Klein]

this technique will always leave infinity focus at the threshold of sharpness and will never be in true focus.

Ralph, I understand that but isn't DOF a compromise to expand the largest DOF that would be in enough focus? For example, I focused on the back fence support and used a aperture that covered DOF from the front to more or less infinity.

 

[Dali]

No offense intended at all, but you have the fundamental relationships confused. Dobson would agree with me. Recommend reading Hecht Optics for starters. There's some better astronomical optic specific references but I can't recall off-hand. Don't want to derail the thread.

I invite anyone to attend a star party and to see by himself/herself which instrument gives the best image. This way, the debate should find an end.

 

[Trask]

This has been an interesting discussion -- including the telescope aspects. I've thought of another way to think about my question. If lens designer were designing a 65mm medium format lens to have maximum aperture f/5.6, perhaps the thinking would be that the lens would be sharper a stop or two down. Now design the lens to be f/8 -- would the results wide open with that lens be the same as the f/5.6 lens stopped down to f/8, or would the f/8 lens be less sharp because that's the maximum aperture. Based on what I've gleaned from everyone's comments, the answer is --- it depends! Depends on the lens design, what aspects the designer optimizes for, etc.

 

[RalphLambrecht]

Ralph, I understand that but isn't DOF a compromise to expand the largest DOF that would be in enough focus? For example, I focused on the back fence support and used a aperture that covered DOF from the front to more or less infinity.

Yes, but in landscapes infinity focus is critical for an in-focus background.that can only be achieved with focusing on the background an not with a DOF compromise, which leaves the back and front of the DOF range less than critically sharp.

 

[M Carter]

This has been an interesting discussion -- including the telescope aspects. I've thought of another way to think about my question. If lens designer were designing a 65mm medium format lens to have maximum aperture f/5.6, perhaps the thinking would be that the lens would be sharper a stop or two down. Now design the lens to be f/8 -- would the results wide open with that lens be the same as the f/5.6 lens stopped down to f/8, or would the f/8 lens be less sharp because that's the maximum aperture. Based on what I've gleaned from everyone's comments, the answer is --- it depends! Depends on the lens design, what aspects the designer optimizes for, etc.

Exactly. And the thing is, it's easy to find out. If it's a 35mm lens, shoot half a roll and develop it if money is tight.

I don't know about everyone else, but doing decently-planned testing makes me feel a little badass, too. Like I've left the realm of horseshoes and hand grenades and throwing spaghetti at the wall. Doing B&W film and development tests with a spot meter has had a big impact on my keeper negs and ease of printing - I test to my final output, not like "can you read a newspaper through the highlights" BS.

I really like Rollei IR400 with a deep red filter, not full-IR (seen one white tree, seen 'em all..) And I don't know about you guys, but in my tests as a straight film, it's barely ISO 100. A quick afternoon test and I was like "sheesh, no wonder!" (BTW, a neat trick with Rollei 400? Wash, dry, and re-roll it before you shoot it. You can get that Efke glow to a good extent. That was another test that went well).

 

[Alan Edward Klein]

Yes, but in landscapes infinity focus is critical for an in-focus background.that can only be achieved with focusing on the background an not with a DOF compromise, which leaves the back and front of the DOF range less than critically sharp.

Ralph, I forgot to post the link with my last post. https://www.flickr.com/photos/alanklein2000/30334168051/in/album-72157626597775701/

The point mis the back fence post was the focal point and the DOF was to cover to the front of the fence. While I wanted to get the infinity as focused as possible, it was not the main subject, the fence was. So I calculated the DOF I wanted and then stopped down one additional stop. That's my usual procedure. Does this make sense now?

 

[Nodda Duma]

This has been an interesting discussion -- including the telescope aspects. I've thought of another way to think about my question. If lens designer were designing a 65mm medium format lens to have maximum aperture f/5.6, perhaps the thinking would be that the lens would be sharper a stop or two down. Now design the lens to be f/8 -- would the results wide open with that lens be the same as the f/5.6 lens stopped down to f/8, or would the f/8 lens be less sharp because that's the maximum aperture. Based on what I've gleaned from everyone's comments, the answer is --- it depends! Depends on the lens design, what aspects the designer optimizes for, etc.

Yes, it really does depend on the requirements. I designed a very high performance lens -- though not fast at f/2.5 -- a couple years ago with best (predicted and measured) performance wide open. Plotting MTF vs. f/# showed a monotonically descending graph, not the typical peak 2 stops down. This was an interesting result, but ultimately was driven by the requirements for the optic.

 

[Marco Gilardetti]

Ralph, I forgot to post the link with my last post. https://www.flickr.com/photos/alanklein2000/30334168051/in/album-72157626597775701/

The point mis the back fence post was the focal point and the DOF was to cover to the front of the fence. While I wanted to get the infinity as focused as possible, it was not the main subject, the fence was. So I calculated the DOF I wanted and then stopped down one additional stop. That's my usual procedure. Does this make sense now?

Dear Alan, the topic you're touching is controversial (in a positive way). Most books, papers and people (myself included) affirm that infinity should be focused, and then eventually the DOF calculated in order to have the fence reasonably in focus. However, eminent photographers (like Ansel Adams) used to say that the main subject of the picture should be focused, and the DOF adjusted accordingly. So there isn't a general consensus on this. For how odd and snooty this may sound, experience has taught me that I "should not agree with Adams at least on this matter"...

Coming to your picture, although it's hard to tell by a scan, I would say that the trees at infinity ar not really in focus, which is exactly what I would expect from DOF calculations. DOF is an approximation, it just tells you what would be "reasonably sharp" at a "prefixed enlargement". If you change the enlargement ratio, or your concept of "reasonably sharp" is not on par with that of whoever calculated the DOF table, you're fried. If you want to put it another way, my personal concept of "sharp" almost never conicides with that of "reasonably sharp".

As a more general note, in your picture there is too much distance from foreground to background, there is no DOF that would *really* bring all that in focus (at least in *my* concept of "in focus"). If I were you, I would have kept the fence as the main subject, and blurred the background by intentionally reducing the DOF. In my humble opinion, that background doesn't add anything special to your picture, but it diverts attention from the subject. But of course this is just personal taste and you may disagree.

 

[Sirius Glass]

And engineer at Kodak had a telescope with a 40" mirror and was grinding a 60" mirror when I worked there.

Telescope requirements are different: it's not sharpness, but light-gathering ability in order to see faint stars, nebulae, &c. Regardless of mirror or main objective size, sharpness is probably hampered mostly by our atmosphere: equivalent to looking through 30 feet of water. Hence the Hubble platform.

His mirrors were ground to 1/4 wavelength across the spectrum. They were ground by that Kodak used for its spacecraft telescopes.

 

[Nodda Duma]

His mirrors were ground to 1/4 wavelength across the spectrum. They were ground by that Kodak used for its spacecraft telescopes.

1/4 wave would have been measured interferometrically at a single wavelength...most likely 632.8nm. When surface accuracy in wavelength is provided, the reference wavelemgth is also specified (also usually specified either P-V or RMS). At a different wavelength, the surface figure would have a different value...1/8 at 1265.6nm for example...although I'm pretty sure I understood what you meant.

As for the value itself...this goes back to an earlier point I tried to make but gave up on: Larger mirrors are more difficult to figure and, in a real system, more difficult to maintain. I've made 8" 1/50 wave mirrors (RMS, measured at 632.8), but I've never seen 48" 1/50 wave mirrors. We were pretty flipping happy with 1/15 at that size, but an 8" f/15 would be unacceptable.