f64

[Armin Seeholzer]

Would be interesting if every lens holds up same quality stoped all way down or if it would also give some strong differences!
I used my APO Ronar 360mm not to long ago at f 64 because it was needed and wanted and I was really surprised about the sharpness I got out of it and I had the impression it was a tick sharper then my 360mm Sironar N would have been at the same f stop!
So somewhere in the future I have to do a little test!

Cheers Armin

[jb7]

I'd agree about the Apo Ronar 360, though I've only used it at f/45...
sufficient for my needs at the time-

Regarding the picture in the op-

I presume the detail is of a section roughly one inch wide?
It has a pleasing tonality, but the issue of sharpness must surely be be read as a detail, a small part of a larger whole- as has been stated-
that detail, as a 1/100th part of a larger picture, must show impressive detail over a significant depth of field-

It also looks very clean, undoubtedly a very important quality, and it doesn't look particularly grainy on my monitor.

It doesn't look particularly sharp in itself, but it also doesn't contain the kind of high frequency detail that would be the true test of resolution- or at least, I can't see it...

Thanks for sharing this Lenny. I think that the people who are saying this crop is soft have no idea what size print it actually represents or how good of a scan it really is.

(I'm sure the same people will tell you how good their V750 could do!)

This claim reminded me of an earlier discussion, and some entrenched views regarding scanning, pixels, resolution, and diffraction-
maybe worth another look-

http://www.largeformatphotography.in...t=44797&page=2

Given a correctly processed 8x10 negative, I'm quietly confident about pulling at least 72MP of usable data from my V750; the larger negative works to the strengths of the machine.

[Brian K]

All one need do is look at resolution charts to see, without any subjectivity that most modern LF lenses have better resolution at f 11, 16, and 22 than they do at f64. And decidedly so. It is not open for interpretation, it is fact.

That said, if one were to make a 2x print from an 8x10 negative shot at f64 versus a 2x print shot at f22, it would be hard to discern a difference. At greater magnifications however the differences would increase steadily.

With a 4x5 negative, and I think most LF shooters shoot 4x5 versus 8x10, the differences become evident sooner as the 4x5 requires twice the magnification that a 8x10 does.

A 5x enlargement of a 4x5 negative required to make a 20x24" print would make the differences between f22 and f64 apparent. Let's assume that you are using a high resolving film like T-Max 100, which resolves about 60-65 lp/mm. That resolution is about the typical limit for an excellent lens at it's sharpest aperture (there are lenses that resolve as much as 100 lp/mm but not in LF) so the resolution of the film is in itself not a limiting factor in most LF situations.

A lens that has a resolution of 65 lp/mm will commonly drop to a resolution between 20 and 30 lp/mm due to diffraction, so let's use 25 lp/mm as the resolution at f64. And if one assumes that the average resolution of the human eye is 6 lp/mm (estimates range from 5 to 14 depending on the distance and the color spectrum), then at 5x the resolution of the f64 lens will visibly fall off. For someone with excellent vision if one uses the 10 lp/mm factor, then focus will fall of at 2.5 times magnification.

Of course all of this depends on viewing distance. I've had my work enlarged billboard size and from the much greater distance that is required to view the whole image at that size, the image looks just as sharp as an 8x10 print in my hands.

But understand as a fact, that at f64 a LF lens will NOT resolve nearly as well as it does at f22. But understand further, that this loss of image quality might not be apparent depending on viewing distance and size of magnification. And one should have this knowledge and use it to determine if f64 serves the situation and outcome that you desire.

[Drew Wiley]

A 16x20 contact print taken at f/256 will probably look sharper than a 16x20 print
taken with a 4x5 at f/45. And as far as lens performance goes, all you have to do is
look up the mfg charts with most modern lenses from any of the four major mfg.
There's no substitute for field testing in terms of creative uses of lenses, but the mfg
has already done the hard work on an optical bench, with the variables accounted for. As far as presenting visual resolution evidence over the web, that's like attempting nuerosurgery with a crowbar and chainsaw.

[Armin Seeholzer]

All one need do is look at resolution charts to see, without any subjectivity that most modern LF lenses have better resolution at f 11, 16, and 22 than they do at f64. And decidedly so. It is not open for interpretation, it is fact.

This is all clear to me, but maybe there are lensdesigns which hold better then others way closed down, did somebody do such a test?

Cheers Armin

[sanking]

Brian, I think you made a mistake. According to this source http://www.flickr.com/groups/techtal...7606727844641/ .... Tmax is 150 lp/mm, if we can trust that.

The actual resolution of Kodak Tmax-100 can be found in the Kodak document F-4016. It is 63 lines/mm at TOC 1.6:1 and 200 lines/mm at TOC 1000:1.

See http://www.kodak.com/global/en/profe...16/f4016.jhtml

Sandy King

[Policar]

I think depth of field as an aesthetic concern always trumps resolution. If all we only cared about was resolution, we'd be stitching digital images. Whether to stop down beyond what depth of field demands to get a long exposure versus using an nd filter is a valid question, and at that point you have to determine what you consider acceptably sharp, but if you're sacrificing the quality of the image for "image quality," that's the wrong priority.

The crop looks good to me, but it's been scaled down to 33%, I think. So that's 900dpi about? For 900dpi it seems a little soft but not terribly.

72 megapixels from 8x10 out of the Epson is kind of a disturbingly low figure... That's 18 megapixels out of 4x5. My $750 slr does that no problem.

[Jiri Vasina]

I think depth of field as an aesthetic concern always trumps resolution. If all we only cared about was resolution, we'd be stitching digital images. Whether to stop down beyond what depth of field demands to get a long exposure versus using an nd filter is a valid question, and at that point you have to determine what you consider acceptably sharp, but if you're sacrificing the quality of the image for "image quality," that's the wrong priority.

IMO very well said. There are much more variables at play that result in the success or failure of an image. Resolution (sharpness??) is only one of them...

Jiri

[Lenny Eiger]

Test, then know for sure what the acceptable range is. In the case of the Rodenstock 300mm apo Sironar S, f64 looks great to me.

This, of course, is the point. Testing one's lens and making a print is the easiest test you can do. All you have to do is take two pieces of film and expose them differently, then make prints, however you do it. My guess is that most of you will be quite surprised at the higher fstops given the standard line.... All lenses won't make it to 64, but I am sure a bunch of them will. The others might make it to 45....

The calculated numbers in a previous post about 72mp being the max for an 8x10 are patently ridiculous. They don't include the size of the chip, it's only parallel lines and numbers of pixels. I would say an 8x10 is more like 300, rather than the 568 it calculates to, but the statement is total conjecture, of course. It's only a measure of resolution, not particularly important to most LF'ers. After all, we use big cameras not because they are sharper, but because of their ability to render more textural information. Yet there is no measure for that, certainly not one that gets discussed regularly here. The measure for "quality" is strangely missing.

The only thing you can do, as Van has stated, is test.

Lenny

[jb7]

The crop looks good to me, but it's been scaled down to 33%, I think. So that's 900dpi about? For 900dpi it seems a little soft but not terribly.

72 megapixels from 8x10 out of the Epson is kind of a disturbingly low figure... That's 18 megapixels out of 4x5. My $750 slr does that no problem.

Well yes, quite.

At 900 dpi, it looks soft-
so why the extra effort in scanning to resolutions beyond the combined resolution of lens and film?

Or rather, why draw attention to it?

I think the point about combined resolution is that it takes account of your lens, aperture, and sensor, so reading a number off the box your DSLR came in is more about marketing than actual resolution.
There is a lot of good information in the thread I linked to, hard earned by those who have tested and calculated, and it might need a second reading to get to the finer details.
If you can get 18 DSLR megapixels from an image exhibiting similar depth of field for $750,
then I've wasted my money by buying a D700. What should I have bought?

I think depth of field as an aesthetic concern always trumps resolution. If all we only cared about was resolution, we'd be stitching digital images. Whether to stop down beyond what depth of field demands to get a long exposure versus using an nd filter is a valid question, and at that point you have to determine what you consider acceptably sharp, but if you're sacrificing the quality of the image for "image quality," that's the wrong priority.

Yes, of course. Perhaps the internet is not the best place to test these claims, a good print in a gallery, (where the picture is more important) will always have more value.

Perhaps we shouldn't discuss it at all, and just accept the numbers produced by those whose business it is to produce millions of pixels? Be they Epson or Nikon...

As I mentioned in my post, it's a good looking scan, very clean, and even at 72(combined) megapixels, highly detailed over it's whole area, not just the detail.
I'm willing to accept that the depth of field is covered by the aperture, at the stated print size, so I would be more than happy to accept 72 combined MP...

Perhaps we're too concerned with the large numbers produced by scanners and computers, to the point that we consider 72MP to be unacceptably small-

The calculated numbers in a previous post about 72mp being the max for an 8x10 are patently ridiculous.

Why? Have you tested, because I can imagine that if you'd like to support your claims, that a test like this would be quite easy to set up.

I have no interest myself, because, a: others more qualified than me will do it for me;
and b: I'm only interested in the picture, and I make no claims for my pictures, beyond what they look like.

I have far more faith in 72MP being a very large number than a very small number, because the denominator of the business we're involved in requires the use of ridiculously and increasingly large numbers.

72 is good, still larger than the combined resolution of any currently available digital back.

I believe my brief critique of the scan in my earlier post will get me off the hook regarding the important qualities (imo) of the image-


j