Zeiss rolling.

[rjlittlefield]

Tesselator, that morning glory is a jewel -- lovely texture on the flower! Thanks for posting out the raw images. It's interesting and informative to be able to process the original data myself and also to look at the whole image area. Turns out, for example, that even though it's not perfectly focused, that textured white curtain just above image center is notably sharper in the f/5.6 image than it is in the f/22. (I can post out the comparison that I'm looking at, if you'd like.)

Chris, thanks for clarifying about the GH-1 sensor size. Obviously I didn't get into it as deeply as you have. I'm happy to see that your final number is close to what I used. In terms of evaluating diffraction blur, I'm quite sure I can't spot by eye the 6% difference between 4.13 and 3.9 microns.

--Rik

[ChrisLilley]

Chris, thanks for clarifying about the GH-1 sensor size. Obviously I didn't get into it as deeply as you have. I'm happy to see that your final number is close to what I used. In terms of evaluating diffraction blur, I'm quite sure I can't spot by eye the 6% difference between 4.13 and 3.9 microns.

I agree it makes little practical difference. Its just that 'urge to be correct' thing...

[rjlittlefield]

Its just that 'urge to be correct' thing...

Hey, I know that one!

--Rik

[missgecko]

Well......I've just scrolled through this topic Tesselator and will honestly admit nearly all of it went straight over my head. But I will say I love the second butterfly on the first page but my favourite is the bee with the pollen sticking to it
Cheers

[PaulFurman]

Hope you don't mind but this spot seemed to show the comparison about right by my reckoning, which is a near tie between f/5.6 & f/8 with f/8 winning. These are 500% so sort of obsessive pixel peeping, which means there isn't much difference. F/11 is the next in line, f/4 & 16 are close too. I darkened the exposure -1.60 in ACR.




and at 100% in the most tricky flare/contrast spot:

[Tesselator]

No, I don't mind! That's an excellent comparison! Very well done!

And it's changed my opinion about this len's sweet spot - which prior to your post I thought was at f/4.0. Typically sweet spots hit in around f/8 and I was surprised to think that this lens was 4.0. I'm less surprised now.

This became a nice thread and a good testament for someone looking for an inexpensive close-up solution (assuming they get it with the optional 7.5mm tube.)

I'm kinda waiting for myself to print out the res chart before posting much more but as requested here's a few more from the same day:



1/160s, f/4.0, ISO100



1/200s, f/4.0, ISO100




1/60s, f/4.0, ISO100, Subject size 2cm, Subject distance: 30cm




1/500s, f/4.0, ISO100

[PaulFurman]

I expected f/4 or 5.6 because of diffraction with such small pixels. It could be misleading because the road stripe at 500% is in front of the focus and majority of the image's detail in the building is behind the focus, so that stopping down will improve those. The car is so simple & contrasty it's hard to see really fine detail on it. Then there's the possibility of focus shift when stopping down if you left the lens focused between shots. These tests are maddening :-)

[Tesselator]

Then there's the possibility of focus shift when stopping down if you left the lens focused between shots.

Oh, what's this? Focus shift? I should refocus between stops? Could you explain about this? I've never heard of it.

But I might say beforehand; the smaller the aperture becomes the more impossible it becomes to focus with any accuracy. I suppose a split prism focusing screen would remedy this but my GH1 is EVF so I guess there's little chance of applying that tech. In fact, with the majority of these MF lenses I'm finding that I need to open the aperture all the way, focus, and then stop down for the desired DOF affect.

[rjlittlefield]

Then there's the possibility of focus shift when stopping down if you left the lens focused between shots.

Oh, what's this? Focus shift? I should refocus between stops? Could you explain about this? I've never heard of it.

This occurs because of spherical aberration. Briefly, if a lens has residual spherical aberration, either under- or over-corrected, then light rays focus in progressively slightly different planes depending on whether they pass through the center of the lens or farther out. When the lens is wide open, most of the light comes through its outer portions, so when you focus wide open, you are focusing for the outer rays. But when you stop down, the outer rays are blocked, leaving only the inner rays that focus in a different plane. The net effect is that the focus shifts.

This effect is easiest to see if the scene includes a continuous range of depths, so that at each aperture you can easily see the limits of DOF and thus accurately estimate where the center focus is. For an extreme macro example, see HERE, second panel, second column, and note that the in-focus area for the Mamiya Sekor 55 mm f/1.8 shifts significantly backward between f/2 and f/5.6. My Sigma 105 mm f/2.8 macro lens exhibits a similar shift. In that case, the rule of thumb is that if I'm focusing wide open around 1:1, I should focus nearer the front of the range I actually want focused when stopped to f/11, as opposed to dead center in the f/11 range as it would be if focus did not shift.

Some lenses are specifically designed so that this is not an issue. Nikon made a selling point that their EL Nikkor enlarging lenses did not shift focus when stopped down.

But many lenses do shift, and I don't know any way of predicting when or how much except by measuring it via some aperture series involving a suitable target. Ideally the measurement should be made at focus distances similar to where you're going to use it. This is because the shift depends on spherical aberration which depends on focus distances.

The car-and-building scene you posted earlier does suggest that your lens exhibits some shift. Here is a side-by-side comparison of the f/2.8 and f/5.6 frames, processed by dcraw from your .RW2 files and reduced to 70% so they'll fit fairly well here. Notice the pavement at bottom of frame gets only a little bit sharper on stopping down, while the background gets a lot sharper. Usually this sort of asymmetric improvement indicates a focus shift. In this case it's not sure thing because I'm not sure what that pavement would look like when it's as crisp as possible, but it's a hint at least.

--Rik

[Cactusdave]

I'm impressed by a number of very nice, beautifully lit, 'luminous' images shot with older lenses designed for film SLR cameras mounted on DSLR bodies, that have been presented. I wonder how much of this is lens and how much is good technique? I assume these shots were all shot on a tripod or other support.

[ChrisLilley]

Oh, what's this? Focus shift? I should refocus between stops? Could you explain about this? I've never heard of it.

If the aperture used for focssing is different from the aperture used for taking the photo, and the depth of field is shallow so that small errors in exact focus are visible, then some lenses will have a visible shift so that a different part of the subject is in focus.

Some lenses show more of this than others. And when the depth of field is large, small errors are not noticed.

In fact, with the majority of these MF lenses I'm finding that I need to open the aperture all the way, focus, and then stop down for the desired DOF affect.

Yes, its easier to focus with the aperture wide open as the image is brighter. On an SLR, most lenses are designed to work like this automatically. The aperture is held wide open all the time, for focus and for metering; the aperture is only closed down to the selected one when actually taking the shot.

I made some tests of focus shift with a Zeiss ZF 50/1.4; I will look for them.

[ChrisLilley]

1/60s, f/4.0, ISO100, Subject size 2cm, Subject distance: 30cm

That mushroom shot is lovely.

[Tesselator]

RJ,
Awesome explanation and exemplification. I (think) I understand it completely now. Testament to a great teacher! Seriously, that explanation makes perfect sense! Thank you!



Dave,
No pods or anything, just my usual goofy-fingered style. I know the basics though. What shutter speeds I can get away with, how many brackets I need to take at what speeds in order to insure that at least one of them is sharp, how to hold the camera (although I don't always follow the rules here), how to frame an image (although that's personal preference), what lighting is likely to look good/interesting, when and how to use flash slash off camera flash, etc.

I'm of the opinion that a lot of it is lens tho. Lenses that were designed to produce an image circle to cover the FF of the 35mm format almost all have very nice centers. Even really cheap and crappy lenses. So on my M43 format with a crop factor of 2, I'm using only the very best part of the lens. Thus I have been ecstatically pleased with lenses I would normally dump if I were on a FF camera like the D700 or something. The down sides are of course that AF lenses don't AF, and the aperture control that Chris mentioned becomes a manual operation. I personally think it's totally worth it though. Some lenses I KNOW are losers on a FF system are producing pristine results on my GH1.

That said, I'm just now getting up the courage to take my pod out with me when I go out shooting the GH1. Until now I wanted to get to know the camera hand-held. I think I'm about there. I even recently took out my 1.5 meter long telescope and shot that hand-held. Yeah, everyone was looking at me like I was nuts! Here's a shot from that:


30% scale crop. 1/1000s ISO640, f/10 (I think), 1000mm



Chris,
Thanks for the compliment! And: Ah-Ha! That's what's missing on these adapters! They need to open and close the aperture blades with actuation! Dang! Well, it seems easy enough to design. So either I create a better product or I just wait. I think I'll wait. I wonder if anyone will do it? I wonder what the demand is for adapters? I know I bought 14 right off the bat before I even received delivery of the camera. Yet when I frequent the M43 forums everyone is talking about the kit and optional m43 maker lenses and almost no one talks about Zuiko, Nikkor, Rokkor, Asahi, or etc. so maybe it won't ever happen unless I make it happen myself.


BTW, anyone have an ID on that mushroom?

[PaulFurman]

Looks like a Russula...

[Tesselator]

Yup, I think you're right. http://en.wikipedia.org/wiki/Russula and it's probably eatable too (if it's not bitter when tasted raw). Neat!