Ray, I think you are missing the main reason for doing the shift this way. It does not involve tilts at all, and it is important to note that the lens stays in one place while the camera shifts. I am using a lens with a big image circle, and I am shifting the camera across the stationary aerial image.
This has huge advantages over shifting the lens in the normal way. Most importantly, there is no perspective change. This could still go bad when stacking, because of possible variations in the way that stacking algorithms line up frames in different stacks from the same aerial image. But it is much better than shifting the lens, unless the lens is perfectly telecentric.
The lenses we use
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
- Posts: 3417
- Joined: Sat Nov 20, 2010 10:40 am
- Location: Santa Clara, CA, USA
- Contact:
Ahh, I knew I was missing some simple thing. I was just thinking of camera-lens geometry, not lens-subject geometry. From camera-lens perspective, it does not matter if you shift the lens at the end of the bellows, or the camera, you get the same thing. This is different from tilt, where tilting the lens at end of bellows adds a lot of shift at the same time as tilt. So it's better to tilt at the camera end if all you want is tilt. But of course, if you want to "scan" an aerial image by shift, the lens must stay put.Lou Jost wrote:Ray, I think you are missing the main reason for doing the shift this way. It does not involve tilts at all, and it is important to note that the lens stays in one place while the camera shifts. I am using a lens with a big image circle, and I am shifting the camera across the stationary aerial image.
This has huge advantages over shifting the lens in the normal way. Most importantly, there is no perspective change. This could still go bad when stacking, because of possible variations in the way that stacking algorithms line up frames in different stacks from the same aerial image. But it is much better than shifting the lens, unless the lens is perfectly telecentric.
-
- Posts: 48
- Joined: Thu Aug 18, 2011 9:48 pm
Lou, I have been using similar system that you described. I encounter severe vibration using the Horseman LX.Lou Jost wrote:Mike, there are cheap Fotodiox shift adapters from Nikon mounts to MFT mounts. No need for tilts. Also, the good Nikon bellows have shifts in one direction (and tilts). For scanning a large aerial image it needs to be used backwards (shifting camera rather than lens), as I discussed in the thread. For Nikon cameras that is the best solution. For MFT cameras you can use both the bellows and the shift adapter so you can have three overlapping rows and columns. Both approaches are limited if the aerial image you want to scan is big. That limitation goes away if you use a Nikon sliding back on a view camera, or even a fixed Nikon back (or Sony, both backs are available) on a view camera with geared rear standard. I recently bought a Horeseman 4x5 camera and several kinds of 4x5 Nikon backs (camera of choice mounts directly to it) for this purpose. I can now scan anything and use lenses of any format from MFT to FF to 6x7 to 4" x 5". This will Be especially handy for my photolithography lenses, with 125mm image circle, m=5 and m=10 with NA up to 0.28. The detail in these aerial image is staggering.
Mike and Ray, I've just tested the PN105A using the reversed Nikon shifting bellows. Unfortunately it does not work at m=1, because there is not enough room to extend the bellows enough to use the shift.
So I tried the Scitex 110mm lens instead, which is a much more compact lens. With this lens I need lots of bellows extension to reach m=1, so there is no problem using the shift. There is no vignetting even when fully shifted. The shift adds about 11mm in each direction, so the FOV would be 57mm at 1:1 for this lens.
So I tried the Scitex 110mm lens instead, which is a much more compact lens. With this lens I need lots of bellows extension to reach m=1, so there is no problem using the shift. There is no vignetting even when fully shifted. The shift adds about 11mm in each direction, so the FOV would be 57mm at 1:1 for this lens.
- enricosavazzi
- Posts: 1474
- Joined: Sat Nov 21, 2009 2:41 pm
- Location: Västerås, Sweden
- Contact:
I had a Sinar P that I adapted to Nikon DSLR and toyed with on and off for a few years, but it is not even close in stiffness and precision of movements to good bellows for 135 cameras (unless you need ample tilt and shift movements on both standards or the large format).Lou Jost wrote:Yes, in retrospect I think the Sinar P would have been a sturdier choice.I encounter severe vibration using the Horseman LX.
But I am hoping that the Horseman will be good enough for low magnification work (high magnification does not merit large format treatment).
In the end, when I sold the house and had to get rid of a lot of stuff I just gave the Sinar to a friend as a present.
--ES
That's interesting Enrico. Unfortunately the 35mm bellows are far too small to enable use of the large image circles of micro-photolithography or medium-format lenses. With good vibration control I still hope the large format cameras might be usable....It might also be possible to modify them to increase stiffness. I don't care much about tilts so I could give up some degrees of freedom.
-
- Posts: 3417
- Joined: Sat Nov 20, 2010 10:40 am
- Location: Santa Clara, CA, USA
- Contact:
I need to look back at some of Austrokiwi's old posts. He describes how he does the scanning. I don't remember the equipment but it was a shift adapter of some sort that the camera mounts to.Lou Jost wrote:That's interesting Enrico. Unfortunately the 35mm bellows are far too small to enable use of the large image circles of micro-photolithography or medium-format lenses. With good vibration control I still hope the large format cameras might be usable....It might also be possible to modify them to increase stiffness. I don't care much about tilts so I could give up some degrees of freedom.
I can confirm this. I recently had to buy PTGui Pro because none of the free programs I tried would stitch this 24mm shifted panorama from my recent vacation: https://www.flickr.com/photos/trevartha ... ed-public/mawyatt wrote:Ray,
There is a program that does a much better job at stitching than PS. Rik informed me of it long ago when I was trying to do some early massive chip image stack & stitch work which I couldn't get PS to complete.
It is PTGui.
https://www.ptgui.com/
Best
All of the other programs, including Photoshop, stitched with a wavy horizon.
It was expensive, but worth it. I get to look at the canvas print of this awesome scene every morning with my coffee now.
I still prefer Photoshop for normal stitching, but PTGui is there if I need it.