Stopping down objectives – useful or not? (images added)
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Stopping down objectives – useful or not? (images added)
I don't know if this is very useful but I thought it might be worth investigating.
A couple of days ago I slaughtered an old enlarger (presumably) lens I bought for £2 – AICO 35mm f3.5.
The iris assembly conveniently turned out to have RMS (or something close enough) threads in both ends. (the black mount with the aperture prints also turned out to work nicely as a recessed RMS->M39 adapter).
With the dirt cheap Zeiss 3.2X (NA=?) it seemed to deliver quite decent performance, handholding in the 2-5X range. Examples:
coffee bean 1.8X (vingetting is due to the small image circle of the objective – the iris does not affect this as far as I can tell)
pencil tip 2.5X
needle eye 3.8X
drone fly wing 3.8X
All hand held, stopped down (difficult to tell by how much at this point), single exposures (no stacking).
A couple of days ago I received a Nikon BD M Plan 10/0.25 (210/0) and just happened to mount it onto the same adjustable iris (shown in the image above). Since the rear element of this objective is much further in front of the iris than with the humble Zeiss I expected the iris to act more like a mask than an aperture - introducing vingetting when closed down rather than actually stopping it down. But it turned out to work – stopping down increased DOF and diffraction softening without introducing any visible vingetting. If it works like this – without too much compromises when it comes to image quality (disregarding diffraction) – I personally find it quite useful since I'm comfortable dealing with the old DOF-diffraction trade-off.
I need to do more tests before I know if the image quality suffers too much but it would also help a lot if I knew more what to expect and what to look for. I'm sure some of you have played around with similar solutions – experiences? With Riks Stopping down a lens combo in mind it doesn't seem unlikely that the distance between objective and the aperture is a critical factor – any good way to figure out what this distance should be for a particular objective? Perhaps the ideal position would be somewhere inside the objective barrel but when that is not a possibility, like here, does it mean the next best option is always as close to the rear element as possible?
Again – all input on the subject appreciated!
A couple of days ago I slaughtered an old enlarger (presumably) lens I bought for £2 – AICO 35mm f3.5.
The iris assembly conveniently turned out to have RMS (or something close enough) threads in both ends. (the black mount with the aperture prints also turned out to work nicely as a recessed RMS->M39 adapter).
With the dirt cheap Zeiss 3.2X (NA=?) it seemed to deliver quite decent performance, handholding in the 2-5X range. Examples:
coffee bean 1.8X (vingetting is due to the small image circle of the objective – the iris does not affect this as far as I can tell)
pencil tip 2.5X
needle eye 3.8X
drone fly wing 3.8X
All hand held, stopped down (difficult to tell by how much at this point), single exposures (no stacking).
A couple of days ago I received a Nikon BD M Plan 10/0.25 (210/0) and just happened to mount it onto the same adjustable iris (shown in the image above). Since the rear element of this objective is much further in front of the iris than with the humble Zeiss I expected the iris to act more like a mask than an aperture - introducing vingetting when closed down rather than actually stopping it down. But it turned out to work – stopping down increased DOF and diffraction softening without introducing any visible vingetting. If it works like this – without too much compromises when it comes to image quality (disregarding diffraction) – I personally find it quite useful since I'm comfortable dealing with the old DOF-diffraction trade-off.
I need to do more tests before I know if the image quality suffers too much but it would also help a lot if I knew more what to expect and what to look for. I'm sure some of you have played around with similar solutions – experiences? With Riks Stopping down a lens combo in mind it doesn't seem unlikely that the distance between objective and the aperture is a critical factor – any good way to figure out what this distance should be for a particular objective? Perhaps the ideal position would be somewhere inside the objective barrel but when that is not a possibility, like here, does it mean the next best option is always as close to the rear element as possible?
Again – all input on the subject appreciated!
Last edited by morfa on Thu Apr 08, 2010 2:03 pm, edited 1 time in total.
Now I've made some quick test stacks and think I'm ready to draw at least some conclusions.
In my shot of the Ladybird (http://www.photomacrography.net/forum/v ... php?t=9317) made with the Nikon BD M Plan slightly stopped down Chris and Rik pointed out some CA visible in parts of the image and the question was raised weather this could be caused by the iris. Finding an answer to this was my main goal with this test.
I made three stacks:
A) With the iris wide open
B) With the iris stopped down approximately 1 stop
C) Without the iris (but at the same total extension)
Here are the the resulting images (6MP interpolated to 12MP in camera (Fujifilm S5pro) and otherwise untouched).
A: http://farm3.static.flickr.com/2703/447 ... 668e_o.jpg
B: http://farm3.static.flickr.com/2799/447 ... 6749_o.jpg
C: http://farm3.static.flickr.com/2712/447 ... 7681_o.jpg
Not perfect images – for one thing I forgot the focus aid light on and the batteries drained between the first and the last stack so there are some differences in lighting because of this.
But I think they serve their purpose and it's pretty clear that
I) Stopping down introduces more CA
II) Even fully open the iris deteriorates the image to some degree
Edge crop:
I'm not surprised with (II) since even fully open the aperture opening is just about the same size as the rear element and there are several centimeters separating them.
In my shot of the Ladybird (http://www.photomacrography.net/forum/v ... php?t=9317) made with the Nikon BD M Plan slightly stopped down Chris and Rik pointed out some CA visible in parts of the image and the question was raised weather this could be caused by the iris. Finding an answer to this was my main goal with this test.
I made three stacks:
A) With the iris wide open
B) With the iris stopped down approximately 1 stop
C) Without the iris (but at the same total extension)
Here are the the resulting images (6MP interpolated to 12MP in camera (Fujifilm S5pro) and otherwise untouched).
A: http://farm3.static.flickr.com/2703/447 ... 668e_o.jpg
B: http://farm3.static.flickr.com/2799/447 ... 6749_o.jpg
C: http://farm3.static.flickr.com/2712/447 ... 7681_o.jpg
Not perfect images – for one thing I forgot the focus aid light on and the batteries drained between the first and the last stack so there are some differences in lighting because of this.
But I think they serve their purpose and it's pretty clear that
I) Stopping down introduces more CA
II) Even fully open the iris deteriorates the image to some degree
Edge crop:
I'm not surprised with (II) since even fully open the aperture opening is just about the same size as the rear element and there are several centimeters separating them.
Thinking back to Rik's posting re: coupled lenses showing how the position of the aperture changes the path of the light through the lens (he linked to it in a recent post).
You have changed the position of the limiting aperture of the lens by adding the aperture on the back and are probably inducing new/worse aberrations into the system.
Interesting experiment.
You have changed the position of the limiting aperture of the lens by adding the aperture on the back and are probably inducing new/worse aberrations into the system.
Interesting experiment.
- Charles Krebs
- Posts: 5865
- Joined: Tue Aug 01, 2006 8:02 pm
- Location: Issaquah, WA USA
- Contact:
John,
First would be using the older Nikon objectives made for their "Universal Stage". These are long working distance and they actually have an adjustable aperture built into the objective. (Apparently this was needed for some function using one of these exotic universal stages). In any event, my experience was that just about any amount of "stopping down" was clearly noticeable in a reduction of resolution. The second case is with more modern, high quality, high NA microscope objectives that also have a diaphragm built in for slightly adjusting the aperture to achieve proper darkfield illumination with dedicated darkfield condensers. Here again, it was very obvious that if I accidentally "stopped down" in brightfield illumination I could clearly see the reduction in resolution.
The third case was when I added a slightly smaller "waterhouse stop" (simply a black mask with a circular hole in it) to the back of my 20/0.70 S Plan Apo. This was done to see if I could get better results in darkfield, since the high NA of this objective (for it's magnification) made darkfield with 3-dimensional subjects very difficult (huge out-of-focus areas that were brightly illuminated). This actually worked OK. It did seem to reduce resolution a little, but it also helped a great deal in "shrinking" the huge OOF areas. (In the "old days" there were little devices called funnel-stops available for some objectives. These fit into the rear opening of the objective in order to do an aperture reduction for darkfield). I tried my simple "mask" on other objectives and did not get the desired effect, just some vignetting. It seemed to depend on the optical design of the objective, and whether or not an aperture placed at the rear is close enough to what would be (optically... I'll let Rik sketch it out ) the appropriate location for a diaphragm.
My take was this...
With a compound microscope I was adding typically another 1.67-2.5X magnification by using the projection eyepiece needed to place the image on the sensor. With the relatively low NA on the Nikon universal objectives; and because of the high magnifications of my modern high NA objectives with the built in aperture (40X and 100X) I was already, even with the apertures fully open, in the realm where I was losing more resolution to diffraction than I would like. Additional stopping down, even slightly, caused a noticeable loss in the resolution I was getting. With my 20/0.70 I actually had a little "room" to play with. I could stop down a little before the resolution really crashed from diffraction. If I looked closely I could see some loss, but the benefit in the appearance of the darkfield image made the trade-off worthwhile.
My hunch is that with something like a 10/0.30 or (even a 10/0.25) used directly on bellows (with no addition magnificatioon added) there may be times when -- if you can get the aperture in the proper location-- a slight stopping down would result in an acceptable trade-off between DOF and resolution. (Roughly the "effective aperture" of a 10/0.30, at 10X on bellows is f16.7, and a 10/0.25 is f20. So you can see that there may be a small amount of room to make such a compromise).
(edit.. John, right after I posted this message I moved on and saw your post in the image gallery section where you already suggest exactly what I posted below... I think a vaguely remember someone trying something like this with a lens that had an aperture. Might have been lauriek, or you! Definitely worth exploring, so I'll leave it here as well so hopefully more eyes will see, and maybe experiment, with it)
Even if you did not want to take the resolution hit I could see a possible use. Imagine a stack like your ladybug. You don't want to lose anything across the face, but as you get to the back it would be nice to have a more "natural" fall-off in focus (instead of the dramatic and unnatural fall-off you sometimes get if you stop running the stack to the back). It seems it should be possible to run the section you want sharpest at full aperture and then gradually stop down and run larger steps as you get to the rear. A bit complicated but it's an interesting thought.
I've not tried anything like you have here (you come up with some great stuff! ) , but in a couple of ways have a little experience with "stopping down" microscope objectives when used on a microscope.Again – all input on the subject appreciated!
First would be using the older Nikon objectives made for their "Universal Stage". These are long working distance and they actually have an adjustable aperture built into the objective. (Apparently this was needed for some function using one of these exotic universal stages). In any event, my experience was that just about any amount of "stopping down" was clearly noticeable in a reduction of resolution. The second case is with more modern, high quality, high NA microscope objectives that also have a diaphragm built in for slightly adjusting the aperture to achieve proper darkfield illumination with dedicated darkfield condensers. Here again, it was very obvious that if I accidentally "stopped down" in brightfield illumination I could clearly see the reduction in resolution.
The third case was when I added a slightly smaller "waterhouse stop" (simply a black mask with a circular hole in it) to the back of my 20/0.70 S Plan Apo. This was done to see if I could get better results in darkfield, since the high NA of this objective (for it's magnification) made darkfield with 3-dimensional subjects very difficult (huge out-of-focus areas that were brightly illuminated). This actually worked OK. It did seem to reduce resolution a little, but it also helped a great deal in "shrinking" the huge OOF areas. (In the "old days" there were little devices called funnel-stops available for some objectives. These fit into the rear opening of the objective in order to do an aperture reduction for darkfield). I tried my simple "mask" on other objectives and did not get the desired effect, just some vignetting. It seemed to depend on the optical design of the objective, and whether or not an aperture placed at the rear is close enough to what would be (optically... I'll let Rik sketch it out ) the appropriate location for a diaphragm.
My take was this...
With a compound microscope I was adding typically another 1.67-2.5X magnification by using the projection eyepiece needed to place the image on the sensor. With the relatively low NA on the Nikon universal objectives; and because of the high magnifications of my modern high NA objectives with the built in aperture (40X and 100X) I was already, even with the apertures fully open, in the realm where I was losing more resolution to diffraction than I would like. Additional stopping down, even slightly, caused a noticeable loss in the resolution I was getting. With my 20/0.70 I actually had a little "room" to play with. I could stop down a little before the resolution really crashed from diffraction. If I looked closely I could see some loss, but the benefit in the appearance of the darkfield image made the trade-off worthwhile.
My hunch is that with something like a 10/0.30 or (even a 10/0.25) used directly on bellows (with no addition magnificatioon added) there may be times when -- if you can get the aperture in the proper location-- a slight stopping down would result in an acceptable trade-off between DOF and resolution. (Roughly the "effective aperture" of a 10/0.30, at 10X on bellows is f16.7, and a 10/0.25 is f20. So you can see that there may be a small amount of room to make such a compromise).
(edit.. John, right after I posted this message I moved on and saw your post in the image gallery section where you already suggest exactly what I posted below... I think a vaguely remember someone trying something like this with a lens that had an aperture. Might have been lauriek, or you! Definitely worth exploring, so I'll leave it here as well so hopefully more eyes will see, and maybe experiment, with it)
Even if you did not want to take the resolution hit I could see a possible use. Imagine a stack like your ladybug. You don't want to lose anything across the face, but as you get to the back it would be nice to have a more "natural" fall-off in focus (instead of the dramatic and unnatural fall-off you sometimes get if you stop running the stack to the back). It seems it should be possible to run the section you want sharpest at full aperture and then gradually stop down and run larger steps as you get to the rear. A bit complicated but it's an interesting thought.
- rjlittlefield
- Site Admin
- Posts: 23561
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
That's the same one that John linked. He just used a different form of the url tag, that showed the title instead of the URL.
Stopping down an objective is something that I have just barely played with.
I presume that the best place to stop down would be at the position of the current limiting aperture. That is probably near the rear element of the objective -- not a very accessible position. Alternatively, it might give a better image to stop down just in front of the objective, but of course this cuts into working distance. I have played with adding an aperture made of aluminum foil at the front of an objective, but this is not easy to adjust or to center.
I see one other issue not mentioned so far. When the added aperture is located somewhere different from the original limiting aperture, then adding the aperture will also change the perspective. If the change is drastic enough, this could cause problems in stacking images shot under the two different conditions.
John, thanks for clarifying that the earlier images were shot with an added aperture that was oddly shaped and off center. That could easily explain why the CA was not symmetric.
BTW, some people may be wondering how just adding an aperture -- with no refracting elements -- could possibly add CA to an image. The answer is that the aperture does not actually add CA, it just increases the importance of CA that was already present. This happens because the aperture removes the contributions of other light that took better behaved paths through the lens. From the standpoint of the final image there is no difference, but perhaps this explanation will be more comfortable.
--Rik
- enricosavazzi
- Posts: 1474
- Joined: Sat Nov 21, 2009 2:41 pm
- Location: Västerås, Sweden
- Contact:
Re: Stopping down objectives – useful or not?
A number of macro-zoom lenses for industrial macro/microscopes made by Optem, Navitar, Unitron and Meiji (and probably other brands) are normally sold without an iris, but a built-in iris can be purchased as an option (these are special models, the iris cannot be added to a lens without one). In all cases I am aware of, the iris is mounted at the rear of all optical groups. The same is true of the Wild/Leica M400/420/450 macroscopes, which have the iris in the body, at the rear of the zoom objective but in front of the tube lens. I have also seen Lomo low-power microscope objectives with an iris at the rear of the optics. So what you have found confirms that the iris in a photomacrography lens does not need to be placed between lens elements but can be placed at the rear.morfa wrote:
A couple of days ago I received a Nikon BD M Plan 10/0.25 (210/0) and just happened to mount it onto the same adjustable iris (shown in the image above). Since the rear element of this objective is much further in front of the iris than with the humble Zeiss I expected the iris to act more like a mask than an aperture - introducing vingetting when closed down rather than actually stopping it down. But it turned out to work – stopping down increased DOF and diffraction softening without introducing any visible vingetting. If it works like this – without too much compromises when it comes to image quality (disregarding diffraction) – I personally find it quite useful since I'm comfortable dealing with the old DOF-diffraction trade-off.
This opens the possibility of adding an iris to many lenses for this application. This does not always work, especially at low magnification and with certain lens designs. In these exceptions, the iris causes a vignetting or darkening of the image corners. In some cases an iris in front of the lens does work, but this reduces the available working distance.
Of course this iris does increase diffraction by decreasing the NA of the objective, and may introduce chromatic and other aberrations. This is probably why most macro-zoom objectives for industrial microscopes are sold without an iris. Incidentally, these zoom lenses have a mobile group of elements that moves all the way between the front and rear elements, with no space left for an internal iris, so the rear position may be the only feasible one from a design point of view.
--ES
That was probably me, I don't remember the particular thread here but there's a mini article mentioning the technique on my website - http://www.laurieknight.net/article/view/6Charles Krebs wrote: (edit.. John, right after I posted this message I moved on and saw your post in the image gallery section where you already suggest exactly what I posted below... I think a vaguely remember someone trying something like this with a lens that had an aperture. Might have been lauriek, or you! Definitely worth exploring, so I'll leave it here as well so hopefully more eyes will see, and maybe experiment, with it)
- rjlittlefield
- Site Admin
- Posts: 23561
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Remember that for NA, bigger number is a wider aperture. Just the opposite of f-number.elf wrote:I would have expected the 0.30 to have a smaller effective aperture than the 0.25. Is the 0.30 a better lens than the 0.25 (Nikon BD Plans)?
Roughly speaking, f-number = 1/(2*NA).
That formula is exact, if "f-number" is taken to mean effective or working f-number from the standpoint of the subject, and if the objective is being used at its design point.
It is only approximate, as a way of comparing one lens specified as NA to another lens specified as f-number.
The exact formulas that relate specified NA to specified f-number also involve magnification and pupil factors. Those are too complicated to be useful for most purposes.
--Rik
I've now run a quick experiment with the "combined f-stop stacking technique" briefly discussed here and elsewhere.
First I shot a regular 38 image stack. After the final image I stopped down approximately three stops and took an additional image.
Then I ran two PMax stacks in Zerene Stacker: one with, and one without the final stopped down image.
The effect I was after here was to get a more gentle transition from sharp to blurry as well as a slightly more defined out of focus region.
I could immediately see the desired effect in the 39 image stack but the final, stopped down image also added some ugly "double vision halos" as shown here:
When I played around with the alignment/scaling parameters (turning them on and off or changing the numbers) these artifacts changed but never seemed to disappear.
Perhaps Rik was onto the explanation for this even before the problem presented itself:
On the left you see the "standard" 38 image stack and on the right is the one with the 39th stopped down exposure added. Click the images for 1024px views. The objective used was a Nikon E Plan 10x/0.25 160/- (I used a different iris than the one shown above and the 10x BD M Plan wouldn't fit in this one).
It may not seem like a significant difference but for me this is definitely enough to keep me interested in the technique and experiment more with it. The effect could quite easily be varied/amplified by using more than one stopped down exposure (separating them with small focus increments) so I think this method has the potential of giving us much better control over the out of focus regions of our stacks.
First I shot a regular 38 image stack. After the final image I stopped down approximately three stops and took an additional image.
Then I ran two PMax stacks in Zerene Stacker: one with, and one without the final stopped down image.
The effect I was after here was to get a more gentle transition from sharp to blurry as well as a slightly more defined out of focus region.
I could immediately see the desired effect in the 39 image stack but the final, stopped down image also added some ugly "double vision halos" as shown here:
When I played around with the alignment/scaling parameters (turning them on and off or changing the numbers) these artifacts changed but never seemed to disappear.
Perhaps Rik was onto the explanation for this even before the problem presented itself:
Anyway, since the good parts of the 39 image stack in this case were easily separated from the problematic parts I opened the two stacks as layers in a single Photoshop document and simply masked out the ugly bits. Now I had two images that were identical except from the out of focus areas:rjlittlefield wrote: I see one other issue not mentioned so far. When the added aperture is located somewhere different from the original limiting aperture, then adding the aperture will also change the perspective. If the change is drastic enough, this could cause problems in stacking images shot under the two different conditions.
On the left you see the "standard" 38 image stack and on the right is the one with the 39th stopped down exposure added. Click the images for 1024px views. The objective used was a Nikon E Plan 10x/0.25 160/- (I used a different iris than the one shown above and the 10x BD M Plan wouldn't fit in this one).
It may not seem like a significant difference but for me this is definitely enough to keep me interested in the technique and experiment more with it. The effect could quite easily be varied/amplified by using more than one stopped down exposure (separating them with small focus increments) so I think this method has the potential of giving us much better control over the out of focus regions of our stacks.
Repeated the test: same subject and gear, new stack.
Top left: PMax(38 wide open)
Top right: PMax(38 wide open + 1 stopped down)
Bottom left: PMax(38 wide open + 4 stopped down)
Bottom right: PMax(PMax(38 wide open) + PMax(12 stopped down))
No retouching at all – all artifacts left. Click image for 11MP view of the composite.
Top left: PMax(38 wide open)
Top right: PMax(38 wide open + 1 stopped down)
Bottom left: PMax(38 wide open + 4 stopped down)
Bottom right: PMax(PMax(38 wide open) + PMax(12 stopped down))
No retouching at all – all artifacts left. Click image for 11MP view of the composite.