I'm a new member here but have been lurking for a while. This is an amazing forum, one of the most useful specialty forums I've seen.
I have lots of questions but I will open with my most urgent one. I think I understand why microscope lenses are better than regular reversed lenses. Microscope lenses don't have to cover such a wide field as a camera lens, so they can be better optimized for high sharpness in their narrow field of view at their largest aperture. A camera lens with a large aperture needs to be stopped down a lot to get maximal potential sharpness, and then diffraction kills the sharpness when magnification is high.
But there are a few fast lenses that are really sharp even at their widest or next-to-widest apertures. The Zeiss Otus 50mm f/1.4 is one. The new Sigma 50mm f/1.4 Art lens is almost as good. It is well-corrected and has very little curvature of field.
So my question is: If this were reversed, and stopped down one stop, it would have about the same effective aperture as a 10x/0.25 microscope objective, which is I think equivalent to f/2.0. So could they give equally-good results?
A related question: Could a fast, well-corrected 50mm lens like that, stopped down one stop, mounted reversed on a good Nikon 400mm f5.6 ED lF lens, have an even lower effective aperture at 8x magnification than if the 50mm lens were reversed without the telephoto lens? Might it give results as good as or better than a microscope lens?
Thanks!
Fast flat-field reversed lens versus microscope lens?
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Welcome, Lou!
I don't think anyone has run a direct comparison You may find that chriomatic aberrations are more of a problem than outright reolution.
I'd be surprised if a reversed normal lens would be as good, because they're designed for a field which is so much wider. Technoilogy moves on though, and old rules of thumb go out of the window.
One significant difference is price, of course, Nikon's 4x and 10x were as here
Combos can be excellent, or otherwise.
If you're after the best possible then you "should" be better to use a higher NA objective, such as a 10x NA 0.45.
I don't think anyone has run a direct comparison You may find that chriomatic aberrations are more of a problem than outright reolution.
I'd be surprised if a reversed normal lens would be as good, because they're designed for a field which is so much wider. Technoilogy moves on though, and old rules of thumb go out of the window.
One significant difference is price, of course, Nikon's 4x and 10x were as here
Combos can be excellent, or otherwise.
If you're after the best possible then you "should" be better to use a higher NA objective, such as a 10x NA 0.45.
Thanks for the welcome, Chris. I'm on a budget so a 10x NA 0.42 is beyond me (as is the Otus). But I do want to push the boundaries of my field, orchid taxonomy. My main genus, Lepanthes, has really tiny fake fungus gnat genitalia on their flowers. The pollinating male fly actually mates with the fake genitalia and ejaculates. The exact structure of these fake genitalia is taxonomically important, but hardly ever imaged well.
The two fast 50mm lenses I mentioned are very well corrected for chromatic aberrations. But in my work, I think I can completely eliminate chromatic aberrations, by using blue laser light. This will also push the diffraction limit of any lens up almost one whole stop, because the wavelengths are almost twice as small as some of the wavelengths in white light.
Another idea I had was just to zero all the red values in the RGB image. That alone should eliminate a lot of aberrations. For publication I mainly need black and white photos so the color doesn't matter. Alternatively I might use a bluish-green light so I am using the blue and green sensors in the camera but not the red ones.
Another idea, which might also work for full-color images, was to use white light, but to separate the final R, G, and B values of each image into three different monochromatic images, and stack each color separately, and then merge them. Many chromatic aberrations are caused by colors focusing in different focal planes. But if one makes a stack of just red images, or just blue or green ones, these should be sharp, even if the lens focuses each color in a slightly different plane. Then maybe they could be aligned together after stacking/merging to make a full color image free of chromatic aberration?
By the way, do you know what is the formula for the numerical aperture or f/stop of the combination of a lens mounted in reverse on a telephoto?
The two fast 50mm lenses I mentioned are very well corrected for chromatic aberrations. But in my work, I think I can completely eliminate chromatic aberrations, by using blue laser light. This will also push the diffraction limit of any lens up almost one whole stop, because the wavelengths are almost twice as small as some of the wavelengths in white light.
Another idea I had was just to zero all the red values in the RGB image. That alone should eliminate a lot of aberrations. For publication I mainly need black and white photos so the color doesn't matter. Alternatively I might use a bluish-green light so I am using the blue and green sensors in the camera but not the red ones.
Another idea, which might also work for full-color images, was to use white light, but to separate the final R, G, and B values of each image into three different monochromatic images, and stack each color separately, and then merge them. Many chromatic aberrations are caused by colors focusing in different focal planes. But if one makes a stack of just red images, or just blue or green ones, these should be sharp, even if the lens focuses each color in a slightly different plane. Then maybe they could be aligned together after stacking/merging to make a full color image free of chromatic aberration?
By the way, do you know what is the formula for the numerical aperture or f/stop of the combination of a lens mounted in reverse on a telephoto?
-
rjlittlefield
- Site Admin
- Posts: 23603
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Lou,
Welcome aboard!
Using those fast 50 mm lenses is definitely worth a try, but I would strongly recommend investing in a known good inexpensive 10X microscope objective such as the Nikon CFI BE 10X NA 0.25. An example using that lens (pushed down to 5X) is shown HERE.
My expectation is that the lenses with full frame coverage will not compete favorably with a microscope objective, but unfortunately the only way to know for sure is to try it. Without a head-to-head comparison, you'll be left with deciding whether the lenses you have are good enough, and then you risk investing a lot of time in making images that while "good enough", aren't as good as they could have been for only a little more money. It's a cost/value issue.
You would do better to start with incoherent blue light. LED's should work OK.
1. Relating NA to effective aperture on the same side of the lens:
2. Relating effective aperture at the subject, to effective aperture at the sensor, based on magnification:
3. Magnification with one lens reversed in front of another, with the rear lens focused at infinity:
If the rear lens is not focused at infinity, then there's no good formula because the behavior depends on aspects of the lens design that you don't know.
4. Effective aperture of a lens that is focused only by extension:
(Formula 4 is exact only if the lens has pupil ratio 1. Otherwise it's just an approximation that is pretty good in most cases. The only case where it really falls down is with certain bellows lenses that have pupil ratios far from 1 and are designed to be used non-reversed at high magnification. In that case the formulas http://www.photomacrography.net/forum/v ... =8895]HERE are needed to get accurate calculations.)
For other cases, you can combine these formulas using standard algebra and/or stepwise computations.
For example, an f/2 lens reversed on bellows and operated at 10X, giving f/22, results in an effective aperture at the subject of 22/10 = f/2.2, which is equal to NA 0.227. To get the equivalent of an NA 0.25 objective, by extending an lens on bellows, you would need to start with about f/1.8.
Another example: reversing a 50mm f/2 in front of a 200 mm rear lens focused on infinity. Assuming that the 200 mm rear lens is wide enough that the 50mm f/2 provides the limiting aperture, then the effective aperture on the subject side will be f/2 = NA 0.25. That, combined with the magnification of 4X, will give an effective aperture on the camera side of f/8.
--Rik
Welcome aboard!
Using those fast 50 mm lenses is definitely worth a try, but I would strongly recommend investing in a known good inexpensive 10X microscope objective such as the Nikon CFI BE 10X NA 0.25. An example using that lens (pushed down to 5X) is shown HERE.
My expectation is that the lenses with full frame coverage will not compete favorably with a microscope objective, but unfortunately the only way to know for sure is to try it. Without a head-to-head comparison, you'll be left with deciding whether the lenses you have are good enough, and then you risk investing a lot of time in making images that while "good enough", aren't as good as they could have been for only a little more money. It's a cost/value issue.
Blue light is a good idea. But laser light is not, unless you're prepared to invest in a vibrating diffuser such as these "speckle reducers". Otherwise, you're very likely to end up with nasty speckle patterns superimposed on whatever actual structure your subject has. The problem is that the coherence length of laser light is so long that normal non-moving diffusers don't work as you might expect. A normal diffuser will spread the light out, but it will still be speckly. To make the speckles disappear, you have to move them around (hence the vibrating element) and use a long enough exposure time to catch them in enough different positions that they integrate out.I think I can completely eliminate chromatic aberrations, by using blue laser light.
You would do better to start with incoherent blue light. LED's should work OK.
Stacking the colors separately is an attractive idea, mentioned periodically. But I know only a few people who have tried it, and they did not report favorable results. I don't know where the disconnect occurs between theory and practice.Another idea, which might also work for full-color images, was to use white light, but to separate the final R, G, and B values of each image into three different monochromatic images, and stack each color separately, and then merge them. Many chromatic aberrations are caused by colors focusing in different focal planes. But if one makes a stack of just red images, or just blue or green ones, these should be sharp, even if the lens focuses each color in a slightly different plane. Then maybe they could be aligned together after stacking/merging to make a full color image free of chromatic aberration?
Here are the formulas that I think you need:By the way, do you know what is the formula for the numerical aperture or f/stop of the combination of a lens mounted in reverse on a telephoto?
1. Relating NA to effective aperture on the same side of the lens:
Example: an f/2 lens, focused at infinity, is NA 0.25 on the sensor side. 0.25 = 1/(2*2)NA = 1/(2*F_eff)
F_eff = 1/(2*NA)
2. Relating effective aperture at the subject, to effective aperture at the sensor, based on magnification:
Example: microscope objective at NA 0.25, operating at 10X, gives NA 0.025 = f/20 at the sensor.NA_subject = mag*NA_sensor
F_eff_sensor = mag*F_eff_subject
3. Magnification with one lens reversed in front of another, with the rear lens focused at infinity:
Example: 50X lens reversed in front of 200 mm lens, with the 200 focused at infinity, gives magnification 4 = 200/50.magnification = FL_rear / FL_front
If the rear lens is not focused at infinity, then there's no good formula because the behavior depends on aspects of the lens design that you don't know.
4. Effective aperture of a lens that is focused only by extension:
Example: an f/2 lens reversed on bellows and operated at 10X, gives effective aperture f/22 at the sensor.F_eff_sensor = F_nominal * (magnification+1)
(Formula 4 is exact only if the lens has pupil ratio 1. Otherwise it's just an approximation that is pretty good in most cases. The only case where it really falls down is with certain bellows lenses that have pupil ratios far from 1 and are designed to be used non-reversed at high magnification. In that case the formulas http://www.photomacrography.net/forum/v ... =8895]HERE are needed to get accurate calculations.)
For other cases, you can combine these formulas using standard algebra and/or stepwise computations.
For example, an f/2 lens reversed on bellows and operated at 10X, giving f/22, results in an effective aperture at the subject of 22/10 = f/2.2, which is equal to NA 0.227. To get the equivalent of an NA 0.25 objective, by extending an lens on bellows, you would need to start with about f/1.8.
Another example: reversing a 50mm f/2 in front of a 200 mm rear lens focused on infinity. Assuming that the 200 mm rear lens is wide enough that the 50mm f/2 provides the limiting aperture, then the effective aperture on the subject side will be f/2 = NA 0.25. That, combined with the magnification of 4X, will give an effective aperture on the camera side of f/8.
--Rik
Rik, thanks very much for your answers, and for your work on the rest of this forum.
What if the long lens is f5.6? I guess then that would be the limiting aperture. But with a reversed lens in front, is the sensor even seeing that f/5.6 aperture? Would the limiting aperture still be that of the front lens, reduced by (1+m)?
I think a little warm milk will do the trick there. Milk is a colloidal suspension of fast-moving particles that scatter blue light. Just add water to taste, pass the beam through it, and it will be ready to go. But you are right, a blue LED would be simpler.Blue light is a good idea. But laser light is not, unless you're prepared to invest in a vibrating diffuser such as these "speckle reducers".
So a reversed lens that is sharp at f/2 is theoretically competitive (or even better) than a 0.25 microscope objective, at magnifications up to about x8. At least, the laws of physics don't decide the issue any more, like they used to when there weren't any affordable fast lenses that were sharp wide open....An f/2 lens reversed on bellows and operated at 10X, giving f/22, results in an effective aperture at the subject of 22/10 = f/2.2, which is equal to NA 0.227.
Under that assumption about the 200mm lens, the effective aperture of this combination is very much larger than the diffraction limit, and it will again be competitive with the microscope objective. But it will cover a far wider area on the sensor without corner degradation.Assuming that the 200 mm rear lens is wide enough that the 50mm f/2 provides the limiting aperture, then the effective aperture on the subject side will be f/2 = NA 0.25. That, combined with the magnification of 4X, will give an effective aperture on the camera side of f/8.
What if the long lens is f5.6? I guess then that would be the limiting aperture. But with a reversed lens in front, is the sensor even seeing that f/5.6 aperture? Would the limiting aperture still be that of the front lens, reduced by (1+m)?
Rik and Chris, I should add that I did already invest in some cheap Nikon microscope objectives, the Nikon E Plan 10/0.25 160/- and the Nikon M Plan 10/0.25 210/0. Both have rather shorter working distances than I prefer, though, so I am about to buy the Mitutoyo 10x after studying the posts on this forum. I already have both Raynox lenses that Rik has recommended for this objective, so I will use them as tube lenses. I also bought some cheap lenses whose glass I will remove to make long rigid focusing extension tubes for the Raynox lenses.
I also will buy a crap lens to use as a stage for my subjects. I will be able to raise and lower the stage by focusing it, and will be able to illuminate it from the bottom through the lens.
I also will buy a crap lens to use as a stage for my subjects. I will be able to raise and lower the stage by focusing it, and will be able to illuminate it from the bottom through the lens.
-
rjlittlefield
- Site Admin
- Posts: 23603
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
You're very welcome. Thank you for the kind words.Rik, thanks very much for your answers, and for your work on the rest of this forum.
In the center of the frame, surely the sensor will not see the f/5.6 rear aperture. The limiting aperture in that case will be the front lens, reduced by a factor of m, not (1+m).Under that assumption about the 200mm lens, the effective aperture of this combination is very much larger than the diffraction limit, and it will again be competitive with the microscope objective. But it will cover a far wider area on the sensor without corner degradation.Assuming that the 200 mm rear lens is wide enough that the 50mm f/2 provides the limiting aperture, then the effective aperture on the subject side will be f/2 = NA 0.25. That, combined with the magnification of 4X, will give an effective aperture on the camera side of f/8.
What if the long lens is f5.6? I guess then that would be the limiting aperture. But with a reversed lens in front, is the sensor even seeing that f/5.6 aperture? Would the limiting aperture still be that of the front lens, reduced by (1+m)?
Away from the center of the frame, the sensor may start seeing the rear aperture in addition to the front aperture. This is a form of vignetting. How bad it gets depends on how far inside each lens the respective apertures appear to be (pupil positions). One of the reasons that the Raynox DCR-150 works so well on full frame is that its pupil is far forward so it is very resistant to vignetting.
If you have occasion to test the objectives head-to-head against say the Sigma 50mm F1.4 Art lens, we would of course be very interested to see the results.
--Rik