Theoretical question on image circles

[jerseyinHK]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?


Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

 

[bokesan]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

That does not work, because the focus would be far beyond infinity.

However, there are adapters with optics that do exactly what you're thinking of. MetaBones called them "Speed Boosters." Hasselblad also offers one, the XH Converter 0.8.

- Chris

 

[Alexander6x6]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

You would just illustrate the lens optical design.

 

[Alexander6x6]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

That does not work, because the focus would be far beyond infinity.

However, there are adapters with optics that do exactly what you're thinking of. MetaBones called them "Speed Boosters." Hasselblad also offers one, the XH Converter 0.8.

- Chris

An optical reducer might introduce weird artifacts as I described in other discussion:

https://www.dpreview.com/forums/post/68251649

 

[dbateman]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

Lenses are designed with a specific flange back focal distance or back focus distance (for projector lenses). This is the distance from the flange or rear most element to the film or camera sensor. If you are greater than this distance, you are into macro. If you are less than this distance you are beyond infinity (were the aliens live or the twilight zone).

Image circle doesn't matter. I have been looking for a sharp 50mm lens with 32mm image circle. That will project a full circle on the gfx sensor.

 

[Oren Grad]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

Lenses are designed with a specific flange back focal distance or back focus distance (for projector lenses). This is the distance from the flange or rear most element to the film or camera sensor. If you are greater than this distance, you are into macro. If you are less than this distance you are beyond infinity (were the aliens live or the twilight zone).

Image circle doesn't matter. I have been looking for a sharp 50mm lens with 32mm image circle. That will project a full circle on the gfx sensor.

To be clear: image circle expands with focus extension, by simple geometric projection. Large format photograpers, who use lenses that are not dedicated to specific camera lines and need to think about which lenses will cover which formats with which degrees of subject magnification, are well aware of this.

But as you point out, if you reduce the extension to less than the flange focal distance "designed into" the lens, you cannot achieve focus at all.

 

[Rob de Loe]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

As others have said, the lens needs to be at its designed flange focal distance to be focused at infinity. Double that distance is 1:1. Closer than flange focal distance is a potentially attractive blurry mess.

This is another example of an essential point that people getting into lens adaptation often miss: the lens design doesn't change when you put it on a different sensor.

The first two sections on this page may clear up some of your questions (or may just raise more!) https://www.robdeloephotography.com/Pages/Lenses

 

[Oren Grad]

This is another example of an essential point that people getting into lens adaptation often miss: the lens design doesn't change when you put it on a different sensor.

There is a subtle but important qualification to this point, which you understand very well and have wrestled with in your own adaptations: the varied sensor cover glasses on different digital cameras mean that the optical system that delivers an image to the sensor amounts to the lens plus cover glass, and the lens design does need to change for optimal results. "Design change" can include the after-the-fact adaptation of changing optical group spacing, or can be what the lens designer does up front with his computer in optimizing the design for a particular sensor sandwich.

 

[Rob de Loe]

This is another example of an essential point that people getting into lens adaptation often miss: the lens design doesn't change when you put it on a different sensor.

There is a subtle but important qualification to this point, which you understand very well and have wrestled with in your own adaptations: the varied sensor cover glasses on different digital cameras mean that the optical system that delivers an image to the sensor amounts to the lens plus cover glass, and the lens design does need to change for optimal results. "Design change" can include the after-the-fact adaptation of changing optical group spacing, or can be what the lens designer does up front with his computer in optimizing the design for a particular sensor sandwich.

Yes indeed. I've had to go the cell spacing adjustment route, but I believe early adopters of wide Leica M lenses on Sony A7-type cameras also faced this problem and solved it with correction "filters" screwed onto the front.

I never tried correction filters on the wide symmetrical lenses I adapted, so I don't know if that would have worked. My gut feeling is that cell spacing adjustment was the best choice -- the evidence for that being that 10 micron differences in cell spacing made a photographically significant different; the correction filter route always struck me as making do with the available filters.

So far, I've only seen this cover glass problem in the context of wide range finder lenses and wide symmetrical lenses. Retrofocus designs for "system" cameras like Pentax 67 and Pentax 645 seem to be fine.

 

[Oren Grad]

I never tried correction filters on the wide symmetrical lenses I adapted, so I don't know if that would have worked. My gut feeling is that cell spacing adjustment was the best choice -- the evidence for that being that 10 micron differences in cell spacing made a photographically significant different; the correction filter route always struck me as making do with the available filters.

I do wonder whether adjusting the cell spacing also amounts to making do, not that you have much practical alternative. That is, if you were starting from scratch and optimzing for different sensor stacks, would cell spacing be the only difference in the designs, or would something more fundamental about the elements/groups be adjusted too? Related, what exactly are independent lens manufacturers like Sigma, Tamron or Cosina doing when they offer a given lens in multiple mounts and claim to be optimzing for each one? Are they truly optimzing or are they compromising with a simpler adjustment like a cell spacing? Or maybe there is no further benefit to be gained by a more complex redesign, or the benefit is so small as to be practically not relevant.

I hope one of our optical experts can shed some light here.

 

[Rob de Loe]

I never tried correction filters on the wide symmetrical lenses I adapted, so I don't know if that would have worked. My gut feeling is that cell spacing adjustment was the best choice -- the evidence for that being that 10 micron differences in cell spacing made a photographically significant different; the correction filter route always struck me as making do with the available filters.

I do wonder whether adjusting the cell spacing also amounts to making do, not that you have much practical alternative. That is, if you were starting from scratch and optimzing for different sensor stacks, would cell spacing be the only difference in the designs, or would something more fundamental about the elements/groups be adjusted too? Related, what exactly are independent lens manufacturers like Sigma, Tamron or Cosina doing when they offer a given lens in multiple mounts and claim to be optimzing for each one? Are they truly optimzing or are they compromising with a simpler adjustment like a cell spacing? Or maybe there is no further benefit to be gained by a more complex redesign, or the benefit is so small as to be practically not relevant.

I hope one of our optical experts can shed some light here.

It's a good question Oren. Adjusting the cell spacing is definitely making do. It's problematic because I've now locked the lens to a specific camera design. For example, I can no longer use my adjusted lenses on a Hasselblad, or with film. If I switched from Fuji to Hasselblad, it would be a hassle but I could simply re-adjust the lenses, but it's not possible to run multiple systems.

As to manufacturers making one design for multiple systems, my hunch is that if the design is retrofocus, they won't have to worry about calibrating for different cameras. I have a Samyang 24/3.5 T-S lens. Samyang's approach is simply to bolt the same basic front part onto different rear parts that contain the movement mechanisms; the length of the rear part varies based on the system's flange focal distance. This is a clever way to keep costs down. Do they customize the optics for each camera system? That I do not know.

 

[dbateman]

What would happen if you mounted an lens with a larger image circle closer to the sensor than it is designed? Think of a macro helicoid, but in reverse (decreasing the distance rather than increasing it).

As an example, a Pentax 67 lens has a larger image circle than the GFX. If the lens was mounted closer to the sensor to where the image circle matched that of the GFX, would the sharpness and transmitted light increase, maybe at the expense of the minimum focus distance?

Would this work, or would there be color fringing and other shifts? I don't have the setup to try it, but wondering if one of you fellows with a view camera has done something like this.

Lenses are designed with a specific flange back focal distance or back focus distance (for projector lenses). This is the distance from the flange or rear most element to the film or camera sensor. If you are greater than this distance, you are into macro. If you are less than this distance you are beyond infinity (were the aliens live or the twilight zone).

Image circle doesn't matter. I have been looking for a sharp 50mm lens with 32mm image circle. That will project a full circle on the gfx sensor.

To be clear: image circle expands with focus extension, by simple geometric projection. Large format photograpers, who use lenses that are not dedicated to specific camera lines and need to think about which lenses will cover which formats with which degrees of subject magnification, are well aware of this.

But as you point out, if you reduce the extension to less than the flange focal distance "designed into" the lens, you cannot achieve focus at all.

Generally yes, but not always. Lens designs can be quite complex, depending on floating elements and group shifts for zoom lenses. I have at least one lens that's image circle actually decreases as you focus closer.

The rear throat of the lens can also cause hard vignetting as you focus closer. Many macro lenses suffer from this and need Gf-mount extension tube to avoid hard vignetting. The Sigma 105mm macro lens, Loawa 100mm 2x macro and Nikon 105mm f2.8 AF-D macro lenses suffer from this, that I own. I will use an 18mm extension tube to avoid the hard vignetting. It would be nice if I could find a shorter tube, as the Loawa doesn't need much extension.

 

[Oren Grad]

To be clear: image circle expands with focus extension, by simple geometric projection. Large format photograpers, who use lenses that are not dedicated to specific camera lines and need to think about which lenses will cover which formats with which degrees of subject magnification, are well aware of this.

But as you point out, if you reduce the extension to less than the flange focal distance "designed into" the lens, you cannot achieve focus at all.

Generally yes, but not always. Lens designs can be quite complex, depending on floating elements and group shifts for zoom lenses. I have at least one lens that's image circle actually decreases as you focus closer.

The rear throat of the lens can also cause hard vignetting as you focus closer. Many macro lenses suffer from this and need Gf-mount extension tube to avoid hard vignetting. The Sigma 105mm macro lens, Loawa 100mm 2x macro and Nikon 105mm f2.8 AF-D macro lenses suffer from this, that I own. I will use an 18mm extension tube to avoid the hard vignetting. It would be nice if I could find a shorter tube, as the Loawa doesn't need much extension.

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

 

[Alexander6x6]

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

This statement must be supported by some examples.
Here is the lens design of 40mm Distagon released in 1994:

 

[Oren Grad]

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

This statement must be supported by some examples.
Here is the lens design of 40mm Distagon released in 1994:

I wrote "large format", not "medium format", though Fuji is certainly doing its best to muddy the waters with its recent marketing. You will not find internal focus or floating element designs in lenses designed and marketed for use with 4x5 and larger film.

 

[JimKasson]

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

This statement must be supported by some examples.
Here is the lens design of 40mm Distagon released in 1994:

Extinction resolution. Not a metric particularly applicable to real world photography.

--

the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.

Photography meets digital computer technology. Photography wins -- most of the time.

blog.kasson.com

 

[JimKasson]

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

This statement must be supported by some examples.
Here is the lens design of 40mm Distagon released in 1994:

Not a large format lens.

--

the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.

Photography meets digital computer technology. Photography wins -- most of the time.

blog.kasson.com

 

[Alexander6x6]

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

This statement must be supported by some examples.
Here is the lens design of 40mm Distagon released in 1994:

I wrote "large format", not "medium format", though Fuji is certainly doing its best to muddy the waters with its recent marketing. You will not find internal focus or floating element designs in lenses designed and marketed for use with 4x5 and larger film.

I just wanted to point out that the "many recent lens designs for digital camera systems" in your statement do not include non-recent 6x6/6x7 medium format systems, which were also far more complex too.

From what I have learned from articles about modern lens design that it tends to be less complex than before. The following example of "recent lens design" may be interpreted as being more complex than it actually is:

 

[Alexander6x6]

Good points all, thanks. Many recent lens designs for digital camera systems are far more complex than even the most modern large-format lens designs, which by their nature are unit focus.

This statement must be supported by some examples.
Here is the lens design of 40mm Distagon released in 1994:

Extinction resolution. Not a metric particularly applicable to real world photography.

I'm wondering if the microfilming photography (which is referred to in the above description) no longer belongs to the realm of real-world photography.