MF Lenses with Shift Adapter or RhinoCam on Full-Frame/35mm Digital Body

[jng]

I'm not saying you're wrong about your Hasselblad 80/2.8 covering 102mm, but I would find that astonishing because that would mean Hasselblad's lens designers decided to make an image circle vastly larger than needed to cover the film. You could cover an 85mm x 85mm piece of film with a 120mm image circle. I don't see why they would do that.

The subject raised here regarding image circles of medium format legacy lenses piqued my interest, as I continue to use some of these to complement my dedicated tech camera lenses. Indeed, some of the V system lenses have image circles that one would consider generous considering the requirements of the 56x56mm film format. According to Wilde's 7th edition of the Hasselblad Manual (the Hasselblad user's bible), one can safely shift the 80/2.8 by 10mm, which by my quick calculation works out to an image circle of 107mm. Similar values are listed for the 60 and 250, with even greater shift (14mm) allowable for the 100/3.5 and 120/4. I've shifted the 100/3.5 pretty close to its limits and the image quality holds up remarkably well out to the edges. Perhaps the designers were not aiming for such large image circles (indeed, lenses like the earlier 40/4 and 50/4 can tolerate minimal if any shift) but rather this is how things turned out as they optimized around other parameters?

John

EDIT: I just saw subsequent posts - Graham and Rob: your calculations are consistent with Wilde.

 

[rdeloe]

The subject raised here regarding image circles of medium format legacy lenses piqued my interest, as I continue to use some of these to complement my dedicated tech camera lenses. Indeed, some of the V system lenses have image circles that one would consider generous considering the requirements of the 56x56mm film format. According to Wilde's 7th edition of the Hasselblad Manual (the Hasselblad user's bible), one can safely shift the 80/2.8 by 10mm, which by my quick calculation works out to an image circle of 107mm. Similar values are listed for the 60 and 250, with even greater shift (14mm) allowable for the 100/3.5 and 120/4. I've shifted the 100/3.5 pretty close to its limits and the image quality holds up remarkably well out to the edges. Perhaps the designers were not aiming for such large image circles (indeed, lenses like the earlier 40/4 and 50/4 can tolerate minimal if any shift) but rather this is how things turned out as they optimized around other parameters?

John

EDIT: I just saw subsequent posts - Graham and Rob: your calculations are consistent with Wilde.

I briefly felt some Hasselblad envy on the part of my humble Pentax lenses... But then I put my Pentax 645 75/2.8 -- the normal lens for Pentax 645 -- on my VX23D to see how large the circle of illumination is. I was able to shift 28mm in landscape on my GFX 50R sensor, at f/16, before hard vignetting occurred. That translates to a circle of illumination of 105mm -- so in the same zone as Hasselblad's 80/2.8 (which I take it is the normal lens for that system?)

In hindsight, this isn't surprising. My 35mm Pentax 645 lens has a much tighter image circle. This is typical for wide angle lenses. The same seems to be true for the 40mm Hasselblad based on the figure in the Flexbody manual that Graham shared. Similarly, the longer focal length Hasselblad lenses have even larger image circles according to that manual, and the same is true empirically for my longer Pentax 645 lenses.

The catch of course is image quality with a shift that large. The edge of the 105mm circle of illumination is not something that Pentax expected anyone to use. Based on one quick test, I was not surprised to see that at the edge of the 105mm image circle on my 75/2.8, image quality was poor, but I was surprised to see that it's not a total mess. I've never shifted this much on my 75/2.8 because I use my Schneider Kreuznach Apo-Digitar 80/4 at this focal length, but it's interesting to see that it does as well as it does out there.

As always when asking lenses to do things they weren't designed for, the proof of the pudding is in the eating. I've had some nice surprises, and a lot of disappointments when trying out different lenses for shift. In the nice surprise department, this is +/- 15mm of shift on my GFX 50R using my new Mamiya G 50mm f/4 at f/11. This is a test shot from a location I use for evaluating infinity performance. The point of focus is a cell tower in the centre of the frame, 4.28km away. This is a near-symmetrical lens designed for a rangefinder camera, so I was expecting strong lens cast. Happily, there's none that I can detect. Light falloff was noticeable in the shifted images. I used the Flat Field Correction tool in Lightroom to tidy that up. A full resolution JPEG version of the frame is here if you're curious.

 

[Audii-Dudii]

View attachment 182282

This level of performance is nothing short of amazing! And even more so when you consider this is an old, film-era lens that's been adapted for use on a camera that's significantly different than the camera for which it was designed.

 

[lowep]

So is the snow!

Now I understand how you can know what you do, given your willingness to test camera lenses out in that kind of weather - unless you are shooting through the window of your office or Porsche speedster?

 

[rdeloe]

So is the snow!

Now I understand how you can know what you do, given your willingness to test camera lenses out in that kind of weather - unless you are shooting through the window of your office or Porsche speedster?

I think it was -12C when I made that picture! Worse, to get to my favourite shooting position for these test frames, I had to jump into some deep snow, which filled my boots. So cold temperatures, cold and wet feet, and freezing fingers -- all in the quest for lens evaluation!

I love the snow though. I'm of the view that there's no such thing as bad weather, just poor clothing choices.

From the forest near where I live, late afternoon on a sunny winter day. I wanted every grain of snow, every tiny piece of material that fell off the trees, to be sharply in focus. That's what I can do with a lens like this and an outfit that allows tilt and shift.

 

[wallpaperviking]

I briefly felt some Hasselblad envy on the part of my humble Pentax lenses... But then I put my Pentax 645 75/2.8 -- the normal lens for Pentax 645 -- on my VX23D to see how large the circle of illumination is. I was able to shift 28mm in landscape on my GFX 50R sensor, at f/16, before hard vignetting occurred. That translates to a circle of illumination of 105mm -- so in the same zone as Hasselblad's 80/2.8 (which I take it is the normal lens for that system?)

In hindsight, this isn't surprising. My 35mm Pentax 645 lens has a much tighter image circle. This is typical for wide angle lenses. The same seems to be true for the 40mm Hasselblad based on the figure in the Flexbody manual that Graham shared. Similarly, the longer focal length Hasselblad lenses have even larger image circles according to that manual, and the same is true empirically for my longer Pentax 645 lenses.

The catch of course is image quality with a shift that large. The edge of the 105mm circle of illumination is not something that Pentax expected anyone to use. Based on one quick test, I was not surprised to see that at the edge of the 105mm image circle on my 75/2.8, image quality was poor, but I was surprised to see that it's not a total mess. I've never shifted this much on my 75/2.8 because I use my Schneider Kreuznach Apo-Digitar 80/4 at this focal length, but it's interesting to see that it does as well as it does out there.

As always when asking lenses to do things they weren't designed for, the proof of the pudding is in the eating. I've had some nice surprises, and a lot of disappointments when trying out different lenses for shift. In the nice surprise department, this is +/- 15mm of shift on my GFX 50R using my new Mamiya G 50mm f/4 at f/11. This is a test shot from a location I use for evaluating infinity performance. The point of focus is a cell tower in the centre of the frame, 4.28km away. This is a near-symmetrical lens designed for a rangefinder camera, so I was expecting strong lens cast. Happily, there's none that I can detect. Light falloff was noticeable in the shifted images. I used the Flat Field Correction tool in Lightroom to tidy that up. A full resolution JPEG version of the frame is here if you're curious.

View attachment 182282

That is awesome, looks great! I know you are using your FrankenCam with such a lens but do you see it being possible to adapt to a smaller footprint "Shift" adapter? I have the Kipon Eos - GFX Shift, do you think you could attach the lens somehow to this adapter? Thanks!

 

[wallpaperviking]

Also, how did you change the aperture? I thought these worked electronically?

 

[rdeloe]

That is awesome, looks great! I know you are using your FrankenCam with such a lens but do you see it being possible to adapt to a smaller footprint "Shift" adapter? I have the Kipon Eos - GFX Shift, do you think you could attach the lens somehow to this adapter? Thanks!

I'm using a stock, unmodified Toyo VX23D -- so no Frankencamera. The Master of Frankencameras is forum member Audii-Dudii, who has done amazing things with bits and pieces of his VX23D.

Let me start with a plan for using Mamiya G lenses directly on a GFX (so a "straight" adapter). No such adapter exists, of course. Therefore, I'm going to build using components that are all available on eBay:

• GFX to M65x1 adapter
• M65x1 focusing helicoid, 17-31mm
• M65x1 to M58x1 adapter ring
• K&F Concept Olympus OM to Canon EOS adapter

I'll screw the K&F Concept OM-EOS mount to the M65 to M58 ring, screw that ring into the helicoid, screw the helicoid into the M65 to GFX adapter, adjust the helicoid to set infinity focus, and then lock it down. Focus would then be by the lens; the only purpose for the helicoid is to make it easy to get the finished piece to the exact length needed for infinity focus. The inside will have to be treated to eliminate stray light that causes glare. I usually use telescope flocking, which I buy in sheets from Edmund Optical. I'll also have to mount a locking pin because the OM lenses had the lens release and lock mechanism in the lens; it's not on my converted lens (no room). All the parts can be purchased on eBay -- for just over $100 USD.

You might be wondering I'm using an OM-EOS adapter. I think I haven't supplied this link before, where I explain how I did the conversion. Here it is: https://www.dpreview.com/forums/thread/4555617#forum-post-64867841 In a nutshell, there are no adapters for Mamiya G lenses, so you have to replace the mount with something else. You can use a threaded adapter, but I wanted a bayonet and had the parts on hand.

OK, now for your question. Tilt-shift adapters might be out. To get the lens on a tilt-shift adapter of some kind, you'd just have to figure out how to attach it using some combination of the kinds of bits and pieces I listed above. As far as I know there is no such thing as an EOS to GFX adapter. The only ones I've seen are for Pentax 645, Mamiya 645, Pentax 67 and Hasselblad. These all have flange distances much longer than the Mamiya G lenses. It might be possible to hack one of these. For example, I've looked at the Mamiya 645 adapter and wondered about removing the Mamiya 645 mount part, machining the rim down, and attaching a replacement mount. I'm not sure the result would be short enough, and I'd worry about the projecting rear part of the Mamiya G 50/4, which sticks out 20.75mm beyond the flange. Its 35.96mm wide at the front. That's probably going to clash with the tilt-shift mechanism.



An EOS-GFX shift adapter is a whole other story. Canon EOS has a big throat, so that rear part might fit through. Will you be able to shift with that thing in the guts of the adapter? You'll have to figure out how to mount it, and you'll have to figure out how to fill the gap between the EOS flange distance (44mm) and the Mamiya G flange distance (which is 56.65mm using the Mamiya 6 mount, and a bit more using my OM mount, which is thinner. That's a simple machining problem if you can't use pre-made parts.

You can figure this all out in advance. Just build something out of cardboard and other materials to simulate the lens with its projecting rear end, and see what happens when it's in your adapter at the proper distance (which will be around 12.65mm in front of the face of the EOS mount).

Note that the shutter is electric on Mamiya G lenses, which is why I had to have it locked open by a skilled technician. However, the aperture is totally manual. Communication with the Mamiya 6 camera happened via levers.

 

[dchew]

I love the snow though. I'm of the view that there's no such thing as bad weather, just poor clothing choices.

Rob,
We should head into the mountains together some day.

Dave

 

[rdeloe]

Rob,
We should head into the mountains together some day.

Dave

I look forward to heading out anywhere one day! Mountains would be lovely. I am eternally grateful that I live near a little forest where I can enjoy a bit of nature, even the tame variety.

 

[ggibson]

This could be of interest--fotodiox introduced a new adapter for MF lenses where the camera is offset from center and rotates around a center point in 4 shots to cover a square within the larger image circle and some overlap between frames. There's no "shift" involved other than the camera rotating into a different part of the image circle. If that's confusing, just take a look at the video for their explanation:

Video demo:

RhinoCam Vertex

fotodioxpro.com

They quote 2.5x area increase in the video, which would be around 46mmx46mm. Clearly they need to make a GFX or Hasselblad X adapter to cover 60mmx60mm since we've established that most of these lenses have some wiggle room

Dual shift adapters can achieve this area and a bit more though, not to mention the ability to capture different aspect ratios. It only requires more careful execution to move and preserve overlap. The rotating "vertex" adapter seems pretty simple and quick to capture 4 successive shots.

 

[rdeloe]

This could be of interest--fotodiox introduced a new adapter for MF lenses where the camera is offset from center and rotates around a center point in 4 shots to cover a square within the larger image circle and some overlap between frames. There's no "shift" involved other than the camera rotating into a different part of the image circle. If that's confusing, just take a look at the video for their explanation:

Video demo:

RhinoCam Vertex

fotodioxpro.com

They quote 2.5x area increase in the video, which would be around 46mmx46mm. Clearly they need to make a GFX or Hasselblad X adapter to cover 60mmx60mm since we've established that most of these lenses have some wiggle room

Dual shift adapters can achieve this area and a bit more though, not to mention the ability to capture different aspect ratios. It only requires more careful execution to move and preserve overlap. The rotating "vertex" adapter seems pretty simple and quick to capture 4 successive shots.

Interesting new tool for those who need a big square they can use or crop. In the video the tripod wiggled a lot during movements, so a sturdy one will help keep edges straight.