The Fujinon GF Tilt Shift Lenses Are Finally Here!

[tcdeveau]

Idk the way Fujifilm marching forward withe their system is making me regret the X system a bit.

In what way? Just curious. My wife has an X-T4 we’re still getting good use out of. I had looked into expanding the X system at some point with an X-H2s and more lenses but am thinking about the GFX100II instead.

 

[Ai_Print]

I don't think Ai_Print is ready yet to tilt towards Fuji, but it's becoming more likely.

Actually in practical terms, it would still be a ways off. My 100mm CFI produces fantastic imagery using a tilt-shift adapter and the 30mm would be wider than I normally use. A 55-65mm range T/S lens would have gotten more of my attention.

 

[rdeloe]

Actually in practical terms, it would still be a ways off. My 100mm CFI produces fantastic imagery using a tilt-shift adapter and the 30mm would be wider than I normally use. A 55-65mm range T/S lens would have gotten more of my attention.

For some people 30mm on 33mm x 44mm is not wide, but I'm like you: it's wide for me. I rarely use my 28mm, but I use my 35mm a lot. It always struck me as bit strange that way back when several companies made 35mm shift or tilt-shift lenses for 35mm film (e.g., Canon, Nikon, Minolta, Contax), but then when Canon and Nikon kept going with modern tilt-shift lenses, they ignored 35mm completely. Chinese companies like Laowa are doing the same thing. There are lots of ultra-wide options, but nothing in that moderate wide space that I prefer.

As good as the Fuji 30mm t-s lens looks to be, I'm highly doubtful I'll get one because I really want something excellent at around 32mm that I can use on my F-Universalis with GFX 100S. I suspect there will never be such a lens.

 

[Alan]

It always struck me as bit strange that way back when several companies made 35mm shift or tilt-shift lenses for 35mm film (e.g., Canon, Nikon, Minolta, Contax), but then when Canon and Nikon kept going with modern tilt-shift lenses, they ignored 35mm completely. Chinese companies like Laowa are doing the same thing. There are lots of ultra-wide options, but nothing in that moderate wide space that I prefer.

I was just talking with an architect while scouting today about how far away to shoot overall building exteriors. I showed him how 35mm (on 135 format) is my default "neutral" feeling focal length that doesn't make the building feel too flat or too sharp. (This would be around 45mm on a GFX, 50-55mm on IQ4.) He expressed annoyance with photographers who always push super wide exaggerated shots with the corner of the building projecting up in an acute angle.

Of course, that's not always possible to use due to other buildings across the street, too tall of a subject building, etc. And last week, a different architect was totally OK with the super-wide look. ¯\_(ツ)_/¯

 

[4x5Australian]

He expressed annoyance with photographers who always push super wide exaggerated shots with the corner of the building projecting up in an acute angle.

Of course, that's not always possible to use due to other buildings across the street, too tall of a subject building, etc.

The corner projecting up at an acute angle has been called 'ship's prow effect' or 'ship's prow distortion'. Yes, it looks ridiculous.

Other than backing off and using a longer focal length, which (as you say) is not always possible, the traditional way to overcome it is to combine a large swing with a large lateral shift sufficient to re-centre the building.

That combination is impossible with a tilt-shift lens that allows shift along only one axis, because the shift is already utilised for front rise/rear fall to get the top of the building into the frame.

It's one of the advantages of using a technical camera that allows simultaneous XY shift on two axes and lenses with oversized image circles.

 

[marc aurel]

The corner projecting up at an acute angle has been called 'ship's prow effect' or 'ship's prow distortion'. Yes, it looks ridiculous.

Other than backing off and using a longer focal length, which (as you say) is not always possible, the traditional way to overcome it is to combine a large swing with a large lateral shift sufficient to re-centre the building.

That combination is impossible with a tilt-shift lens that allows shift along only one axis, because the shift is already utilised for front rise/rear fall to get the top of the building into the frame.

It's one of the advantages of using a technical camera that allows simultaneous XY shift on two axes and lenses with oversized image circles.

Good description of what helps to reduce that effect. I use that too. But I would not agree that tech cam is always in advantage. The GF 30mm TS shift element will allow 15mm shift. If you rotate that by 45°, you have 10,6mm rise and 10,6 shift. If you compare that e.g. with a Digaron-W 32mm on a 54 x 40mm sensor, you can apply 8mm of rise and 8mm of shift before you hit the edge of the image circle. At least according to the excellent visualizer tool by digital transitions (https://dtcommercialphoto.com/visualizer-ic-2019/visualizer-ic-2019.html). Of course you could use just a crop of that sensor and have more relative shift. But at the wider angles we were talking about, I don't see the tech cam in an advantage. The tilt-shift lenses just organize rise/fall and shift in a different way.

 

[guphotography]

Given the similar image circle, I wonder how this lens would fare on tech cam with appropriate adapters, a worth alternative to 32 hr maybe? At least without the dreaded distortions I hope!

 

[drevil]

Given the similar image circle, I wonder how this lens would fare on tech cam with appropriate adapters, a worth alternative to 32 hr maybe? At least without the dreaded distortions I hope!

will be interesting to see if cambo does another total conversion like they did with the nikon 19mm

 

[Alan]

The GF 30mm TS shift element will allow 15mm shift. If you rotate that by 45°, you have 10,6mm rise and 10,6 shift.

Yep - just have to change your thinking from XY to Angle-Distance. Works great, even for stitching. You don’t even have to limit yourself to the click-detent angles.

There was a company (Gottschalk?) making a shift camera with that radial strategy several years ago.

 

[lookbook]

Good description of what helps to reduce that effect. I use that too. But I would not agree that tech cam is always in advantage. The GF 30mm TS shift element will allow 15mm shift. If you rotate that by 45°, you have 10,6mm rise and 10,6 shift. If you compare that e.g. with a Digaron-W 32mm on a 54 x 40mm sensor, you can apply 8mm of rise and 8mm of shift before you hit the edge of the image circle. At least according to the excellent visualizer tool by digital transitions (https://dtcommercialphoto.com/visualizer-ic-2019/visualizer-ic-2019.html). Of course you could use just a crop of that sensor and have more relative shift. But at the wider angles we were talking about, I don't see the tech cam in an advantage. The tilt-shift lenses just organize rise/fall and shift in a different way.

... i don't have enough knowledge, but isn't this a comparison between a small medium format sensor and a medium format sensor??
what is the effect of the fact that the image space is "slanted" - the usable area for a straight photo would be much smaller, wouldn't it?

 

[marc aurel]

Yes, it compares two different sensor sizes. But if you want a specific angle of view, you can achieve it with different formats. The concept of equivalence. I'm sure others here can explain it better, but here's a dpreview article about it: https://www.dpreview.com/articles/2666934640/what-is-equivalence-and-why-should-i-care
I'm not sure what you mean by "slanted". Maybe the fact the 3-dimensional objects (like a sphere) look distorted the wider the lens and the further to the edge, the more? That effect is independant of the format you take your image with. Just depends on the angle of view. If you want to use such extreme angles of view, you have to compose your image carefully and avoid such objects far from the image center.

To compare the shift range that's possible between the GF 30mm TS and the Digaron-W 32mm on the same sensor size: you could use an IQ250. Attached you see possible shifts with that sensor. It's a bit more, about 12mm with the Digaron-W 32 vs. 10,6mm with the GF 30mm TS. On the other hand the Fuji lens is a bit wider. Then the distortion correction for the Digaron will eat some more pixels. So results should be pretty close. (those who use this lens – please correct me, if the lens should have a larger image circle than what it's specifications say).

Of course digital backs with a 44 x 33mm sensor are all just 50MP. So if you want the higher resolution, this is a not an option. You could of course crop the IQ4 to 44x33mm and would have 100MP. That would be a direct comparison.

 

[Paul Spinnler]

In practice the 32 gives more shift than 8mm on a full gram sensor …

 

[marc aurel]

In practice the 32 gives more shift than 8mm on a full gram sensor …

Yes. According to the visualizer tool 16mm rise/fall are possible, and 13mm lateral shift. What I was talking about was combinig rise and shift at the same time. Sorry I did not make that clear enough.
I hope that sounds more realistic

 

[Alkibiades]

Digaron-w 32 mm will allow you more than 20 mm movements on the IQ250 and other 33x44 sensors. even Digaron-S 35 mm will allow you more then 15 mm movements.

 

[marc aurel]

Digaron-w 32 mm will allow you more than 20 mm movements on the IQ250 and other 33x44 sensors. even Digaron-S 35 mm will allow you more then 15 mm movements.

That sounds great and is more than the 15mm that the shift mechanism of the Fuji TS lens allows. But the numbers I cited were for a 54x40mm sensor.
It's harder than I thought to be precise about what I talk about....

Does anyone have real life numbers for the 54x40mm sensor?

 

[Alkibiades]

That sounds great and is more than the 15mm that the shift mechanism of the Fuji TS lens allows. But the numbers I cited were for a 54x40mm sensor.
It's harder than I thought to be precise about what I talk about....

I use the 32 HR also with the large sensor (80 MP) that have the same size than the IQ4 150 MP and even 17 mm shift still work fine. These are my experiences in the real work, also with 40 HR, 50 HR- they have the same image circle. the 23 HR can be used about 5 mm on the large sensor, but still more than 10 mm on the 33x44 mm.

 

[rdeloe]

will be interesting to see if cambo does another total conversion like they did with the nikon 19mm

Do you have any experience with the Actar 19? I have to think that lens uses a floating element design. If you take the cells out of the native housing and put them in your own housing without a focusing helicoid, you're ignoring the problem the floating element design solves, which is close focus performance in edges and corners.

I had my Pentax-A 645 35/3.5 apart yesterday and it uses a floating element design. There are only two lens groups that screw into either side of the aperture mount inside the lens. It's a floating element design, but what that means is the two lens parts do not move in a linear fashion at close distances, like they would in a simple unit focusing design. If Cambo is simply putting the front and back parts of the lens in a mount, with focus provided by the Actus, then it's a unit focusing design that is not correcting for close distances anymore.

Anyway, long story short, I've tested how my Pentax-A 645 35/3.5 would do if someone did that, and not surprisingly close focus performance is terrible. If that's what Cambo is doing with the Actar 19, 24 and 35 (all of which are conversion lenses), then I have questions...

 

[Paul Spinnler]

The Fuji weighs almost twice as much, is larger, has less IC coverage, and costs half or less than half than the Digaron W. You cannot use the Fuji on a tech cam given how it works with the shift built-in in the lens.

On the other hand, jointly with the Fuji camera, it is an LCC-free workhorse perfectly suited for production-oriented architectural photography where I can see it having high appeal given price, lack of colour cast and auto-correction metadata. Like go in, make the shot, run C1 preset over it, bill. A good, reliable tool.

Another point - can this Fuji lens do shift on X/Y at the same time? Ie can you do 4,6 shot stitches? The Digarons can be shifted on all tech cam systems on both X AND Y axes at the same time, which is also an added benefit.

This said, I'll keep my 32 HR despite the horrible distortion (which you need to manually correct in C1)

 

[marc aurel]

The Fuji weighs almost twice as much, is larger, has less IC coverage, and costs half or less than half than the Digaron W. You cannot use the Fuji on a tech cam given how it works with the shift built-in in the lens.

On the other hand, jointly with the Fuji camera, it is an LCC-free workhorse perfectly suited for production-oriented architectural photography where I can see it having high appeal given price, lack of colour cast and auto-correction metadata. Like go in, make the shot, run C1 preset over it, bill. A good, reliable tool.

Another point - can this Fuji lens do shift on X/Y at the same time? Ie can you do 4,6 shot stitches? The Digarons can be shifted on all tech cam systems on both X AND Y axes at the same time, which is also an added benefit.

This said, I'll keep my 32 HR despite the horrible distortion (which you need to manually correct in C1)

Paul, if you already have a 32 HR, I can not imagine any reason to sell it ;-)
Stitching is possible with the Fuji lens. You can attach a lens collar for parallax-free stitching.

Stitching 2 images

• That will be simple and will give you a 63x44mm image with about 194MP (I had to edit this part, wrong numbers, sorry. 33mm+2x15mm=63mm, close to twice the height)

Stitching 4 images will be inconvenient but possible:

• Shift diagonally, take one image at +15mm, one at -15mm.
• Change the rotation of the shift unit and take 2 more shots.
• Since you can rotate the lens in 2 independent planes (one between camera and shift unit, one between shift unit and tilt unit) you can keep the camera level while shifting diagonally. Practically the shift unit can be rotated independently while camera and the front part of the lens are kept level. I attach a product photo where everything but the shift unit is greyed out.
• If you choose 45° and stitch the 4 images that should result in a final image of about 65 x 54 mm with 245MP
• That will not nearly be as comfortable as with a tech cam with independent movements in x and y directions, but i'm sure it will work.

 

[P. Chong]

I might have missed it, but will the 110 work with extension tubes, and what magnification will it achieve with either or noth Fuji tubes.?