short test: Fuji 100MP with Zeiss 35 mm or Phase one 50 MP with Digaron HR 35 mm?

[Alkibiades]

Old question, what is more important, more pixel or a better lens?
Afcouse better lens and more pixel will be the best option, but ...

I made now a short comparison of a good lens on 100 MP and an excellent lens on 50 MP back. Afcourse it would be the best methode to use the same lens on both systems, but this is the practical compatibility problem.

Sadly I can use the one lens only on the one system and the other lens on the second.

The first one is the Fuji GFX 100S with the most used wide angle focal length 35 mm.

And the lens that I need must be a shift lens that allows me movements, standart wide lens is useless for me.

As long as Fuji cant deliver the new 30 mm Shift lens I use an old but still very good and very known Zeiss Distagon PC 35 mm.

It was made for Contax, one version as a shift lens for Contax 35 mm and one version as a Wide angle with no movements for Contax 645. Both share the same lensdesign accept one additional lens element conditioned by the different camera mount.

The lens is still one of best existing 35 mm for 645 Format, even it is an old Film lens.

I use the lens on Fuji 100S becouse there is no other option for me and this lens close the gap between canon 24mm TS-E II and 50 mm TS-E.

This combination allows me a very fast and spontaneous work with great mobility.

17 TS-E and 15 Laowa can help at special use, but is mostly too wide.


Zeiss Distagon was a great lens on my Sony FF sensors and allows me still on Fuji100S full 10 mm movements in almost good quality.

But I wanted to know if the 100 MP combo is better than a 50 MP back with the best 35 mm available.
I can use my Phase one 250 with Digaron-W 4/32 mm HR and Digaron-S HR 4/35 mm and both lenses are simply the best lenses in this range. The sharpness, resolution and performance at the edges are simply on other level.
With the smaller 33x44 mm back I even prefer the Digaron-S 35 mm as the lens is much smaller than the 32HR.

So I take a pictures with both systems from the same position, both shifted 10 mm, no correction, no LCC, no sharpening, both files develope in capture one.

 

[Alkibiades]

Fuji100 s and Distagon in practical use:

 

[Alkibiades]

here the test:

 

[Paul Spinnler]

Thanks Albibiades - goes to show how remarkable still tech cam lenses are - they are cleary of higher quality than system lenses.

It seems though that most people are being priced out to play with the best stuff, seeing how much the list prices are nowadays for the most sought after of these lenses.

 

[Alkibiades]

when watched at 100% both files are sharp, the details seems to be the same. but the phase one/ Digaron HR35 mm is clear sharper.
But when scaled to 400% the differance is clear:
- Phase one / Digaron 35 mm ist even at 400% clear razor sharp, it capture more details and much sharper and clear. There are no chromatic aberrations.
- Fuji 100S/Zeiss 35 mm have clear much chromatic aberrations ( you could remove it with C1 but these parts will looks soft anyway), the details are soft ( when scaled-at 100 % looks fine),
the perfomance at the courners are much worst than the center, where there is no degradation on Digaron 35 mm.

 

[Alkibiades]

conclusion:
When I would make big pints there is no question that the 50MP back with a exellent Digaron lens gives me clear a better and sharper file that can be scaled in great quality, then the 100MP with a good but not exellent lens. The difference is clear to see on big prints, it is not silly pixel picking.
Here the lens is clear the main factor, the question if 50 MP or 100 MP is really not important. Also the logarithms to scale up the files becomes better and better...
The Fuji 100MP sensor is back illuminated, that would be a great sensor for Digital back not becouse of more pixes but the possibility to use Schneider wide lenses with less color cast and full movements. Sadly the Fuji body dont allow to use these great lenses, so we have to wait for the back version, maybe from hasselblad?...
Please dont understand me wrong: I love Zeiss Distagon PC lens, I used it with Canon bodies, Sony E bodies and now with the Fuji. This is not bad for a 30 years old lens.

 

[nazdravanul]

Wow, striking results! Great test, thank you!

 

[MGrayson]

There’s a reason DxO never included tech cam lenses in their charts. Everything else would have looked pitiful.

 

[rdeloe]

Did you shift the Contax PC-Distagon 10mm on GFX? According to Zeiss, it allows 10mm of shift on 24mm x 36mm, which means it has a usable image circle of 61mm. A 10mm shift on GFX requires a 72mm circle of good definition. I would expect the PC-Distagon to be terrible that far outside of its designed range, especially as image quality has already crashed at the edge of the part of the image circle that is meant to be used. Your Rodenstock 35mm at least has a 70mm image circle.

 

[ErikKaffehr]

Thanks Albibiades - goes to show how remarkable still tech cam lenses are - they are cleary of higher quality than system lenses.

It seems though that most people are being priced out to play with the best stuff, seeing how much the list prices are nowadays for the most sought after of these lenses.

Did you shift the Contax PC-Distagon 10mm on GFX? According to Zeiss, it allows 10mm of shift on 24mm x 36mm, which means it has a usable image circle of 61mm. A 10mm shift on GFX requires a 72mm circle of good definition. I would expect the PC-Distagon to be terrible that far outside of its designed range, especially as image quality has already crashed at the edge of the part of the image circle that is meant to be used. Your Rodenstock 35mm at least has a 70mm image circle.
View attachment 201998

Quite good up like 25 mm off axis at f/5.6. But I would think that modern designs could be quite a bit better.


Old Hasselblad Distagon 4/40 CFE


New Hasselblad Distagon 4/40 CFE IF. The MTF curves are much higher on the new lens, but 'tan' drops quite a lot. That is often a sign of lateral chromatic aberration.

The MTF data for the XCD 2.5/38V is calculated, not measured, but it is quite impressive.

The Zeiss Loxia 2.4/25 doesn't cover 33x44 mm, but looks pretty good at f/5.6. These are probably measured data.

Best regards
Erik

 

[sunhouwang]

I also have these two sets, GFX100s+PC Distagon 35mm YC mount, Credo 50+Alpagon 35mm, and did the same comparison test last summer.
Fat pixel of Credo 50mm set has sharpest performance from center to corner.
While 100s set has high resolution, similar sharper center performance, a little soft corner, good news is you could get full movement on the 100s body.

 

[ErikKaffehr]

Those technical sheets for the XCD 2.5/38mm and Loxia 2.4/25mm lenses rather conspicuously omit a graph for distortion. I bet they have distortion levels higher than we'd like to see.

Up to and including the CFE Distagon T* 4/40 FLE lens of 2000-2003, Zeiss had boasted in its technical sheets for the retrofocus Distagon lenses that the level of distortion was "well controlled". For example, in the technical sheet for the CFE Distagon 4/40mm FLE, Zeiss wrote: "Distortion is particularly well controlled. So the Distagon®T* 4/40 CFE lens can handle demanding architecture, product shots, and industrial jobs with fully professional results". The distortion graph for the CFE 40mm FLE lens shows the distortion peaking at around 1.4%. As a reference, symmetrical wide-angle lenses typically have distortion levels of 0.5% or less.

However, when Zeiss introduced the CFE 40mm IF FLE lens in 2003 to provide increased resolution for digital sensors, the distortion level went up to 3.5%. In the technical sheet, Zeiss dropped any discussion of distortion, boasting only of the increased resolution at wide apertures and its application for digital sensors. The vertical scale of the distortion graph was halved in scale, presumably to camouflage the increase. That graph and the much higher distortion it revealed attracted a lot of negative comment.

With the push on to increase the resolution of retrofocus designs for digital sensors, it appears that the level of distortion increased to a new, much higher level than was previously acceptable.

The distortion might be addressed, even eliminated, by image processing software, but obviously doing so involves additional steps, especially if the lens is mounted on some other camera or a digital back and shifted.

Zeiss publishes distortion data for the Loxia 2.4/25:

Source: https://www.zeiss.com/content/dam/c...n/loxia-lenses/datasheet-zeiss-loxia-2425.pdf

I would guess that the trend is a bit towards more symmetric designs with mirrorless, as that would yield a more compact build. Vignetting and optical crosstalk is reduced with BSE and there is no need for mirror box.

Best regards
Erik

 

[4x5Australian]

Erik, yes, it is clear that I was wrong to conclude that the two newer lenses you mentioned plus others would have higher levels of distortion.

Another member contacted me and pointed me to the technical data sheets on the Zeiss website.

The newer Zeiss lenses have impressive resolution, yet their distortion levels are lower than the old Distagon designs.

Your suggestion that these newer lenses might trend towards more symmetric designs, given that the cameras are mirrorless, is interesting.

Rod

 

[B L]

here the test:

Excellent. Thank you. But I noticed cyan/purple cast, should we consider that as a defect?

 

[Alkibiades]

Excellent. Thank you. But I noticed cyan/purple cast, should we consider that as a defect?

no, as I wrote the zeiss 35 mm lens has apochromatic aberretions, this is the cyan/purple cast. At 100 % the cast is not so clear as here with magnification of 400%.
Only such Apo lenses like here tested Digaron-S HR 35 mm have no cast.