APO SILVETAR f5,6 / 35mm

[diggles]

Just received an email from Linhof Studio promoting this lens. The email stated "...designed and manufactured by Silvestri Cameras..."

The brochure on the Silvestri website:

http://www.silvestricamera.com/assets/files/Silvetar_en-datasheet.pdf

This is a quite interesting lens in terms of focal length and image circle and compatibility. Now we have 3 options for the 35mm focal length/actus/gfx combo. It's also great to see a company manufacturing new lenses for tech cameras.

I'm not that versed in how to read mtf charts, so I don't have a sense for how good it is in terms of sharpness and distortion.

Does anyone here have any experience with this lens? Or other thoughts?

 

[buildbot]

I’m also not very good at reading MTF charts, but it seems decent enough, though perhaps relatively a lot of falloff at 40lpmm.

It’s a retro focal design though, so wouldn’t distortion affect this lens a lot?

 

[Steve Hendrix]

Not much shift latitude. Hard to say if it would improve over a Schneider Digitar or perhaps even a Rodenstock 35 HR-S.


Steve Hendrix/CI

 

[rdeloe]

That actually looks quite good! The MTF chart shows a lot of promise. On big shifts you'll lose fine detail at the edges of the image circle, but contrast is good almost all the way out. The image circle is decent (a bit bigger than 645).

My main concerns as a GFX user is that there's a fair bit of lens sticking out past the flange. But it looks workable. Here's the deal:

• The lens data sheet doesn't give a number, but from measuring off the screen I'd say 38.2mm.
• That's longer than my Mamiya N 43mm f/4.5 L, but the Mamiya has a shorter flange distance. So the APO-Silvestri may be outside, but could also be inside the Rotafoot on my F-Universalis + GFX outfit.
• The opening through the Rotafoot is around 62mm, and the rear of the lens is 23mm (could be a bit more but hard to tell from lens data), so that means 19-20mm of shift before hitting the Rotafoot if it's inside. If it's not inside, no limits.
• So... long story short, you can't shift 20mm on GFX using a lens with a 75mm image circle anyway (max is around 12mm). It should work whether it's inside the Rotafoot or not.

Nice to have new options.

Now the big mystery: what's this based on, because can Silvestri really design and make its own lenses??

 

[diggles]

Not much shift latitude. Hard to say if it would improve over a Schneider Digitar or perhaps even a Rodenstock 35 HR-S.


Steve Hendrix/CI

It would be quite remarkable if it improves on the 35-hr in terms of sharpness or the digitar in terms of distortion.

 

[diggles]

It’s a retro focal design though, so wouldn’t distortion affect this lens a lot?

That's my worry, but Laowa does a pretty good job of managing distortion with the 20mm zero-d. Not zero distortion, but pretty good really.

I'm cautiously optimistic on the Silvetar...a little worried about it though since distortion was not specifically addressed in the data sheet.

 

[rdeloe]

That's my worry, but Laowa does a pretty good job of managing distortion with the 20mm zero-d. Not zero distortion, but pretty good really.

I'm cautiously optimistic on the Silvetar...a little worried about it though since distortion was not specifically addressed in the data sheet.

I just mentioned to someone else that I wanted to see a distortion chart! That is indeed a worry. If it suffers from strong distortion, it won't be much use for architecture. Distortion performance does seem like something that ought to have been mentioned.

 

[kinglang]

I don't know the source of this 35mm lens, but the image field of only 75.6mm is too small even for a 44X33 sensor. 35mm on 44x33 sensor is about 27mm view,
In addition,lack the distance data D from the rear component to the CMOS. On gfx, it may also protrude into the bayonet.

 

[xavieryoung]

I'm very weary of these rebadged lenses after my experience with the Cambo 35mm "Actar". This involved finding a lovely Contax Distagon and then paying Cambo to transfer the optics into a dumb tube and manual aperture. What they don't tell you is that the floating element corrections are lost too. You end up with a lens that works fine at infinity but starts to fall apart very badly at anything closer than about 2m. As far as I can see they've done the same with the 19mm Nikon / Actar. Cambo's argument seems to be that only architectural / landscape photographers use wide angle lenses and they only shoot at infinity - so my bad.
As a retrofocus design I assume this is based on something with a more complex focusing system than just moving the whole lens back and forth. I know it would make these lenses even more eye wateringly expensive but couldn't a simple correction setting be included like the older Hasselblad Distagons?

 

[dchew]

I don't know the source of this 35mm lens, but the image field of only 75.6mm is too small even for a 44X33 sensor. 35mm on 44x33 sensor is about 27mm view,
In addition,lack the distance data D from the rear component to the CMOS. On gfx, it may also protrude into the bayonet.

It does say, “Nearest distance from sensor plane = 37.5 mm, but that leaves only 5.8 mm for that rear element. 43.3 - 37.5 = 5.8

Does not seem proportional to the photo. Maybe I am misinterpreting what they mean.

Dave

 

[4x5Australian]

It's heartening to see the entry of a new 35mm focal length lens having an oversized image circle for use on technical and medium format cameras.

Silvestri seems to be responding to a gap in the current offerings for a compact, lightweight and reasonably 'affordable' 35mm focal length lens of high resolving ability for use with 50mpx and 100mpx sensor mirrorless cameras and digital backs on the new generation of popular lightweight adjustable cameras such as the Cambo Actus.

According to the Linhof & Studio website, Rodenstock has discontinued its HR Digaron-S 35mm f/4, and the Schneider Apo-Digitar 35XL f/5.6 was discontinued in 2016. Those cancellations leave the HD Pentax 645-D FA 35mm f/3.5 AL(IF) as, I think, the only other current 35mm lens with both a mechanical aperture mechanism and oversized image circle.

The Apo-Silvetar 35mm is compact in size. Its diameter is the same as the petite Apo-Digitar 35XL and its length is only 4.2mm longer.

The stated weight of 200g places it in the truly lightweight class. That weight is surprising and a bit hard to believe given the thickness of the lens elements. The Apo-Digitar 35XL weighs 240g, and the elements of this lens look chunkier. The HR Digaron-S 35mm weighs 480g.

Silvestri states that the image circle is 75.6mm in diameter at a working aperture of f/22 and that it allows a 44x33mm sensor to be displaced up to 11mm lengthwise or 14mm widthwise. That amount of shift is certainly useful for many photographers and is comparable to the 15mm of shift provided by shift adapters for medium format lenses from Kipon and Fotodiox.

However, we generally observe that the resolution of lenses at f/22 is degraded by diffraction, mandating use of wider apertures of f/8 to f/11. At those wider apertures, the image circle is likely to be somewhat smaller.

The single MTF graph presented suggests that the lens resolves very well indeed. The centre compares well to the HR Digaron-S 35mm and the outer edge appears to be at least equal to the Apo-Digitar 35XL at the equivalent point (at about 80% of the 35XL's 45mm field radius).

At what distance were the MTF curves calculated? We have to guess infinity. At which aperture? The aperture is not stated either. However, the maximum radius of 47 degrees shown on the horizontal axis is half the stated field of view (diameter) of 94 degrees. On the previous page it's unstated but implied that the field of view is also given at the working aperture of f/22.

Is the resolution of this lens really that good at f/22? Is it even better at, say, f/11? How much smaller is the image circle at f/11?

Being of retrofocus design, we anticipate distortion. The absence of a distortion graph, therefore, is concerning. The second sample photograph contains vertical lines showing some barrel distortion, so it's there to some degree. The amount seems much less than the really conspicuous barrel distortion I recall seeing in test images from the HR Digaron-S 35mm.

For me, the real test is horizontal lines and elements across the frame, and indeed the tendency of an entire vertical facade to bow outwards when photographed in a front elevation view (i.e., at 90 degrees to the facade). The oblique viewpoints shown in these two first samples are unhelpful in that regard.

I think many photographers will find this new lens attractive, useful and affordable. Bravo to Silvestri.

 

[rdeloe]

I don't know the source of this 35mm lens, but the image field of only 75.6mm is too small even for a 44X33 sensor. 35mm on 44x33 sensor is about 27mm view,
In addition,lack the distance data D from the rear component to the CMOS. On gfx, it may also protrude into the bayonet.

I'm not sure what you mean here. Do you not like the angle of view on GFX, or do you not like the size of the image circle?

The former is personal taste of course. I really like 35mm on GFX and use it a lot. I'd like a larger image circle, but it's larger than the one on my Pentax-A 645 35mm, which does give me a very useful amount of movement.

For the money, this new lens would have to be spectacular to replace my Pentax. Now if it was 30mm instead of 35mm, I would have one on the way already...

 

[rdeloe]

I'm very weary of these rebadged lenses after my experience with the Cambo 35mm "Actar". This involved finding a lovely Contax Distagon and then paying Cambo to transfer the optics into a dumb tube and manual aperture. What they don't tell you is that the floating element corrections are lost too. You end up with a lens that works fine at infinity but starts to fall apart very badly at anything closer than about 2m. As far as I can see they've done the same with the 19mm Nikon / Actar. Cambo's argument seems to be that only architectural / landscape photographers use wide angle lenses and they only shoot at infinity - so my bad.
As a retrofocus design I assume this is based on something with a more complex focusing system than just moving the whole lens back and forth. I know it would make these lenses even more eye wateringly expensive but couldn't a simple correction setting be included like the older Hasselblad Distagons?

This has always bothered me -- especially the fact that they're not very up front about the fact that they're not using the floating elements when they rehouse these lenses. I never shoot at infinity!

 

[diggles]

Linhof Studio replied to my inquiry email and told me that this is an entirely new lens design.

 

[rdeloe]

Linhof Studio replied to my inquiry email and told me that this is an entirely new lens design.

Well congratulations to Silvestri then! Thanks for checking.

 

[buildbot]

Linhof Studio replied to my inquiry email and told me that this is an entirely new lens design.

Very interesting. I wonder if it’s becoming more feasible to do this kind of work as a small company or if they contracted the design to nitoh/qioptic/someone else. Very high precision machining is fairly easy to come by these days, and if you have money, optical design tools.

 

[kinglang]

I'm not sure what you mean here. Do you not like the angle of view on GFX, or do you not like the size of the image circle?

The former is personal taste of course. I really like 35mm on GFX and use it a lot. I'd like a larger image circle, but it's larger than the one on my Pentax-A 645 35mm, which does give me a very useful amount of movement.

For the money, this new lens would have to be spectacular to replace my Pentax. Now if it was 30mm instead of 35mm, I would have one on the way already...

I prefer 30mm on gfx(24mm angle of view on 35mm formart) ,image circle slightly smaller. Still makes sense for 44x33 and 36x24 sensors.

 

[rdeloe]

I sent a note to Silvestri with some questions I'm hoping they can answer. I told them I'd post the results here. Let's see what comes up.

1. Is the weight figure of 200grams correct? That would be excellent for my use, but it seems very low.
2. How many blades does the aperture have, and how are they shaped? Do you have a picture?
3. For your MTF chart on page 3,
   1. The X-axis would be 37.8mm at infinity, but yours goes to 50mm. What is the subject distance from this plot?
   2. What aperture was assumed in these calculations?
4. All of my architecture colleagues are wondering what the distortion behaviour of this lens is! Can you provide a graph?
5. We’re also interested in the light falloff behaviour of the lens at f/5.6, and at f/11. Do you have a graph you can share?

 

[rdeloe]

I sent a note to Silvestri with some questions I'm hoping they can answer. I told them I'd post the results here. Let's see what comes up.

1. Is the weight figure of 200grams correct? That would be excellent for my use, but it seems very low.
2. How many blades does the aperture have, and how are they shaped? Do you have a picture?
3. For your MTF chart on page 3,
   1. The X-axis would be 37.8mm at infinity, but yours goes to 50mm. What is the subject distance from this plot?
   2. What aperture was assumed in these calculations?
4. All of my architecture colleagues are wondering what the distortion behaviour of this lens is! Can you provide a graph?
5. We’re also interested in the light falloff behaviour of the lens at f/5.6, and at f/11. Do you have a graph you can share?

** Correction: Rod pointed out that the X axis is in degrees, not mm. I missed that. I've never actually seen that before (and should have read Rod's post more carefully!)

 

[Alan]

Interesting for both 24x36 and 33x44 sensors. As everyone else has said, need some more technical info as well as sample shots with some shift. There's a bit of barrel distortion showing on the right edge of that architecture shot. Question is does it get more complex farther out in the image circle?