7Artisans 50/1.1, or how Vignetting does not matter in the Real World

Sonnar Brian

Sonnar Brian

Product of the Fifties
Staff member
Local time
5:00 AM
Joined
Jan 12, 2004
Messages
19,780
View My Gallery
I've been looking at Vignetting in more detail- went as far as to put together a light box to put out some diffuse white light. I wrote some code to process the DNG files from my M9 and M Monochrom.

In other words, "use/lose Leave", cold outside, and the museums and other places to visit are pretty much closed down.

First up for testing with the new light box - my pair of 7Artisans 50mm F1.1 lenses. I have one optimized for the M9 and a second for the M Monochrom. The first one is just over SN 200 in the series. The second, just over 1200. This lens is a modern incarnation of the Zunow 5cm F1.1, and cost as much as that lens did over 60 years ago: uncorrected for inflation.

Vignetting is worst at infinity, and that is how tests are normally done. Lens coding is turned off. I normally code this as a Leica 50/1.0. The original Noctilux is known for vignetting in the corners.

I used the M Monochrom for these tests. My code processes both M9 DNG files and M Monochrom DNG files. I use an 8x8 average, so 4 Bayer cells of the M9 gets averaged out. A quick test shows the M9 and M Monochrom results were very close with this lens.

INF1DIAG.jpg


INF2DIAG.jpg


At F1.1- farthest corners show ~3.4F-Stop drop off. Things improve up to F2, but still a 1.5F-Stop drop.

Asymmetry in the curves: I looked closely at the shape of the aperture as it stops down. It is not round, and not quite symmetrical. I think that is the reason. My home-made light box is not completely uniform, but much flatter than just holding a sheet of paper or opaque plastic over the lens.

SO- with results shown above, you would think all shots wide-open would look like someone was using a flashlight to illuminate the image.

7Artisans lens #1, wide-open and close-up.



Lens Number 2, wide-open,



 
Lens # 2, wide-open,



F1.5,



At F4,



This is a Sonnar formula lens, stretched to Ludicrous Speed. And sells for less than many of the J-3's on Ebay.
 
Wide-Open, some uniform background:



At F1.5,



I had the C-Sonnar with me for comparisons, so used F1.5- not F1.4.

The "real-World" images are coded as the 50/1.0 Noctilux, and a curve is applied to the values stored in the DNG file. The corrections are not huge at the corners, boost the intensity about a third.
 
These photos largely make your point.

The only one that fails for me is the outdoor shot of two buildings. The two left corners look unnaturally dark. Other images are good. Thanks for posting this. I'll have to use mine more.
 


(The Blue X-Axis, caused by the "HPGL" emulator I'm using.)

This is the plot for lens #2 focused to 0.7m. Still more than a 2 F-Stop falloff at the far corners at F1.1, and an F-Stop when stopped down. The corrections applied by the camera when lens coding is enabled help, but not for all of this.

Maybe vignetting is a feature, not a flaw- helps draw your eye to the center of the image. Too much of it makes the lens a "special-use" device. Not enough character in a lens- back to being a special use device.
 
ColSebastianMoran said:
The only one that fails for me is the outdoor shot of two buildings. The two left corners look unnaturally dark.
Click to expand...

Completely agree- but the dark corners to give the impression of an 19th century image.

Same shot, at F1.5:



Also- the wide-open shot was over-exposed for the center, had max'd out the M Monochrom shutter speed.
 
Sonnar Brian said:
Completely agree- but the dark corners to give the impression of an 19th century image.

Same shot, at F1.5:
Also- the wide-open shot was over-exposed for the center, had max'd out the M Monochrom shutter speed.
Click to expand...

That is precisely why I like vignetting in lenses....for that 19th century feel they impart to the image.
 
Focus shift- I just looked at a test done wide-open and down a stop: typical Sonnar, no worse than the Zeiss C-Sonnar.

For an M3: this lens has an adjustment for the RF Cam. I would set optimal focus at F2. The focus at F1.1 with film will be fine, and same with F2.8.

Using with B&W and a deep contrast filter: focus with this lens is towards infinity. I have one lens optimized for the M Monochrom using a deep yellow, that lens is perfect on the M9 at F1.4.

For the M3: avoid early ones, SN below 1500 or so. The M3 has a collar over the shutter release which will jam up on the earlier lenses, including my ~1200 lens. I filed a little off the mount and corrected the error. The early Nikkor 8.5cm F2 lenses had the same problem, someone milled the collar on my 1949 lens down for the M3.
 
Just for comparison: The Voigtlander Nokton 50mm F1.1, at infinity, uncoded on the M Monochrom.



Vignetting at the edges about 1 stop less than the 7Artisans. The Nokton is a Double-Gauss lens. I've had mine for almost 10 years. The Nokton is much bigger and has a longer optical path compared with the 7Artisans lens. The Nokton is an under-rated super-speed lens. I never got my Canon 50/0.95 converted to M-Mount.

The Nokton has less aperture blades- again, I think that is causing the Asymmetry with the lens stopped down. I'm averaging the values from the two diagonal cuts across the image to help account for this.

Compare the "0.5" mark, 1-Stop down vignetting for the two F1.1 lenses. The distance from the center indicates how far out that you are getting an F1.1 lens. The difference is not as big as I expected it to be.
 
You must log in or register to reply here.
Back
Top Bottom