After much back and forth and then some conversation with their facebook page, I finally ended up getting one of A7rII modified.
I won't have any time to do any real tests until this weekend, but in the meantime, I opened up my window and did some tripod comparisons with my regular A7rII using two 35mm lenses, a Voigtlander 35mm f/1.2 II and a Zeiss 35mm f/1.4 ZM. They're really dumb unstructured tests, but the difference is so vast that I just needed to share the results with someone.
The centers are basically identical, but with a slight edge given to the modded sensor. Here's the ZM35: Zeiss 35mm f1.4 ZM Center by Mike Aubrey, on Flickr
At the time I didn't have any expectation that the corners would be matter given the difference in distance, but I must have been close enough to the hyperfocal distance that it worked.
Even the Voigtlander improved dramatically, though certinaly still unuseable at f/1.2 for landscape, but of course, why would a person do that anyway?
Once I've done formal testing with a solid infinity subject across the whole frame this weekend I'll have better data, but right I'm extremely pleased.
Okay, I had a moment this morning to do something slightly more serious.
These still aren't at infinity, but they're a much flatter subject.
Here's a comparison with the Canon FD 85mm f/1.2L on the two sensor stacks. This is a lens I didn't expect a whole lot of improvement. Yes, it was designed for fimlm, but it's also (1) a longer focal length and (2) an SLR design and that tends to negate the effects of the sensor stack. Still, I was quite pleasantly surprised by the degree that the lens improves.
Notes:
(1) Clear increase in purple fringing--likely a result of more sensitivity to the UV end of the spectrum...something to think about.
(2) Slight increase in center resolution at both f/1.2 and f/4.0.
(3) Larger increase in resolution in the midzone and corner at both f/1.2 and f/4.0.
(4) The results at f/4.0 are giving me the kind of resolution across the frame that I used to have to stop down to at least f/5.6 if not f/8.0 for. That's a big win.
mcbroomf wrote:
Nice tests Mike, thanks. I know there was a lot of discussion on the other thread(s?), but could you post details of the mod you chose?
I did version 3, .75mm "extreme" which also removes the cover class itself, but we did everything via NKIR's facebook page and then paypal rather than.
As for white balance, these are all the camera's auto white balance. If I use the eye dropper they look identical, though the original stack is Temp: 5400 / Tint +9 and the thin stack is 5000 / -5. So there is a difference.
I did version 3, .75mm "extreme" which also removes the cover class itself, but we did everything via NKIR's facebook page and then paypal rather than.
As for white balance, these are all the camera's auto white balance. If I use the eye dropper they look identical, though the original stack is Temp: 5400 / Tint +9 and the thin stack is 5000 / -5. So there is a difference.
I used NKIR mod A7R2 for 2.5 years and now used NKIR mod A7III and A7RIII
as u can see the awb very close to the original one,
Maybe u can use AWB adjustment, B1.5-0.5 , G0.5-1
for me when i use A7RII NKIR v3+ i used B1.5 G0.5
This setting will direct save in RAW file
E1ho wrote:
Is Nkir v3 mode really 0.75mm? I saw his facebook page and the image he uploaded showing that v3 mode is 1.75mm. Which one is right??
If you're talking about this post, it say "reduced by 1.75mm."
So total thickness is 0.75mm anyway? which is thinner than Kolari UT? I was going to have thin filter conversion but didn't decide between Kolari and Nkir. Maybe I should go for Nkir. thank you for reply.
I've been talking with NKIR on Facebook and am planning to have my A7RIII modified with a thin sensor. I'm curious what, if any, problems have been encountered by MAubrey and others who have used this service. Are you all satisfied customers?
There are options for a Version 3 or Version 4. NKIR said the V3 can have problems with IR pollution causing a purple coloring in some images. The V4 then has an IR-cut membrane (of negligible thickness) that solves this, but can lead to a green cast along the edges and corners. Have any of you encountered these problems? I'm debating whether to get V3 and use IR-cut filters on my lenses as with the Leica M8, or if I should go with V4 and risk needing to post-process sometimes to remove green edges.
NKIR also talked about a service they perform to check for any dust after the new sensor stack has been installed. They use a lens at f/16 to take some pictures and then examine these on a large monitor for spots. There's also an option to pay a little extra for NKIR to use f/22 for this process. It seems strange to me that turning the aperture ring one more stop costs extra, so I'm wondering if something has been lost in translation (I don't speak Chinese). Can anyone else who has dealt with NKIR shed some light on this confusion?
That is clearly for the service provider to answer.
If I have to speculate, I would think the f/16 vs f22 Corresponds to how clean a sensor you want back.
Opening the epoxied cover glass and replacing it with another is a very risky operation. Hence the charges.
CrazyLeicaFox wrote:
NKIR also talked about a service they perform to check for any dust after the new sensor stack has been installed. They use a lens at f/16 to take some pictures and then examine these on a large monitor for spots. There's also an option to pay a little extra for NKIR to use f/22 for this process. It seems strange to me that turning the aperture ring one more stop costs extra, so I'm wondering if something has been lost in translation (I don't speak Chinese). Can anyone else who has dealt with NKIR shed some light on this confusion?
!...Show more →
CrazyLeicaFox wrote:
NKIR also talked about a service they perform to check for any dust after the new sensor stack has been installed. They use a lens at f/16 to take some pictures and then examine these on a large monitor for spots. There's also an option to pay a little extra for NKIR to use f/22 for this process. It seems strange to me that turning the aperture ring one more stop costs extra, so I'm wondering if something has been lost in translation (I don't speak Chinese). Can anyone else who has dealt with NKIR shed some light on this confusion?
While diffraction will increase with aperture values, aberration decreases in a similar arc. I would guess they are using a flat field uniform illuminant to detect micro and nano dust that may have gotten into the stack during the process. An increase to F/22 would likely involve higher output illumination that still maintains uniformity - a brighter but equally uniform light source - which would detect even finer aberrations from even smaller dust. This could easily be more expensive to calibrate, operate and maintain.
Who knows? I imagine it is a proprietary process. Good luck with your sensor glass.
realVivek wrote:
That is clearly for the service provider to answer.
If I have to speculate, I would think the f/16 vs f22 Corresponds to how clean a sensor you want back.
Opening the epoxied cover glass and replacing it with another is a very risky operation. Hence the charges.
I've been communicating with NKIR on these details, but some understanding is lost in translations back and forth. I was left confused about this issue and hoped those who have gone through this process could explain why f/22 is different and costs more.
The explanations from both of you make sense, that f/22 requires greater precision and necessitates more work for them to achieve. Thank you!
Now I'm just hoping for any users who tried Version 3 or Version 4 to comment on their experience with the results, and how bad is the purple or green coloring.
After trading some more messages with NKIR, I decided to go with the Version 3 mod. It can result in a purple cast, as mentioned by MAubrey earlier, but that can be avoided by using IR/UV cut filters.
NKIR shared a bunch of photos with me showing the results of Version 4, and the cyan edges were very obvious with all wide-angle lenses, requiring lens correction profiles. He said the IR-cut membrane refracts light coming at a steep angle, which causes the loss of red if the angle of incidence is large, even with a BSI sensor.
So I guess there's no easy solution to all of this until someone invents a type of glass that filters all UV and IR light with only a fraction of a mm thickness.
CrazyLeicaFox wrote:
After trading some more messages with NKIR, I decided to go with the Version 3 mod. It can result in a purple cast, as mentioned by MAubrey earlier, but that can be avoided by using IR/UV cut filters.
NKIR shared a bunch of photos with me showing the results of Version 4, and the cyan edges were very obvious with all wide-angle lenses, requiring lens correction profiles. He said the IR-cut membrane refracts light coming at a steep angle, which causes the loss of red if the angle of incidence is large, even with a BSI sensor.
So I guess there's no easy solution to all of this until someone invents a type of glass that filters all UV and IR light with only a fraction of a mm thickness....Show more →
I had debated going this route, but opted for the Kolari Vision UT (Ultra Thin) sensor modification for my A7rII. This was performed in February of this year. Vivek and I had discussed which option to go prior to performing the modification. To both of our shock and surprise, at the time, Kolari Vision performed the modification and returned the camera to me in a week.
naturephoto1 wrote:
I had debated going this route, but opted for the Kolari Vision UT (Ultra Thin) sensor modification for my A7rII. This was performed in February of this year. Vivek and I had discussed which option to go prior to performing the modification. To both of our shock and surprise, at the time, Kolari Vision performed the modification and returned the camera to me in a week.
Rich
I remember reading that thread. I think that's also where someone (Vivek?) talked about doing their own sensor mod that reduced the stack thickness to just 0.7mm.
That page is in Chinese, but my browser extension translator did a decent job, and the most important information is the pictures. I have many of the lenses he mentions there. Among my collection of rangefinder lenses is the Contax G 21 and 28; Zeiss ZM 15/2.8, 18/4, 21/4.5, 35/2.8, and 35/1.4; Leica SEM 21, Elmar-M 24, 28/2.8 ASPH, and 50 Summilux ASPH; and lots more, including film SLR lenses like the Canon 85/1.2 shown above by MAubrey.
Given that my focus is so totally on rangefinder and film lenses, I'm aiming for the very thinnest mod possible. It's quite possible I'll regret it later when I have to fiddle with IR-cut filters all the time, but right now, I figure the extra 0.15mm reduction below the 0.9mm offered by Kolari is worth it.
