Home · Register · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Leica & Alternative Gear | Join Upload & Sell

1
       2       3       end
  

Archive 2014 · Resolution and field curvature

  
 
Fred Miranda
Offline
Admin
Upload & Sell: On
p.1 #1 · p.1 #1 · Resolution and field curvature


I'm seeing an increase of the "effect" of field curvature on some of my wide lenses when shooting with higher megapixels sensors. (From 10MP to 20MP to 36MP).

The field curvature does not change as it's a lens aberration but the effect increases noticeably even when stopping down.
Something to keep in mind as full frame sensors start to break the 50MP barrier.



Feb 03, 2014 at 05:35 PM
shirozina
Offline
• • •
Upload & Sell: Off
p.1 #2 · p.1 #2 · Resolution and field curvature


Also changes with focusing distance and some lenses which perform well at close up I.e have a flat field then change to a curved field at infinity. Now as most lens tests (DXO etc)are conducted at close focusing distances (practically impossible not to with wide lenses ) it's difficult to assess which lens a real world shooter should choose. If I were cynical I would say lens manufacturers are optimising their lens performance for these tests and not for real world use..........


Feb 03, 2014 at 05:53 PM
RCicala
Offline
• • • •
Upload & Sell: Off
p.1 #3 · p.1 #3 · Resolution and field curvature


Shirozina, I've been having the same thoughts. It's why I bought a small optical bench and now am investing in one of the big benches: I want to be able to test at infinity. With DxO and Imatest wide angles are being tested at 6 to 12 feet shooting distance. I'm not sure that correlates well with real shooting.


Feb 03, 2014 at 09:02 PM
Tariq Gibran
Offline
• • • • • •
Upload & Sell: Off
p.1 #4 · p.1 #4 · Resolution and field curvature


Perhaps the effects of field curvature on resolution might also be exaggerated due to the thicker cover glass on some of the higher MP cameras (such as the A7r for instance). This was brought up previously but I don't recall exactly what was said. So, maybe it's not just resolution increase but other variables at play as well.


Feb 03, 2014 at 09:11 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.1 #5 · p.1 #5 · Resolution and field curvature


It would be great if there was a way to test this out...

On the A7R, the effects are much more pronounced when shooting ultra-wides like my TS-E 17mm and Samyang 14mm.

In Live View, focusing on the center and then the corners at around 20 feet distance, gives either sharp center OR corners. They never get both sharp with the same focusing distance with a flat subject indicating field curvature. What was interesting was that I never noticed such effect with these lenses on 5D Mark III (22MP)

When going from sharp center and refocusing to get sharp edge (normal for lenses with moderate field curvature), the distance travelled (focusing scale) is the same when testing with the A7R or 5D Mark III but the discrepancy between center / corner sharpness is much bigger with the A7R.

Stopping down does not seem to minimize the effect that much on the A7R and on the 5D Mark III, it's not that noticeable.

I have wondered if this was caused by the adapter, cover glass thickness or resolution increase.
To me, it was caused by the resolution increase as the effects of field curvature are much more noticeable nowadays than in the film era.

Perhaps we could ask lens or adapter designers what's going on here.



Feb 03, 2014 at 10:57 PM
ebrandon
Offline
• • •
Upload & Sell: Off
p.1 #6 · p.1 #6 · Resolution and field curvature


I'm so glad someone else said this, because I've been afraid to say it myself for fear of sounding stupid.

I think many lenses have more field curvature on the A7R than they did on other cameras, but I can't be sure if it's real or just more visible because of the increased resolution & detail.

The increased resolution and detail is something else when you're testing lenses. I was testing my TS-E 24mm f3.5 v2 on the A7R and came to the conclusion that it had to be tilted a smidge to the left to be "tilt neutral". Put it on the Canon 6D to test this theory, and the resolution just wasn't there to be sure.

Today I did a big shootout of my 45mm-60mm lenses on the A7R, and noticed that the Konica 57mm f1.2 has wavy field curvature with sharp centers and edges, and softer areas in between. Never noticed THAT before on m43 cameras or the NEX 5N.

Anyway, I'd love to get to the bottom of this increased field curvature on the A7R issue.



Feb 03, 2014 at 11:58 PM
douglasf13
Offline
• • • • •
Upload & Sell: Off
p.1 #7 · p.1 #7 · Resolution and field curvature


ebrandon wrote:
I'm so glad someone else said this, because I've been afraid to say it myself for fear of sounding stupid.

I think many lenses have more field curvature on the A7R than they did on other cameras, but I can't be sure if it's real or just more visible because of the increased resolution & detail.

The increased resolution and detail is something else when you're testing lenses. I was testing my TS-E 24mm f3.5 v2 on the A7R and came to the conclusion that it had to be tilted a smidge to the left to be "tilt neutral". Put it on
...Show more

Maybe try downsampling the size of the A7R's file to the same size as the 6D's file and give it a look??



Feb 04, 2014 at 12:00 AM
Peter Figen
Offline
• • • • •
Upload & Sell: On
p.1 #8 · p.1 #8 · Resolution and field curvature


One factor that might come into play is that almost all modern retrofocus wide angle lenses use floating elements to optimize performance at different focusing distances. Being off even a fraction of millimeter in adapter thickness can put those floating element sets in a less than optimal position for the distance being focused on. Whether this affects curvature of field in infinity shots, I don't know, but it certainly affects test results of lenses being tested at much closer than infinity distances.

And don't forget that those new Canon tilt shifts (the 17 for sure) have two different sets of floating elements to optimize. Where I see that coming into play is a noticeably increase in barrel distortion as you focus closer, but the distance seems to be very critical. You can also see the effect using a 12mm extension on the 24 t/s. The image quality completely falls apart with huge amounts of color fringe, looking more like a plastic Diana lens than the precision optic it is.

Also, have you mic'd your adapter to see that it's actually exactly the same thickness at all points around the mount? You could easily have an alignment issue that is contributing to or masking another defect.




Feb 04, 2014 at 12:27 AM
Gunzorro
Offline
• • • • • •
Upload & Sell: On
p.1 #9 · p.1 #9 · Resolution and field curvature


Besides "normal" curvature, which I'm very familiar with from using Canon DSLRs from 16MP and up, I suspect the aberration is exaggerated by the engineering of the a7R sensor toward the edges, with the slant induced.

I also agree that lab tests at short distances do not conform to real world results from close up to infinity. When I run a lens through its paces it is always in real world distances, usually in conjunction with establishing the MFA needed for best average of all distances. Since most of my shooting will be at middle distance to infinity, that's where I place the greatest emphasis. No magic formula, just a lot of trial and error observations and staring at 100% details on the computer.

I'm very eager to hear what people with real science and measuring devices have to say about all this.



Feb 04, 2014 at 12:36 AM
shirozina
Offline
• • •
Upload & Sell: Off
p.1 #10 · p.1 #10 · Resolution and field curvature


I see it on my 5D2 both in live view and on the screen at 100%. I also see focus planes which are not parallel with sensors. Both can look the same as they have center sharpness but edge softness but one edge is soft due to back focus and the other to front focus. My 24TS-E mk2 does this and needs F11 to cure it. Stopping down will help as the depth of focus increases to cover more of the curvature or slant but remember the shorter the focal length then the shallower the depth of focus (not field). Field curvature with focusing distance is well understood even back to the days of film ( and film was never as flat nor as high a resolution as digital)Floating lens designs were partly to address this but as mentioned the flange to film/ sensor distance needs to be exactly on spec. Wide lenses where the whole group moves were often poor performers up close as they were optimised for long distance. I had a Hassleblad 50mm CF which I replaced with the FLE version which I still use via a mirex adapter on my 5d and this has an adjuster for focusing distance and it does work. The edges can be brought into much higher resolution using it esp when shifted I.e using the outer edges of the image circle. I have a Samyang 14mm which was awful out of the box but I partially dismantled it and found the adjusters for the focusing mechanism and managed to get a better balance between edge and center. This was a compromise as there was some loss of center sharpness in order to gain an improvement at the edges. On the A7R I'd expect losses from diffraction to kick in before you could cover the problems with stopping down. I'm toying with the idea of buying one with an eos adapter to replace my 5D2 (5D3 is not an upgrade to a base ISO shooter) but I've got my reservations. Some M43 lenses are terrible in this respect my 12-35 and 20mm are the worst. They shoot perfect flat test targets a few m away but for landscape the edges are mush. I fear we are living in a world of lens design that is being dictated to by a very narrow and unrepresentative testing and measurement procedure.


Feb 04, 2014 at 02:59 AM
philber
Offline
• • • • • •
Upload & Sell: Off
p.1 #11 · p.1 #11 · Resolution and field curvature


I don't think it affects only wide angles. The Summilux 50 ASPH is a good example of what Fred describes. I couldn't get sharp corners at infinity before f:8.0. I thought it was a mismatch with the sensor and registry, until I focused the corner itself and noticed how much better it was, but the image centre was OOF. That suggested field curvature. A blogger called 3D-Kraft proved it to my satisfaction in a picture of wavelets in a harbour. Massive field curvature can be seen readily, where wavelets are sharp on the sides, and very much OOF at the same distance, but in the centre.
I am right now testing 2 28mm Olympus, a f:2.0 and a f:3.5. The difference in "apparent DOF" is striking, where there should be little or none. This is interesting, because they are DSLR lenses so, in theory, unaffected by the short registry and the "compatibility" issues. Also I shoot both with the same adapter, which equalizes that part of the equation.
That suggests, though I not through with testing, so I am only tentative, that the A7R is a lot more revealing of transitions from in focus to OOF than 5DII and 5D III which I used before, also with FF lenses. That said, the fact that 100% magnification is far greater on the A7R also plays a significant part in revealing the problem IMHO.



Feb 04, 2014 at 04:12 AM
Jon Tainton
Offline
• •
Upload & Sell: Off
p.1 #12 · p.1 #12 · Resolution and field curvature


Fred Miranda wrote:
I'm seeing an increase of the "effect" of field curvature on some of my wide lenses when shooting with higher megapixels sensors. (From 10MP to 20MP to 36MP).

The field curvature does not change as it's a lens aberration but the effect increases noticeably even when stopping down.
Something to keep in mind as full frame sensors start to break the 50MP barrier.


IIRC the Canon 17 & 24 tse's have been tested on Hartlbei and other MDF cameras, no one commented on field curvature.

There's some recent test images of the the Phase IQ250 back +Alpa +17tse here http://www.alpa.ch/en/news/2014/IQ250-review1.html?year=2014&num=1 you could drop an enquiring email to ask if they were aware of any field curvature 'effects' compared with 35mm format. If nothing else it would eliminate mega pixels from the issue you see with the A7R.



Feb 04, 2014 at 04:31 AM
AhamB
Offline
• • • • •
Upload & Sell: Off
p.1 #13 · p.1 #13 · Resolution and field curvature


Peter Figen wrote:
One factor that might come into play is that almost all modern retrofocus wide angle lenses use floating elements to optimize performance at different focusing distances. Being off even a fraction of millimeter in adapter thickness can put those floating element sets in a less than optimal position for the distance being focused on. Whether this affects curvature of field in infinity shots, I don't know


The Samyang 14/2.8 doesn't have floating elements though and it shows the same increase in FC, as reported by Fred and others. I'm guessing the Konica 57/1.2 doesn't have floating elements either.

My guess is that the sensor of the A7R increases astigmatism due to the cover glass Sony used. Maybe the curious ghost reflections in night shots that it has are a related issue (although this is seen on the A7 as well). Perhaps it has something to do with the microlens array too -- the short FFD of the FE mount probably sets different requirements for native lenses, possibly impairing performance of SLR lenses to some degree.



Feb 04, 2014 at 04:57 AM
edwardkaraa
Offline
• • • • • •
Upload & Sell: Off
p.1 #14 · p.1 #14 · Resolution and field curvature


One more vote for the thick cover glass being the main cause. Keep in mind that a thin cover like the one on the Leica M9 and M240 has insufficient IR filtering, and breaks easily. I would prefer a thin cover personally for less field curvature/astigmatism in the corners with wide angles, and would be willing to sacrifice some purple contamination in the blacks and some not very appealing skin tones.


Feb 04, 2014 at 05:44 AM
SoulNibbler
Offline
• •
Upload & Sell: Off
p.1 #15 · p.1 #15 · Resolution and field curvature


Just a bit of link sharing, I saw this a few weeks ago
http://image-sensors-world.blogspot.co.at/2014/01/camera-module-image-quality-testing.html
I think there is some optimization for the tests going on but not for nefarious reasons, simply because they tests that are available and in use don't specifically handle infinity well.

As far as the issue with the A7 cover-glass, I don't think it will be that long before we see naked versions tested. In reality the camera is still quite young.



Feb 04, 2014 at 05:55 AM
BrianVS
Offline
• •
Upload & Sell: Off
p.1 #16 · p.1 #16 · Resolution and field curvature


The cover glass on the M9 is 0.8mm, a little thicker than that used by Canon. Cover glass on the M8 is 0.5mm.

Seeing some test images would be interesting, something taken wide-open that shows the curvature moving in/out of the plane of focus. Sonnar formula lenses have a high field-curvature. I'm used to this on the Leica, discovered repositioning the rear triplet could reduce it somewhat.



Feb 04, 2014 at 05:56 AM
AhamB
Offline
• • • • •
Upload & Sell: Off
p.1 #17 · p.1 #17 · Resolution and field curvature


SoulNibbler wrote:
As far as the issue with the A7 cover-glass, I don't think it will be that long before we see naked versions tested. In reality the camera is still quite young.


Naked? You can remove the IR filter, AA-filter (if present) and even the color filters, but I think you need a cover glass at the very least. Otherwise you can't clean the sensor without destroying it.



Feb 04, 2014 at 06:56 AM
SoulNibbler
Offline
• •
Upload & Sell: Off
p.1 #18 · p.1 #18 · Resolution and field curvature


Yeah but if you are willing to risk destroying the sensor for testing I'm sure taking a few pictures without any cover glass is within boundaries.

Actually I'm uncertain of the terminology. It was my understanding that many sensors have a very thin quartz cover glass beneath the IR filter. Somehow we keep talking about the cover glass in the A7R but it seems we are really talking about the combined hot-mirror (IR filter) cover glass /anti-stick combo, that in many other cameras is a combined hot-mirror, AA filter, and cover glass.

I ask because I have a friend who claims that there was a bonded cover glass beneath the IR/AA filter on the NEX5 that he had to remove when he was trying to remove the color filters.

Also it seems like the sensor surface is not as fragile as one might guess: http://image-sensors-world.blogspot.co.at/2013/08/passivation-strength-test.html



Feb 04, 2014 at 09:00 AM
RCicala
Offline
• • • •
Upload & Sell: Off
p.1 #19 · p.1 #19 · Resolution and field curvature


SoulNibbler wrote:
Just a bit of link sharing, I saw this a few weeks ago
http://image-sensors-world.blogspot.co.at/2014/01/camera-module-image-quality-testing.html
I think there is some optimization for the tests going on but not for nefarious reasons, simply because they tests that are available and in use don't specifically handle infinity well.

As far as the issue with the A7 cover-glass, I don't think it will be that long before we see naked versions tested. In reality the camera is still quite young.



Thank you for the link - it my day. We arrived at a similar conclusion and are having a machine spec-built in a similar way. Rather than 9 collimators we're using 3 collimators with a rotating lens plate. It will get similar information but require a bit more time: 3 total images with the lens rotated 45 and 90 degrees. Hopefully that will give exactly the same information.



Feb 04, 2014 at 10:36 AM
sebboh
Offline
• • • • • •
Upload & Sell: Off
p.1 #20 · p.1 #20 · Resolution and field curvature


shirozina wrote:
I see it on my 5D2 both in live view and on the screen at 100%. I also see focus planes which are not parallel with sensors. Both can look the same as they have center sharpness but edge softness but one edge is soft due to back focus and the other to front focus.


this is decentering not field curvature that you are describing. it could be due to imperfectly aligned lens element(s) or lens/camera mount. the wider a lens is the more obvious such alignment issues become. if it was field curvature, both edges would be back focused or both edges would be front focused.




Feb 04, 2014 at 12:29 PM
1
       2       3       end




FM Forums | Leica & Alternative Gear | Join Upload & Sell

1
       2       3       end
    
 

You are not logged in. Login or Register

Username       Or Reset password



This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.