ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Nikon D850 OVF and Internal Focus Indicator Accuracy with 3 Otus Lenses
- Thread starter ArthurMeursault
- Start date
Righto.
By best optical focus I meant focusing by eye on what passes for a ground glass these days.
I will probably add CDAF for both cameras, since that's easy to do.
Did you use peaking? It makes all the difference if you know how to use it right and you use it at max mag.
http://blog.kasson.com/the-last-word/manually-focusing-the-fujifilm-gfx-50s/
Jim
--
http://blog.kasson.com
Last edited:
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
I did not use peaking. The actual difference turned out to be a stray fray on part of a rope that wasn't visible with the Sony set up or photo and well defined with the Nikon set up and in the photo.
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
Not sure, but using a tripod for all shots, stellar alignment for sensor and target, live view and remote trigger the scores (across the board) were worse than yesterday's hand held.
I have to pay attention to family for a while but I'll try again tomorrow.
I have to pay attention to family for a while but I'll try again tomorrow.
Welcome to testing.
No hurry. We have plenty of time. You've got your priorities right.
Jim
Bill Janes
Senior Member
Jim,Here's what you get with the D850 and the Sigma 50 ART, which is not as sharp as the Otus 55, at f/4 if you let the D850 focus shift feature find the sharpest point. (The steps on the FSS are too large, so this isn't quite in focus). Demosaiced with dcraw with (I think; is the string sticky?) Imatest standard option string.
I don't recommend developing in Lr or C1, since you can't be sure all sharpening is really off (in Lr and ACR, you can be sure all sharpening is not turned off, no matter what you do.) I used the dcraw for this. Imatest makes that easy, and ships with dcraw.
Jim
--
http://blog.kasson.com
MTF at Nyquist is nearly 50%, considerably higher than I have seen with Imatest and my Nikon d800e, which also lacks a blur filter. Is this aliasing due to the lens's high resolution at Nyquist and how does it affect the quality of the observed image? Would incorporation of a burr filter improve the perceived image quality?
Best Regards,
Bill Janes
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
From memory, there is a small range around 16% improvement and this is for C1's import pre-sharpening 1 only. (16% difference between no sharpening and pre sharpening 1).
So it could turn a pigs ear into a pretty pigs ear but never into a silk purse.
I have never understood (for certain) why the bottom score in Imatest is lower than the upper (higher) store except for the pattern of this typically occurs when the image has been sharpened.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
So it could turn a pigs ear into a pretty pigs ear but never into a silk purse.
I have never understood (for certain) why the bottom score in Imatest is lower than the upper (higher) store except for the pattern of this typically occurs when the image has been sharpened.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
Last edited:
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
Hey Jim,
I've got to ask. Are you pulling fast one here ?
The challenge is 1500 cy/ph but I am looking at several of your lab tests, done with all advantages (especially considering compare to hand held and focus indicator only) and many of your tests (even in lab conditions with an Otus 85mm) don't meet the 1500 level.
http://blog.kasson.com/the-last-word/nikon-d850-mirror-shock-with-otus-851-4/
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
I've got to ask. Are you pulling fast one here ?
The challenge is 1500 cy/ph but I am looking at several of your lab tests, done with all advantages (especially considering compare to hand held and focus indicator only) and many of your tests (even in lab conditions with an Otus 85mm) don't meet the 1500 level.
http://blog.kasson.com/the-last-word/nikon-d850-mirror-shock-with-otus-851-4/
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
Last edited:
And the target, and the focusing method. As I said before, the Sigma is not quite as sharp on-axis at best aperture as the Otus 55. This image used a backlit razor blade for the target, which is the gold standard for sharpness. Your D800E should have slightly higher MTF50 at Nyquist because of the lower pixel pitch.Jim,Here's what you get with the D850 and the Sigma 50 ART, which is not as sharp as the Otus 55, at f/4 if you let the D850 focus shift feature find the sharpest point. (The steps on the FSS are too large, so this isn't quite in focus). Demosaiced with dcraw with (I think; is the string sticky?) Imatest standard option string.
I don't recommend developing in Lr or C1, since you can't be sure all sharpening is really off (in Lr and ACR, you can be sure all sharpening is not turned off, no matter what you do.) I used the dcraw for this. Imatest makes that easy, and ships with dcraw.
MTF at Nyquist is nearly 50%, considerably higher than I have seen with Imatest and my Nikon d800e, which also lacks a blur filter. Is this aliasing due to the lens's high resolution at Nyquist
You can certainly see aliasing in some circumstances. It's amazing how little of the time that this causes a visual problem.
I believe so. I am not a fan of AA-less cameras.
Jim
--
the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.
Photography meets digital computer technology. Photography wins -- most of the time.
blog.kasson.com
First off, if you go back and read what I said, it wasn't a challenge. Second, this is using a not-very-sharp target. Third, I have not figured out how to consistently get focus as accurate with the D850 as with the GFX and the a7RII. Fourth, the camera was only focused once for the entire series, it's not a statistically-valid sample.
Jim
Are you sure you can turn it all the way off? Have you compared with dcraw?
I don't know what you mean by bottom score and upper score.
Jim
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
The Imatest charts - top and bottom. Bottom is usually higher.
Notice it isn't on my import sharpened test.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
Notice it isn't on my import sharpened test.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
I still can't tell what numbers you're talking about. Are you talking about the graphs? If so, they are quite different things. The top graph is the reconstructed slanted edge profile. The bottom graph is the Modulation Transfer Function. Neither the vertical or the horizontal axes are the same. The top graph is in the space domain. The bottom one is in the frequency domain.The Imatest charts - top and bottom. Bottom is usually higher.
Notice it isn't on my import sharpened test.
Or maybe I'm completely off base and you're talking about something else entirely.
Jim
--
the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.
Photography meets digital computer technology. Photography wins -- most of the time.
blog.kasson.com
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
Understood. What I am saying is that when the slant edge is artificially sharpened in post then the upper number (PH) is higher than the lower number (LW/PH).
When it's an 'honest' test the lower number (LW/PH) is always the higher number. Try it and see.
At least that's the pattern I've noticed.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
When it's an 'honest' test the lower number (LW/PH) is always the higher number. Try it and see.
At least that's the pattern I've noticed.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
Last edited:
Here's the razor-blade test:
Towards a reproducible MTF testing protocol - the last word
There has been a little interest in developing a protocol for slanted edge modulation transfer function (MTF) testing that is sufficiently standardized and reproducible that the results from various amateur practitioners can be compared fairly. There are several things that conspire to make...
blog.kasson.com
Here's what it looks like from the camera:
It is very difficult for a human to focus on a backlit razor blade. If you're going to try this, turn on a light in front of the razor to focus, then turn it off for the shot.
Here's a documentation of the peaking technique I use, with the GFX as the example camera.
Manually focusing the Fujifilm GFX 50S - the last word
This is the 51st in a series of posts on the Fujifilm GFX-50S. The series starts here. I am reading many complaints about photographers who are having difficulty in manually focusing the Fuji GFX, both with the native focus-by-wire lenses and adapted ones. I have developed what I think are...
blog.kasson.com
You will find that most cameras, including the D850, don't have the fine control of the peaking threshold that you'd really like them to have.
Jim
--
the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.
Photography meets digital computer technology. Photography wins -- most of the time.
blog.kasson.com
Gotcha, I think.
You're talking about MTF50, right? If so, it all depends on the amount of sharpening. Let's say the raw file MTF50 is 0.2 cy/px. That is about 0.4 lw/px. To make the math easy, let's say the PH is 5000 px. So the MTF50 of the raw file is 1000 cy/ph, or 2000 lw/ph. Now I can sharpen the image a little (and many demosaicing algorithms sharpen as they demosaic) and make the MTF50 2000 cy/ph or 4000 lw/ph. I'm still under your threshold. Or I can sharpen the image more, and make the MTF50 3000 cy/ph or 6000 lw/ph. Now I'm over your threshold.
So, that's not a good test.
JIm
I should be clear on one point which may provide important context for this whole discussion. I have found no focus mode on the D850 as reliably sharp as Focus Shift Shooting with minimum step size, even though that method is deeply flawed as a focus bracketing mechanism.
In fact, I've not found a focusing method on any camera except the Betterlight Super 6K that is as consistently good as FSS.
Jim
In fact, I've not found a focusing method on any camera except the Betterlight Super 6K that is as consistently good as FSS.
Jim
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
So we got some results today and once again as deeply disturbing as the election results last year. (More exactly, the candidate choices were the only thing worse than the results).
On the tripod and using OVF and indicator for focus and racking between shots, this series of 5 (f2-f8 whole stops) best result 1337 at f2 but with f4 and f8 very close behind
Tripod shots f2-f8 (l-r)
Now for the crazy part - 2 handheld shots using the OVF and indicator only scored above the best tripod shots.
Here's they f2 and f4.
My opinion. The Nikon focus indicator works very well and is reliable. As you know, this is a lens with a lot of focus shift and at 3.3 meters from the target.
In my experience these scores are more than sharp enough.
Handheld at same distance.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
On the tripod and using OVF and indicator for focus and racking between shots, this series of 5 (f2-f8 whole stops) best result 1337 at f2 but with f4 and f8 very close behind
Tripod shots f2-f8 (l-r)
Now for the crazy part - 2 handheld shots using the OVF and indicator only scored above the best tripod shots.
Here's they f2 and f4.
My opinion. The Nikon focus indicator works very well and is reliable. As you know, this is a lens with a lot of focus shift and at 3.3 meters from the target.
In my experience these scores are more than sharp enough.
Handheld at same distance.
--
"A gentleman is one who never hurts anyone’s feelings unintentionally."
Last edited:
ArthurMeursault
Forum Enthusiast
- Messages
- 312
- Reaction score
- 462
Do you see FFS bursts as a way of ensuring optimum focus with a live subject ?
Forget about stacking, just picking the best image of the bunch?
Such a waste of shutter count though.
Forget about stacking, just picking the best image of the bunch?
Such a waste of shutter count though.
Similar threads
Keyboard shortcuts
- f
- Forum