NikonGear'23
Gear Talk => Camera Talk => Topic started by: Akira on March 14, 2017, 12:04:46
-
It has just posted.
You can compare GFX, Nikon D810, Sony A7RII and Pentax 645Z.
https://www.dpreview.com/reviews/fujifilm-gfx-50s
To my eyes, at ISO 100, the image from Fuji is the most ordinary. The peripheral area (green and brown feathers) are the least sharp among these four.
In short, it seems to coincide with our Michael's findings.
-
Wow - What an amazing amount of pixels from the GFX,,, :)
To my eyes the Nikkor 85mm AFS 1.4G at f/5.6 helps the D810 produce an admirable comparative image to the GFX with the 63mm,,,
-
The D810 images show a lot more false colour (moire) in the white text than the images from the GFX.
-
Yes, the fidelity of the printed letters is impressive.
I wonder whether the sensor has the AA filter? I find no mention of the AA filter on their official webpage...
-
I have read that they are without AA filter,,,
-
While the Fuji is quite impressive, it is equally impressive at what Nikon is capable of pulling from the 135 sensor in the D800 series cameras.
-
The D810 images show a lot more false colour (moire) in the white text than the images from the GFX.
Apparently Pentax645Z shows less moire in the same white test.
I have read that they are without AA filter,,,
Thanks, Erik. Pentax645Z doesn't have AA filter either, so Pentax seems to be able to control moire better.
While the Fuji is quite impressive, it is equally impressive at what Nikon is capable of pulling from the 135 sensor in the D800 series cameras.
Indeed. Also, I'm surprised how Nikon retains rich, saturated color at high ISO. D750 is even more impressive, despite the lower resolution.
-
Maybe it's just that the Nikon lens is so much higher resolution that with 36MP there is moire whereas the MF lenses are not. Not a fault of the camera as such but shows that the lenses would do better with a higher resolution sensor.
-
Possibly. And maybe that proves the difficulty to retain the image quality (or the flatness of the field) in the larger area.
-
Are you guys now seeing those problems with dynamic range that have been talked about recently on this forum?
-
Yes and they mention it in the text,,, The histogram looks different than D810, more compressed and ETTL, when opened in Adobe RAW and this is at ISO 100 the D810 goes down to ISO64,,,
-
If I may ask: how can you infer the dynamic range from the degree to which the histogram is bunched up in ACR (note that ACR does not display a RAW histogram, but a histogram after conversion)?
I do not refer to the dpreview text since I cannot test their findings now. I'm asking whether you are seeing the problems yourself.
-
Are you guys now seeing those problems with dynamic range that have been talked about recently on this forum?
In general, I don't think it really possible to assess the dynamic range from the studio scene shots. The metal palettes on the above left and the below right corners are always clipped abruptly and the details are lost. I mostly assess the resolution on the feathers, high-ISO noise on the wine bottle labels and the grey background and color saturation on the Macbeth color chart.
-
Sorry, but I don't understand,,, 'bounced up' ? I do, but what is the problem, this is not a sciences lab,,,
I don't have the camera at hand, I have not shot any images with it, I'm just giving my opinion on what I see in ACR. And what people put up online. You are entitled to see something else with the same converter or another RAW converter :)
Why will you not refer to what they write, this is just their opinion and thoughts, just as I have mine, but OK feel free to not comment on it,,,
To my eyes the samples are pretty close regarding most things, I said the GFX has a lot of pixels but the 85mm AFS 1.4 G looks like it is helping the D810,,, implying they should have done the test with the 120mm that seems to be a better performer according to reviewers all around, but I can only state what I see from looking at these samples,,, and what Michael said, and I agree with him on his exposures and the look and feel of the files,,,
You will not see me make a statement on if the GFX has a DR of 12 or 14, you should know that - I leave that for the number crushers out there ;)
My point is sort of; We hear this GFX is a new wonder and much better than old obsolete cameras like the vintage D810 - Many people say they would like a new D820 since it would be much better,,, Not so, new is not always better I have learned over the years - just look at the D810 it has unforgivable flaws compared to some of it's predecessors,,,, for some users,,, Chassis/mount weak, small battery and need for an grip to mention just obvious ones,,,
-
Here is the quote
Dynamic Range
,,,
However, the exposure latitude test (https://www.dpreview.com/reviews/image-comparison/fullscreen?attr144_0=fujifilm_gfx50s&attr144_1=nikon_d810&attr144_2=fujifilm_gfx50s&attr144_3=nikon_d810&attr146_0=100_6&attr146_1=100_6&attr146_2=100_5&attr146_3=100_5&normalization=full&widget=488&x=0.142606214&y=0.5005349), where we lift the shadows in images shot at progressively lower exposures shows that its performance is only slightly better than that of the D810, despite receiving more total light (double the exposure time and half the light per square cm, captured on a sensor with more square cm of area). Now consider the fact that the D810 has an ISO 64 mode, which would allow you to use a 2/3EV brighter exposure before the sensor clips. We expect this will give a real-world result similar to when we pitted the Pentax 645Z against the Nikon (https://www.dpreview.com/reviews/image-comparison/fullscreen?attr226_0=nikon_d810&attr226_1=pentax_645z&normalization=compare&widget=479&x=0.25890258&y=0.269753247).However, Fujifilm does claim to have increased the 'Photic Saturation Point' (by which we assume they mean 'full well capacity'), by 1/3EV so we'll need to conduct a full expose-to-the-right side-by-side test to be sure.
I can't see an issue with this statement. Quite the opposite, it's a mighty fine sensor.
-
Sorry for any misunderstandings...
There is no obvious problem with the dpreview statement. It is just that I cannot verify it now (it would be easy to verify, given enough time).
What I was asking is whether people that were commenting on some dynamic range issues based on JPEGs, can now see the same issues in these RAW files.
Nothing more, nothing less! :)
You said that you can see it from the ACR histogram, but I'm not able to follow your reasoning. How do you do that?
That's why I asked for clarifications.
P.S. I said 'the histogram is bunched up' which is a synonym for your expression 'the histogram is compressed'.
-
Now I understand you.
Comparing the D810 and GFX histogram samples at ISO 100, to me it looks like the GFX is done ETTL this would to me suggest a too dark image, as compared to a neutral one, where you don't take advantage of the GFX sensor to the fullest. The D810 looks more neutral or toward ETTR.
Could explain the 'dark shadow' JPG files we have seen so far,,, since people didn't have the RAW converters until now,,,
-
I see.
The problem with this is that simply using a different camera profile in the 'Camera Calibration' tab you could move the histogram around as you please.
I.e. the histogram depends on what Adobe deemed a 'standard conversion'. This is a fairly arbitrary concept.
I was also wondering why the Fuji file is darker than the D810 file when opened in ACR with 'Adobe standard' profile.
You might be surprised that the Fuji file has slightly brighter RAW levels than the D810 file (only about 0.1-0.2 stops as measured in RawDigger).
Any ideas?
EDIT: added RawDigger screenshots.
-
I understand that its just a 'standard' or an educated guess at a 'normal' convertion,,,
Ok we observe the same issues then :)
It seems strange I agree, my first guess was lack of highlight recovery/tolerance of the GFX sensor,,, to avoid blown highlights and blooming,,,
-
Now I see the RAW Digger histogram;
Look at the spikes on the right,,,
Look at the saturated areas left,,,
Exactly what I am trying to say with words,,, I'm no expert but this looks strange to me,,,
I have never seen a histogram mapping before :o
-
It seems strange I agree, my first guess was lack of highlight recovery/tolerance of the GFX sensor,,, to avoid blown highlights and blooming,,,
You might be onto something. You can see in the RawDigger histograms that some separations between peaks are different. This means that the linearity of the light response of the two sensors is different.
-
Look at the spikes on the right,,,
Look at the saturated areas left,,,
Please note that the test images are different. The Nikon image has some additional checkerboard patterns that the Fuji is missing due to different aspect ratios. This checkerboard pattern is what I think is causing the higher number of peaks on the right side of the histogram.
I think for differences in dynamic range and such you have to look at distances between peaks that are visible in both histograms. If the distance is smaller on the EV scale then it means that one sensor produces data which is nonlinearly distorted wrt. the other sensor.
-
Is the Adobe Standard, or ACR per se, immune to the encrypted white balance data in NEF? If not, the difference of the exposure in each color channel could be affected.
-
ACR does read white balance set in the camera from NEF files.
-
Thanks for the confirmation. Then the difference of the exposure of each color channel could be caused by the different white balances.
-
White balance is done by multiplying the channels with some numbers in order to achieve the same value in all channels at the white balance target. But white balance has two degrees of freedom while there are three colors, RGB. So one of the multipliers can be considered to be one while the other two are being tuned relative to that. An overall multiplication of all channels by the same number corresponds roughly to the exposure slider in ACR. However, how bright an image the zero position will result in depends on other controls as well as the camera profile, which includes nonlinear transformations of all channels (gamma curve, to get from the linear response of the sensor to something the human eye can digest, i.e. logarithmic) as well as individual channels (color profiling, which consists of channel mixing and a subsequent curve per channel).
The white balance in ACR is different between the Fuji and Nikon, but note that for the same white balance target there need to be different sets of multipliers for the two cameras because the color filter arrays are likely to be different. I can see from the RAW histogram that the blue response is stronger in the D810 while the red response is weaker. But all of these things, while interesting, IMHO are only weakly related to dynamic range.
-
Isn't there any possibility to tweak the contrast or the tone curve of the individual color channels in addition to balancing the levels?
By the way, I'm also interested in the comb-like pattern on the large and wide peaks in the histograms of the blue and the red channels. I wonder if they would suggest the posterization or lack of gradation due to the fewer pixels for these colors?
-
After reading a lot about the GFX I feel that the new sensor might not be perfectly embedded yet into the firmware and RAW converter environment. Due to its bigger area and lower pixel density it should outperform the a7r2 sensor easily. That seems not to be the case.
I feel a fair test will be to put both the D810 and the GFX on a bellows and put a Schneider APO Digitar on the front standard.
Michael can easily do that. The lens easily outperforms both sensors and with the same light and the same scene we have a much better starting point for discussion.
For better firmware and RAW converter support we can only wait.
Yes. We should calibrate the workflow on a per sensor / standardized target basis also using the same lens.
Thank you.
-
I think Michael returned the camera,,,
-
The A7R II has a BSI sensor so it probably collects a bit more light per area. The D810 has the ISO 64 which too works to its advantage when there is enough light to use it.
When I was shooting 35mm film in the 1990s, I looked at images printed from 6x4,5cm film, 6x6cm , 6x7cm , 4x5 inches and 8x10 inches. I felt the 6x4,5 was a bit better than 35mm but to get really blown away, 6x7 cm and bigger did that. I got a 6x7 rangefinder and shot with that for a few years, and yes, the image quality was great, the detail in print was amazing once properly scanned (LS-9000 in my case), but most lenses were designed with compactness in mind and I wasn't that happy with the relatively small apertures. The 6x7cm diagonal is sqrt(56*56+72*72) = 91mm vs. 43mm for FX; the ratio between the diagonals of the two formats is 2.1:1. For the 33x44mm sensor the diagonal is 55mm; ratio between the diagonals of the two formats 1.28:1. So the sensor sizes are not different enough to make a really clear difference in image quality especially given the variation in lens characteristics in each system. Yes, the images shot with the larger sensor are probably a little better, but I would consider the system as a whole and see if it offers features which together with the larger sensor size make it worthwhile.
For example, the X1D has the leaf shutter lenses which are useful for outdoor portraits with flash in particular, the GFX has a tilting EVF which I would imagine would be very useful for some kinds of technical photography and, e.g., shooting portraits in sunlight if you want the camera to be at waist level without kneeling down; the D810 and 645Z have optical viewfinders etc. All these differences make the systems different in what you can do with them even though image quality differences are not that large. It is possible that Fuji, Hasselblad and Pentax make 100MP variants of the camera bodies in the near future; perhaps the differences will be more pronounced at that point.
-
Akira,
I completely agree with your math for "medium format" vs 35mm and likewise for the relative difference in image quality. Doubling the film/sensor size is quite noticeable. An increment of 28% is not likely to be something that makes you shout.
In the days of film I started with 35mm but soon went to 6x6, 6x9 and 4x5 because they offered such a different look and tonality.
Each has a different purpose and way of making photographs.
-
By the way, I'm also interested in the comb-like pattern on the large and wide peaks in the histograms of the blue and the red channels. I wonder if they would suggest the posterization or lack of gradation due to the fewer pixels for these colors?
The comb patterns could be the light used to take the sample. It could be the subject particularly if it contains printed materials rather than a natural subject. It could be post processing if harsh but in this case I don't think so. RAW images get some processing in camera. Ii could be low bit depth, e.g. 8 bit JPG files. It could be a strong "S" curve added somewhere as the mid tones would get stretched. I'm probably missing some possibilities. It's probably a combination of some of the above. Some don't apply to a RAW image and some do.
The last place I'd look for a cause of the comb pattern with these cameras is the availability of pixels. There are too many available.
Dave Hartman
-
Isn't there any possibility to tweak the contrast or the tone curve of the individual color channels in addition to balancing the levels?
We have to distinguish between RAW data and data that has been mapped to a target color space (e.g. Adobe RGB or sRGB).
When you mix channels or apply curves to them in the RAW data, you get the degrees of freedom that are typically used for color profiling.
This is significantly more than just white balancing. White balancing tries to equalize the channels in a single target.
In color profiling you have a bunch of target, e.g. a color checker chart and you then solve an optimization problem where you try to minimize the deviations of each patch from the target value. Then there are different ways of doing it depending on the software, e.g. DNG profiles in ACR. They are actually computing something that does not interfere with white balance. So you can apply the profile but have to set the white balance regardless.
When we talk about the data of the exported RAW conversion that is in sRGB or similar, even applying a white balance involves nonlinear transformations.
For example, when you use the grey eyedropper in the curves adjustment in Photoshop, you will observe that a different curve is applied to each channel.
This is one reason why it is much better to white balance the RAW file instead of the exported RAW conversion; for RAW files, the white balance is a mathematically simpler operation that does not stretch or squeeze anything.
By the way, I'm also interested in the comb-like pattern on the large and wide peaks in the histograms of the blue and the red channels. I wonder if they would suggest the posterization or lack of gradation due to the fewer pixels for these colors?
Do you mean this stuff?
(http://www.suracephoto.com/wp-content/uploads/2017/03/Fuji-copy.jpg)
This is simply an artifact of the histogram building process. Let's zoom into it a little bit and show only RAW levels 64 to 1024 (i.e. -8 stops to -4 stops relative to saturation)
(http://www.suracephoto.com/wp-content/uploads/2017/03/logzoom.jpg)
The histogram is built on a log scale. Since the levels are integers (1 through 2^14), the bins in the log domain (1/96 EV in this case), don't map to even levels. For instance, let's consider the first bin from the left that is standing out (just right of -7EV). That bin runs from 148.957 to 150.036. So the pixel counts at level 149 and 150 are counted as belonging to that bin. The bin just left of that runs from 147.885 to 148.957; so it counts only pixels that had a level of 148. So even if there are the same number of pixels each for levels 148, 149, and 150, the right one of the two bins would have double the number of counts. You can check for yourself that the count is exactly double of what it should be.
If you display the same range on a linear scale, the problem is not present:
(http://www.suracephoto.com/wp-content/uploads/2017/03/linearzoom.jpg)
Here, the bin size is 1, so all the little jaggies in the histogram are 'real'.
The comb-like patterns are purely a matter of histogram display, and have nothing to do with either the scene or the camera sensor.
-
Please remember that all the Hasselblad H-series-leaf-shutter-optics do work on the GFX and the GFX can use the leaf shutter.
Only trouble is they are all turned in to manual focus due to Fuji not wanting to pay a license fee for their adaptor it seems.
-
Simone, thank you very much for the thorough explanations. Yes, that was exactly what I meant with "comb-like pattern"!