Author Topic: Nikon D850 previews, commentary - first reviews  (Read 13317 times)

longzoom

  • NG Supporter
  • **
  • Posts: 463
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #165 on: September 10, 2017, 15:10:24 »
Thanks God, there is no extra room for flash. So using the  shift lenses will be much more convenient! So far, step-by-step, this one looks very good. DR, let me be sure...  LZ

chambeshi

  • NG Member
  • *
  • Posts: 301
  • Woody
    • Chambeshi Photography FCotterill Photographic Explorations
Re: Nikon D850 previews, commentary - first reviews
« Reply #166 on: September 12, 2017, 20:01:43 »
Yet another review but a report of initial field tests - link to full Report [I value this guy's tests of exotic telephotos & the D500, D5 etc ]

http://www.naturalart.ca/voice/blog.html#anchor_blog_archive

11 Sept 2017: Nikon D850 - Very First Impressions...

OVERALL ONE-SENTENCE SUMMARY: The Nikon D850 is an EXCEPTIONAL high-resolution, low-ISO DSLR.
D850 Df, D500, 20 f4 AI, 45 f2.8AIP, 55mm f2.8AIS Micro, 85 f1.4D, 105 f2.5AIS, 105 f4AI Micro, 135 f2DC, 180 f2.8D, 200 f2G VRII, 300 f2.8G VR II, 24-120 f4G, 24-85 f/3.5-4.5G ED, 70-180 Micro f4-5.6D, Zeiss 15 f2.8 & 21 f2.8 Distagon, Zeiss 135 f2 APO

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 831
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #167 on: September 12, 2017, 20:44:18 »
I don't understand why he'd compare D850 image noise with the D5 at pixel level. A D5 pixel is much bigger in area than a D850 pixel and thus if the subject and application are the same, those smaller pixels occupy a smaller area and thus the impact of the noise is less. It makes more sense to compare the image noise at equal presentation size of the whole image.

When comparing D850 with D500 images there is some sense in comparing DX crops and using the actual pixels since the pixel size is approximately equal. And one can evaluate how the quality of cropped images compare. However, in normal applications one would seek to maximize quality and fill the frame with the subject close to the optimal composition and end up using only some of the resolution in a typical application.

Most sites which compare noise do not consider the per pixel noise the primary metric of interest.

I do like Brad Hill's analysis and he gives useful information. However I think the noise analysis is a bit too subjective and not really well defined methodology.

Les Olson

  • NG Member
  • *
  • Posts: 401
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #168 on: September 12, 2017, 21:23:46 »
I don't understand why he'd compare D850 image noise with the D5 at pixel level. A D5 pixel is much bigger in area than a D850 pixel and thus if the subject and application are the same, those smaller pixels occupy a smaller area and thus the impact of the noise is less. It makes more sense to compare the image noise at equal presentation size of the whole image.

No, because one would not use a D5 and a D850 for the same subject and application.  There is no reason to have all those pixels if you do nothing with them.  And what else are they for but to print much larger than you can with a D5?   

Measurements of noise are superfluous, let alone massaged measurements.  Read noise and dark noise will be (near enough to) the same in sensors of the same generation, but they will not be zero.  Smaller sensels must have smaller full well capacity and lower maximum signal to noise ratio, and at any given light intensity each D850 sensel captures fewer photons than a D5 sensel.  The photon noise is the square root of the number of photons captured, so below FWC the D850 and the D5 have the same signal to photon noise ratio for each sensel but the absolute signal is lower in the D850, and then the same absolute read and dark noise are added, so the final SNR must be lower for the D850. 

Whether that is photographically important is another question, but answering it needs a photographic argument, not one about noise metrics.

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 831
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #169 on: September 12, 2017, 22:24:33 »
If the task is not the same any comparison between performances of two products is invalid. You would not compare the fuel costs of driving a car 20km vs. flying an airplane 100km would you, and then complain how the airplane trips cost more in fuel. You'd at least make a fair comparison and divide the costs by the length of the journey. Which is why noise comparisons within the same format are better carried out after resampling to the same pixel count.

Lots of people are going to buy the D850 for other reasons than the high pixel count. It's half the price of the D5 and that may be all they can justify, this is a very good reason to buy it. Some are like myself who will buy it mainly because of the low ISO dynamic range (which is exhibited in images which are not necessarily printed large; all you need is contrasty light in the scene and the need to do local adjustments to display it well), the AF, and the flash control system, and the fact that I need several camera bodies to reach the necessary redundancy in case a piece of equipment fails and the backup should closely mimic the primary in functionality so that the necessary routine exists. But the D5 has a very different performance profile in terms of the ISO range than the D850 which is why the the two cameras would complement each other nicely. The D850's pixel count gives some extra range with the subject which can be useful when working with primes and shooting subjects that can be at different distances in bright light. If you are shooting at high ISO then there is really no extra detail to talk about since to make an acceptable print you'd need to do something to the noise which then cancels what detail you might have had if you had had enough light to give the sensor. So a high ISO comparison really isn't about pixel level noise but noise that is relevant in the final presentation (which is application, not camera determined).

A photographer who uses both D850 & D5 is likely to choose to shoot different subjects with the two cameras but a person reading an evaluation of the D850 is interested in how the same subject would be rendered by the different cameras. If the subjects are different or the magnification is different then there can be no valid comparison since it's all flown to pieces by uncontrolled parameters. If you make an A3 print from the D850 and an A4 print from the D5 then compare noise at the same viewing distance obviously it is a ridiculous comparison since the visual impact of a twice-as-large print is totally different. Normally one scales the viewing distance so that one can appreciate the whole image - which renders any comparison at different scales irrelevant. The maximum print size is not really related to pixel count. You can make practically any size print from any size original. Large prints are typically made because the viewer is some distance away. You set it up so that the image fills much of the visual field of the observer.

However, as I said if the intention is to use the D850 cropped in place of a D500 then a pixel level comparison is valid and relevant. But only because the pixels happen to be close in size.

Quote
below FWC the D850 and the D5 have the same signal to photon noise ratio for each sensel 

That's not correct. Per area, not sensel, the two have approximately the same signal to noise ratio (if we consider only photon shot noise).

armando_m

  • NG Supporter
  • **
  • Posts: 2146
  • Guadalajara México
    • http://armando-m.smugmug.com/
Re: Nikon D850 previews, commentary - first reviews
« Reply #170 on: September 12, 2017, 23:22:15 »
I do not quite agree with the  last review regarding noise

After editing a few D850 files I disagree the D800E will have a similar level of noise than the D850, a D800 iso 12000 is near useless and has ugly colors, and that is not my observation with the D850 files
Armando Morales
D800, Nikon 1 V1

Erik Lund

  • Global Moderator
  • **
  • Posts: 4129
  • Copenhagen
    • ErikLund.com
Re: Nikon D850 previews, commentary - first reviews
« Reply #171 on: September 13, 2017, 10:17:17 »
The first and the tenth shot in the ten-shot stacking sequence with smallest step size; the difference thus covers 9 steps. The (metric!) ruler is approx. 45 degrees to the lens axis.

This is not a valid af target at all!

The af target must be parallel to the sensor! Not at 45 deg.

the ruler can be in the back ground off center,,,
Erik Lund

Eddie Draaisma

  • NG Supporter
  • **
  • Posts: 189
Re: Nikon D850 previews, commentary - first reviews
« Reply #172 on: September 13, 2017, 12:03:37 »
This is not a valid af target at all!

The af target must be parallel to the sensor! Not at 45 deg.

the ruler can be in the back ground off center,,,

It is not an AF target, it is meant to show the smallest stepsize of the D850 stacking feature when used with the 105VR.

Les Olson

  • NG Member
  • *
  • Posts: 401
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #173 on: September 13, 2017, 20:09:34 »

A photographer who uses both D850 & D5 is likely to choose to shoot different subjects with the two cameras but a person reading an evaluation of the D850 is interested in how the same subject would be rendered by the different cameras.

Per area, not sensel, the two have approximately the same signal to noise ratio (if we consider only photon shot noise).

So the person reading the evaluation is not a photographer?  What you seem to be saying is that evaluation is an end in itself, quite apart from how the camera will be used.

Sorry, I was careless about the SNR.  You are quite right.  For the D850 sensel and the D5 pixel to have the same SNR they must capture the same number of photons, meaning that the D850 sensel must be exposed to more light because its area is smaller. But why should we assume that light exposure is equal?  After all, we don't care about the SNR approaching FWC, because it is very high in both cases, and at the opposite end, where the number of captured photons is zero, we can't do anything about the SNR - and it is not dependent on area. In between, why shouldn't we change the exposure to equalise SNR?  Another way to put this is that the smaller your sensels, the more you need to think about ETTR.


Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 7645
  • Bonn, Germany
Re: Nikon D850 previews, commentary - first reviews
« Reply #174 on: September 14, 2017, 05:26:22 »
Concerning ETTR in practice:

I do not see that over exposure with the D850 (which is still possible, though it should not be) is well contained. There is only a small headroom and then the area posterizes. I still have to check whether this is a problem of NX-D and highlight recovery works better in ACR or PhotoNinja

Also the test says color noise sets in earlier compared to other cameras. Yes, I see that, but I need time before I can put my finger in it.

Ilkka Nissilä

  • NG Member
  • *
  • Posts: 831
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #175 on: September 14, 2017, 15:19:22 »
So the person reading the evaluation is not a photographer?  What you seem to be saying is that evaluation is an end in itself, quite apart from how the camera will be used.

I'll try to explain it from my applications perspective. I realize that a wildlife photographer would have a different perspective.

I photographed an event in old 14th century castle ruins a few weeks ago. The organizers had had a competition for a light art installation that would be displayed on the castle outer walls and there was various musical performances inside, dancing, costumes, illustration for kids, the opportunity for kids to lay coloured LED lits floating on the water etc. Lots of activity of different kinds and they had a lot of visitors far beyond expectations. The event took place  from 7pm to midnight.

All the material I shot would be resized for printing up to A4 size and uploaded for the organizers. I didn't need the extra pixels from a higher resolution camera for this. However, to choose which camera to use for which task does to some extent depend on how the image quality is dependent on ISO and what the requirements for a particular part of the event are. For the light art show I wanted to include the lights in the hands of the visitors as well as the lit castle but there was a dynamic range problem. The light show included elements of the history of the site and the figures were moving and changing quickly so I ended up shooting at f/2.8, ISO 2000, and a shutter speed around half a second if I recall correctly. I was concerned about the story being blurred, which happened to an extent but not too much, but in reality I got a problem because of the large difference in the brightness of the light art show and the visitors.  The choice between camera would be best made with knowledge of the dynamic range at the ISO I would be using. In this case the D5 has a bit better PDR than the D810 at ISO 2000 (7.9 vs 7.1) but the PDR of the new D850 and the D5 are equal at ISO 2000.  In this case although the D850's pixel level noise at ISO 2000 is no doubt greater than that of the D5, the D850 image resized for the application likely is as good or a bit better than the image from the D5.

Similarly for the indoor part I was using ISOs from 2000 to 102400 and to make the right choice of camera the pixel level quality was not a criteria. There is no way to make details at individual pixel level from D810 images at (say) ISO 25600.  I still want to have a camera that excels in low ISO dynamic range as sometimes events occur in bright sunlight and I may want the lower ISO of 64 to be able to shoot at a wide aperture. If there are two performers on an outdoor stage one is in the front the other in the back, there can be a large difference in brightness of the two in the final image. To bring them into the same print I may need to do a local adjustment, and for this a camera like the D810 (or D850) excels.

Personal feelings about image quality can always be reported but they tend to be dependent on the observer and even differ from day to day or image to image. Another person using different software or different subject matter could come to different conclusions. I'm not saying measurements should not be validated with real world experience but in the absence of strongly differing real world evidence from multiple accounts I will trust the numbers derived from the raw data first. This is especially because they eliminate the effect of the algorithm used for raw conversion from the evaluation.

In the end, in actual use I pick the equipment to use based on my own experience. However, prior to actually owning and using the camera there is no possibility to have the benefit of that experience. Camera choice has a lot to do with other parameters than sensor image quality but I try not to be burned too badly by not being aware of what I'm getting also in the sensor characteristics. 

 

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 7645
  • Bonn, Germany
Re: Nikon D850 previews, commentary - first reviews
« Reply #176 on: September 14, 2017, 21:08:52 »
The D5 is the best high ISO Nikon, the D850 the best low an mid ISO Nikon, so a great combination. The D500 has the best AF point coverage and is great for fast action when edge or corner points are needed and for lenses that perform better in the center than the periphery. So all three cameras have their merit and if one is in the lucky position to have all three in the bag, good for her/him. For me super high ISO performance would be nice to have but does not pay my bills, so the D5 can not be justified economically. The D850 plus D500 is currently the  best combo for my purpose. Shall Nikon make a D5 Sensor in a cheaper and slower housing like a D650 or D780 or Df2 I might go for it though...

David H. Hartman

  • NG Member
  • *
  • Posts: 1798
  • I Doctor Photographs... :)
Re: Nikon D850 previews, commentary - first reviews
« Reply #177 on: September 15, 2017, 03:28:03 »
The D5 is the best high ISO Nikon, the D850 the best low an mid ISO Nikon, so a great combination. The D500 has the best AF point coverage and is great for fast action when edge or corner points are needed and for lenses that perform better in the center than the periphery. So all three cameras have their merit and if one is in the lucky position to have all three in the bag, good for her/him.

Does Charles answer prayers or only preyers?

Dave (who resents being named after an ungulate).

----------

I'd like to see Nikon drop a D5s image senor in a D850 body. Nikon could call it the D850h.
Dave Hartman The Handy Tool Maker.™

Ethan

  • NG Supporter
  • **
  • Posts: 124
  • You ARE NikonGear
Re: Nikon D850 previews, commentary - first reviews
« Reply #178 on: September 15, 2017, 09:03:51 »
Does Charles answer prayers or only preyers?

Dave (who resents being named after an ungulate).

----------

I'd like to see Nikon drop a D5s image senor in a D850 body. Nikon could call it the D850h.

1- Who is Charles?

2- Thank you for the vocabulary lessons. I learned two new words today: Preyers and Ungulate.

About the only two things I learned so far in this thread.


elsa hoffmann

  • NG Member
  • *
  • Posts: 3735
  • Cape Town, South Africa
    • Elsa Hoffmann
Re: Nikon D850 previews, commentary - first reviews
« Reply #179 on: September 15, 2017, 12:24:19 »
1- Who is Charles?

2- Thank you for the vocabulary lessons. I learned two new words today: Preyers and Ungulate.

About the only two things I learned so far in this thread.

tsk tsk.... frustration... what kind I might wonder.
"You don’t take a photograph – you make it” – Ansel Adams. Thats why I use photoshop.
www.phototourscapetown.com
www.elsa.co.za. www.intimateimages.co.za