Home › Nikon DSLR cameras › Df
Custom Search
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Categories
- 3.4K All Categories
- 34 Nikon Z lenses
- 77 Nikon Z cameras
- 1.4K Nikon DSLR cameras
- 11 Df
- 237 D6x0/D7x0/D8x0
- 27 D300/D300s/D500
- 29 D3x00
- 45 D3/D4/D5
- 31 D5x00
- 94 D90/D7x00
- 387 Nikon Lenses
- 9 Videos, Nikon et. al.
- 52 Nikon 1
- 18 Nikon Film Cameras
- 56 Nikon Flashes
- 25 Nikon Compact Cameras
- 48 Nikon Software
- 26 Nikon Deals
- 767 General Discussions
- 49 Gear Reviews
- 182 Other Manufacturers
- 105 Fun & Weird
- 128 Photo-a-day
In this Discussion
Who's Online (0)
Disclaimer: This site has no affiliation with Nikon Corporation or any of its subsidiary.
Disclosure: [NR] is sponsored by displaying advertising banners and links.
All trademarks and brands belong to their respective owners.
Want to advertise on NikonRumors.com? Contact [NR].
Copyright © 2008-2017 NikonRumors.com.
See our Privacy Policy.
Disclosure: [NR] is sponsored by displaying advertising banners and links.
All trademarks and brands belong to their respective owners.
Want to advertise on NikonRumors.com? Contact [NR].
Copyright © 2008-2017 NikonRumors.com.
See our Privacy Policy.
Comments
Our posts in this thread is to compare the different models to show it's no better than D600/610 functionality wise so that that people browsing can make an informed decision. That's all.
It had been shown the graphs TTJ provided are not valid for sensor level comparisons. One can not use the pixel level data to compare, when the sensors have different MP/densities. Ade explained why normalized data should be used for comparison instead, and supplied the normalized graph which showed the 16MP is no better at 6400 and below. That's why all 3 Nikon FX sensors are a wash in term of low light performance.
If you have further technical questions, feel free to ask.
http://www.bythom.com/nikond4review.htm
The section "D800 versus D4" may be of particular interest. In summary, Thom's empirical testing agrees with our theoretical discussion remarkably well.
The main advantage of the D4 sensor is speed. However, in the Df, Nikon has taken away this advantage.
Technically, the D800 is very close to the equivalent of the D4 at the most useful higher ISO values (e.g. ISO 1600 through 6400). That's assuming, of course, that you downsize the D800's results to 16mp, and do so with care
Why would I want to do that? Actually, how do I do that? Image size only effects TIFFs and JPegs. I shoot Raw 14 bit lossless compressed plus JPEG Large Fine (might switch to JPEG Basic Small and save hard drive space, not sure I need the finer given my Raws). I suppose that your answer is post-processing in Lightroom/PS and when I become motivated enough, I will explore that.
So Ade and TC88, it seems that you are saying that I have to through away about 20 megapixels to get D4 lowlight performance on my D800 and I that sounds reasonable if that is your point. But I am not in the habit of throwing away megapixels, so TTG's argument and DXo's non-normalized results seem more relevant to me, which indicates that I won't get the D4's lowlight performance on my D800. I am still happy with my D800's lowlight performance however.
It seems like the argument is comparing apples to oranges..........
Your point also seems to suggest that I can improve noise by "throwing away 20 megapixels" in Lightroom/PS somehow? Am I understanding you correctly? If so, I might motivate myself to explore that further.
And all this suggests that the Df really will have superior lowlight performance to my D800 unless I go through the trouble of downsizing my D800's results to 16mp (and MsMoto, this is why this is all related to the Df body, sorry about going back on my word that I was finished above).
@jshickele - Neither Ade or Tc88 are living in the real world so be careful falling them down the rabbit hole. Ade just enjoys attacking whatever I post and TC88 seems to enjoy jumping on the pile this week. This has been going on for the better of 6 months and they enjoy the good ole' "life in a vacuum" arguments.
Both are leaving out on glaring attribute to their argument - for their statements to be true you have to be printing at 8x10 or lower with zero cropping, and zero editing what so ever. Also all the "tests" do not actually measure anything but just colors (saturation, light, hue, contrast) with various algorithms. The tests to not indicate what so ever if you can make out individual hairs, pores, sharp details, or concentrations of noise, or how that all combines for the actual image quality. It also doesn't take into account at all the ability to edit files and where the thresholds are on those. All of that equals the output quality of the file. Some call that subjective to judge an image by how we look at it and base our opinion of quality by knowing this color is an eyelash, and this one is a blimpish seems rather unequivocal.
I hope MsMoto doesn't mind but she has some great examples of what the D4 sensor can accomplish.
D4 • Manual Exposure • 1/200th • f/4.5 • ISO 12,800 •
http://www.flickr.com/photos/fantinesfotos/8709443733/in/set-72157633412302667#
(Link is not working - I think she has the sharing turned off - there is the image address above)
No downsizing makes my D800 look that good - there are details that the D4 retains that would be turned to mush from my D800.
donaldejose, I'm willing to give you that the 16MP can perform better at 12800. However, that does not automatically translate into advantages at any other ISO. I may even give you a slight leeway at 6400. But at 3200 and below, the D800/D600 sensors should do as good if not better. And considering the 39pt used in Df, I'm not sure that it can focus reliably at 6400, it's a moot point to argue any sensor advantage of Df at 12800.
TTJ, I understand not everyone has a scientific background. However, the problem I see is that you talk about the things you don't know as if you know everything. What you are doing is called pick and choose, hand waving arguments. Your arguments about price/build/AF have all been proven false, so you are hanging last hope on sensor advantage through all those subjective arguments.
Thank god photography is cheaper than that hobby was.
You can't see the forest for the trees sir, not at all.
Really funny thing is, that the base of all your arguments is that the D4 sensor is inherently better. Make sure you don't think too hard on that one, I would hate to hear your head exploded.
--------------------------------
I am finding it funny that no one is talking/calling for a Sony a7/r comparison to see how they will match up. Objectively that is what I want to see. Depending on what someone may already own, they are a viable option for cheaper.
--------------------------------
As for the rest of the "arguments", sorry but I am correct - I'm choosing not to call some morons. For what photographers it is designed for (and what they wanted), everything is just as good as anything else. There are bunch of people acting like babies stomping their feet and just can't accept the reality of the world we live in. It is a consumer camera (had for $2,000) with a PRO sensor (had for $6,000) in it, and it is cheaper than the other pro camera ($3,000). A basic average would put this at $4,000 - but Nikon said the premium sensor was only 12% of the D4 price and a premium price due to lower production, and added that to the D610 price. It doesn't matter that silicone wafers are made from sand, it doesn't matter if the same parts are used in other bodies, it doesn't matter if what your opinion is on any of those things are. It is what it is.
Some on here want to prove how wrong Nikon is, how wrong this camera is, how wrong the marketing it, because they can't afford it. That is childish - and most are just waiting for you to cry yourselves to sleep so the adults can have a real discussion on the system.
It is an amazing system - one that most on this thread have been wanting for years. And yes I have been searching and pulling people's past comments, and many are are those dogging this camera or not wanting it - only for price. Why do I know that, because this camera is exactly what people said they would buy in a heartbeat. Retro body, D3s/D4 sensor, save some $$ by using the D7000/D90 AF and Meter. If it is a smaller body, drop video since it may overheat the the chips. Basic weather sealing and build quality of the D7000/D90. The prices people were willing to pay? $1,500 - $2,500. So Nikon's $2,700 is just outrageous! Ahhhhhh! The world is going to end!!! It was $200 more than people's top price!!!!!!! How could a company that relies on profits do something like that!!!!!
The Df isn't garbage or is Nikon headed to an ill fated existence just because you can't or choose not to afford it.
I believe that is called "champagne taste on a beer budget."
Allow me to expand-
What I have noticed is that more pixels have noise (or in better terms, incomplete values/not reading enough or any light), and that is in part due to their size. So in the average of 3,4,6,9 pixels combined where 1 out of 3, 3 out of 5, etc didn't read a value, the resulting section has more incomplete data to create the correct color. The D4's pixels are larger and record "a value" or "more values" for the average so the calculations have a better set of data to create an image. The argument is, that down sampling 4.5 times the D800 image, that result creates the same image.
The issue with the assumption is that the value of a 3 pixel average that is a direct result of the view-able image the D4 took, (where each pixel has a better chance to record a value) will equal the D800's image where first recording of the image (where many more pixels are not read) is outputted and then is recombined again by the computer which is a second average of the same image. Basically it is like taking a photo of the existing photo the D800 already took, and assuming the results, (details, color, etc.) are the same as the D4 image without this middle step. The belief, hopes, & some people's assumptions is that the extra processing doesn't change the re-produced image. When in reality, we know they do.
This is where it get's interesting.
DxO's Screen values vs Print (so called normalized) show "change" that shouldn't be ignored.
All results of the Print show better Noise handling, yes, but also,
Better dynamic range, Better Tonal range, & Better Color Sensitivity.
The file improved? WTF? No, The file changed.
The output of the sensor is the output of the sensor. That doesn't change.
But the Print values became better. Why?
Easy, the computer adds values during it's re-processing to "create a better image." And 99% of the time it does But we are not evaluating the computer, just the camera. The Print image readings, is not only the success of the camera, but the success of the camera AND computer.
That is why the Print values are total BS.
I'll illustrate:
The SNR 18% @6400 (higher dB indicates better noise handling)
---Camera-----dB Screen----dB Print
- D3s ------26.2 ------28
- D4 ------24.7 ------27.9
- D800 ------21.4 ------28
Noise diminishes - hoorah! Makes since, more averages eliminates noise.
The Dynamic Range @6400 (higher Ev indicates better noise handling)
---Camera-----Ev Screen----Ev Print
- D3s ------9.25 ------9.56
- D4 ------9.21 ------9.73
- D800 ------8.07 ------9.1
That's correct, the D800 GAINS 1 stop of DR. The D3s gains 1/3 a stop, & the D4 gains 1/2 a stop of DR. Remember, the Screen is direct from the camera to sensor, no processing. After processing (the "normalization" process) each image gains color. (And also corresponds with MP I might add.)
Begs the question - "What & Where, was color added?"
This is the same with Tonal range and Color Sensitivity.
That means when the "Normalization" happens, COLOR IS ADDED = IMAGE IS CHANGED = LOSS OF ORIGINAL IMAGE as colors (which result in details) are replaced. THAT IS NOT "NORMALIZING" DATA that is changing it.
That is why Screen is a better evaluation - because it does not have the added computer processing where data is added.
Because the D4's pixels records more actual values, less is guessed by the processor, & downstream through computer processing. With the D800, we can see more noise, which equals to more guessing by the processor & subsequently more processing downstream. If you say that "downsizing" the image creates the same noise output, that is incorrect. It is like a wall getting a second coat of paint. Sure it covers more "gaps" but it also covers up more texture of the wall.
What is happening is that the computer knows cobalt blue or fire engine red, surrounded by black should probably be black, so the color is replaced or that color is removed when downsizing. That is great, until it happens on the defining edge of details. Like the separation of hair. Now it is really guessing, what to remove and leave. It is also trying to smooth jagged edges (My guess is this is where most of the color data is added to avoid posterization) out which also softens details as it does it across the whole image. (Why we utilize masks in Photoshop.)
It is simple, Less processing = better quality of an image = D4.
Everyone: As to the question of whether or not the D4 sensor is superior to the D600/D800 sensors at high ISO I would ask "What does Nikon itself think?" Nikon must have some criteria for acceptable IQ and they label their ISO settings with ISO numbers for the range they feel is acceptable. Beyond that range Nikon uses letters/numbers such as H0.3. The ISO numeric range ends with 6400 for the D600/D800 sensor but ends at 12,800 for the D4 sensor giving the D4 sensor a one stop IQ advantage above the D600/D800 sensors. From the photos I have seen I would agree. If it is one stop better at ISO 12800 wouldn't you think logically the IQ will also be noticeably better at ISO 6400 and at ISO 3200?
(**) the labels are also related to marketing. Back in reality, image processing software work quite differently from what is being described above.
Typically, there are two algorithms in common use for image downsizing, known as bi-linear and bi-cubic.
Bilinear. Bilinear is simple (linear) averaging in two-dimensions (x and y), using the 4 closest pixels (2x2). It is called bilinear because the algorithm takes into account of the pixel distances (positions) using two linear interpolation steps. Using lines for averaging allows for downsizing on non-multiples (such as from 3 to 2).
Bicubic. In bicubic, instead of using lines to average between 4 points, the algorithm uses curves ("splines") to smoothly interpolate between any number of points. Think of bicubic as a "weighed average" instead of the simple average used in bilinear. For practical reasons, most bicubic implementations will interpolate from 16 of the closest pixels (4x4). Bicubic takes much longer to process but can produce smoother, more accurate results.
Other algorithms exist (e.g., nearest neighbor, fractal) but they are not typically used for downsizing.
Whats more consider the DR numbers involved, by ISO 1600 its around 10.5 vs 11.5, that's a difference your going to see without pushing shadows hard so seems well suited for a street/social shooting style.
I think this shows it pretty well...
http://photographylife.com/nikon-df-vs-d610
The difference in the well exposed parts of the shot at minimal but go into the shadows(the piece of wood behind the ship here) and the Df(well D4) sensor clearly has less noise at higher ISO's.
Here's a direct quotation from the article you linked:
"Now keep in mind once again that we are purely looking at pixel level performance here. Once 24 MP images from the D610 are resized to 16 MP, the performance differences shrink more. Only at ISO 12800 and above the Nikon Df seems to offer an advantage when down-sampled."
So even the author of your article agrees with normalization.
As to where the advantage lies that's in the eye of the beholder, to me ISO 12,800 is the point where the D610 image becomes worthless not where the difference is first noticble.
A few posts ago, you claimed cropping with have affect image quality differently on D800/D4. If you take the same composition and crop the same amount, they should create the same image with same quality left.
Now you claimed that increased DR, tone range after noise reduction means color has shifted. I don't know when you get some connection. All it means is that the color like highest red stays the same, but the weakest red that was overwhelmed by noise before is now detectable. So the range is increased. That's well understood that with lower noise you can pull out more from shaddows if needed. Somehow your claimed better than us scientific background didn't allow you to comprehend.
There is a saying that the most dangerous people are those who don't know anything and think they know everything. To be honest, I think you are approaching that. Somehow, in your mind, people like Dxomark, Tom Hagan, Ade, Pitchblack and me who pointed out the theoretical explanations are all wrong at the same time and you are the only one correct. Figureatively, your posts are baring yourself to show how much knowledge you only have, just like the emperor without clothes.
I probably won't be able to respond to all your mis-cited refereneces and long rumblings, some of which I can't even summarize what the points are. The following was my first response to your postings before. I'm pretty confident that most people reading this thread will agree with me. Msmoto, I think it may be time to lock this thread. There are enough information presented, correct and incorrect, that people should be able to make their own informed decisions. I'm not sure there is much to be added anymore.
As I said to you earlier, it may be a moot point that you may not even be able to AF using the 39pt at 12800.
Threads get closed too quickly around here as it is. Please don't encourage it.
Being a photographer is a lot like being a Christian: Some people look at you funny but do not see the amazing beauty all around them - heartyfisher.
My, this is getting personal. Remember that we are arguing about nuances that only us 1 per enters would notice or care about. I would love to start a thread one how the 1 percent's benefits trickle down to the 99 percent.
I am in a business role where I have a decent technical knowledge and deal with people that have a much greater technical knowledge than I have. I don't try to match them. I channel my educational efforts into enhancing my "smell detector" so that I can more efficiently administer the "smell test".
I am in the same situation in this forum and am administering the smell test.
On balance, TTJ's arguments smell less to me than Ade's and tc88's. Because the arguments are personal, there is smell all around. I always remember that an argument should be evaluated based on debate about the merits of the argument, not the merits of the debaters.
What is the basis of the conclusion about my smell test? There are a few points?
One is the unspoken assumption that Nikon are evil capitalistic liars dominated by there marketing department. I reject that. Do they make mistakes? Sure! Do they spin? Sure! But overall I think they are competent, even brilliant, and truthfully.
Second, and this is connected with the first when people think that Nikon is spouting marketing bs to justify charging double for the D4. If it is just marketing bs, there must be a way to get something for nothing. D4 for D800 price, D800 for D600 price, Df for D600 price. "If I can only prove that the low light ability of the Df is not any better than the D800, then I can justify why it is only worth the price of a D610." In my experience, if you want more, you usually have to pay for it. Sorry Ade's and tc88. While I doubt that TTj is the brilliant engineer that none of us are, his argument smells right and yours don't because you are implying that something can be had for nothing. If that was true, why wouldn't Nikon just do it and do it first before Canon or Sony.
I love my D800, but I am not going to kid myself that I will get the same low light performance as a D4 or Df. Maybe some day I will buy the Df and use a short fast lens for street photography and my manual focus in m-mode when I want to live "pure photography".
I love the Burberry Df. It would perfectly coordinate with the coat I am wearing.
:-))
@Tc88 - try re-reading what I wrote and comment on that, attacking me doesn't help anyone else. So far your arguments consist of "because I, Tc88, said so" and have not proved anything but that you assume more than you practice or test. Now that is funny!
tc88: you say the D4 sensor and the D800 sensor are a "wash." I would like to hear from someone who has both sensors. My expectation would be that the D800 would clearly be superior in IQ (especially detail rendering) at lower ISOs and the D4 would clearly be superior in IQ (especially dynamic range) at higher ISOs. But, as I said before, I don't have both sensors so I cannot state this from personal knowledge, just from my viewing of images I have seen posted.