Understanding the Nikon Df, amazing jpegs & other things

[Mark B UK]

This is mashy, posterized, and unnatural light distribution.

--
http://www.libraw.org/

I'm grateful to the OP for sharing his views about the Nikon Df and other models in the line-up. But without wanting to be overly critical, I don't see the point in arguing the relative merits of cameras all of which are of very high quality while shooting only JPEGs, then compounding the problem by subjecting them to post-processing in software designed to handle RAW files. To apply a film-era analogy, it's akin to debating the merits of Leica Ms, Nikon Fs and the like while putting out-of-date film through them and processing the images in chemicals at the wrong temperatures. Much better to get the basics right first.

Hi Mark,

The point, basically, is we all know how excellent IQ these new cameras are capable of - in RAW. There is a strong bias towards RAW shooting in these forums.

Outside the forums, even among professionals, jpeg is very often the used file format.

Add to that, jpeg out-of-the-camera has changed dramatically in the past 4-5 years. The D600 is absolutely amazing. Still, I am not sure anyone on this forum has even noticed. Have you?

For me, jpeg image quality is one of the aspects i would consider if ever buying another body. I was curious about the Df and, as it seems, the Df does not dissapoint.

RAW processing, for me, has been a short interlude between 2005 and 2012:

Before 2005, I used film.I almost never did any darkroom work. Too boring. I shot Kodachrome and Fujichrome, and projected big, colourful pictures on the screen. No post processing then.

After 2012, I stopped shooting RAW. Too much hassle, for very small gains. This way, I am sort of back to shooting slides. If you allow another analogy, shooting jpeg in 2013 would be like shooting slide film instead of developing colour prints in your darkroom.

Please note, this is one of the aspects I am trying to highlight in this thread - the others being size, handling, build quality, af performance in the dark, viewfinder, and more. I felt many aspects were not understood; and I was generally appalled at the level of both reviews and debate.

Gabriel

It's true that press photographers mainly shoot JPEGs. Their incomes depend on sending images back to the office quickly, these days mainly over 3G/4G data, and their work is published on newsprint or as small JPEGs on websites. Other types of professionals, including editorial, fashion, advertising, stck, commercial and domestic (weddings, families) mainly shoot RAW.

A JPEG is analogous to a commercial enprint made from negative film: it's a computer's interpretation of how the scene may have looked. Modern cameras' JPEG engines are better than older ones', but they are still only computers.

To make matters worse, each time a JPEG is opened, edited and saved, it is degraded in quality. Putting a JPEG through PP software is a highly destructive process.

 

[luxor2]

Interesting thread, shows how a poorly illuminated scene makes a poor photo,

despite post processing attempts. Raw imaging would have still made a dismal image

due to the technique used. Great example.

 

[HyperCypher]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

 

[Ruud Wilschut]

It's true that press photographers mainly shoot JPEGs. Their incomes depend on sending images back to the office quickly, these days mainly over 3G/4G data, and their work is published on newsprint or as small JPEGs on websites. Other types of professionals, including editorial, fashion, advertising, stck, commercial and domestic (weddings, families) mainly shoot RAW.

A JPEG is analogous to a commercial enprint made from negative film: it's a computer's interpretation of how the scene may have looked. Modern cameras' JPEG engines are better than older ones', but they are still only computers.

To make matters worse, each time a JPEG is opened, edited and saved, it is degraded in quality. Putting a JPEG through PP software is a highly destructive process.

This is true unfortunately.

This degradation is indeed a big disadvantage of shooting jpeg, but when the files does come good out the camera, and there is no need for PP, then there is really no big problem.
When jpegs are copied, moved or anything else than edited, there is no degradation yet.

It's a pity why it isn't possible to let the camera apply (at least) the same basic lens based in-camera corrections such as vignetting, distortion and CA, to NEF's in the same way as it does with jpegs. When this is also possible with NEF's it can save a lot of PP time.

Modern Nikon camera's know which lens is mounted, so why not using this data for the correction of NEF's too?

So far I know AWB setting is applied to both jpegs and NEF's, so why not also with the other settings?

Nikon should make this a setting which can be turned on or off at will to please everyone.

 

[brightcolours]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

Or just using the basic core of the D600?

 

[Iliah Borg]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

What is the vertical dimension of CAM module in D7100? Are you saying Df is created not to compete with D7100?

 

[Iliah Borg]

It's true that press photographers mainly shoot JPEGs. Their incomes depend on sending images back to the office quickly, these days mainly over 3G/4G data, and their work is published on newsprint or as small JPEGs on websites. Other types of professionals, including editorial, fashion, advertising, stck, commercial and domestic (weddings, families) mainly shoot RAW.

A JPEG is analogous to a commercial enprint made from negative film: it's a computer's interpretation of how the scene may have looked. Modern cameras' JPEG engines are better than older ones', but they are still only computers.

To make matters worse, each time a JPEG is opened, edited and saved, it is degraded in quality. Putting a JPEG through PP software is a highly destructive process.

This is true unfortunately.

This degradation is indeed a big disadvantage of shooting jpeg, but when the files does come good out the camera, and there is no need for PP, then there is really no big problem.
When jpegs are copied, moved or anything else than edited, there is no degradation yet.

It's a pity why it isn't possible to let the camera apply (at least) the same basic lens based in-camera corrections such as vignetting, distortion and CA, to NEF's in the same way as it does with jpegs. When this is also possible with NEF's it can save a lot of PP time.

Modern Nikon camera's know which lens is mounted, so why not using this data for the correction of NEF's too?

So far I know AWB setting is applied to both jpegs and NEF's, so why not also with the other settings?

Nikon should make this a setting which can be turned on or off at will to please everyone.

Pretty much nothing is applied to NEF, only some types of noise reduction and underexposure+shadow lifting to protect highlights.

 

[Horshack]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

What is the vertical dimension of CAM module in D7100? Are you saying Df is created not to compete with D7100?

I think it's more that the Df was created not to compete with Nikon's margins.

 

[j_photo]

It's a pity why it isn't possible to let the camera apply (at least) the same basic lens based in-camera corrections such as vignetting, distortion and CA, to NEF's in the same way as it does with jpegs. When this is also possible with NEF's it can save a lot of PP time.

Modern Nikon camera's know which lens is mounted, so why not using this data for the correction of NEF's too?

So far I know AWB setting is applied to both jpegs and NEF's, so why not also with the other settings?

Nikon should make this a setting which can be turned on or off at will to please everyone.

Nikon does what you are wishing for. You just need to use their post processing software.

(PS: white balance is not applied to raw files. It's just that some non-Nikon processing software is able to read the white balance setting and apply it in post.)

 

[David314]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

What is the vertical dimension of CAM module in D7100? Are you saying Df is created not to compete with D7100?

--
http://www.libraw.org/

No that is NOT what he is saying at all

what he is saying is that the op said that the multicam 3500 sensor would have made the df larger and hence nikon did not use it

from the pictures and the fact it is in the d7100 and the d300, that logic seems to be false

 

[gabriel foto]

Good morning,

Wanted to check what View NX2 can do, compared to Photoshop, if you stay away from pulling the sliders all the way and instead just try to create a punchy picture, with slightly better contrast.

Again, just pulling two sliders in Nikon View - extremely simple, anyone can do this, takes 10 seconds:

Same jpeg sooc as first picture, Shadow Protection +77, Brightness -8

Guess it can be improved with more work, and that LR5, Photoshop could do it better as well.

What do you all think about ACR? - this can be used for jpegs.

Want more examples of D600 files? Anyone who has an example from Nikon Df?

Gabriel

Technically, dark corner on the red brick (black halo) and too much blew in the grass in the lower left corner need to be taken care of. How much to open shadows in the greens is matter of taste, but as they are now they attract much attention robbing it from the main perspective.

I do not use ACR or LR in my work. Simply do not need them. But I use high quality gradient filters when shooting.

I have my own Df cameras and I tried them in JPEG mode. My experience is the same as it used to be - for even mildly high dynamic range scenes shooting JPEGs results in more time spent in postprocessing and with a subpar result compared to 14-bit lossless NEFs.

--
http://www.libraw.org/

Iliah Borg wrote:

Now that you have tweaked a jpeg, lifting shadows and found 'nothing funny going there, no additional artifacts' - do you have any thoughts?

Same, those artifacts that are already baked at jpeg would need a lot of touch-up and cloning work before I would be printing the image. All I said is that processing did not add artifacts - just brought to visibility what is already there. Not a good thing IMHO. When I said 90 secs it was from opening the file to saving it. Touch-up took me a little more than 40 minutes, and I work in Photoshop since 2.5 (1992), and worked as a scanner operator retouching hundreds of scans daily.

--

http://www.libraw.org/

Iliah,

Thank you for your taking your time to respond!

I did have a feeling that you actually had a wealth of knowledge about sensor output, processing and more so I had to check... also checked your website - impressive.

So, hope you don't mind, but I felt this is an opportunity to ask more - I would be immensely grateful if you are interested in trying to straighten a couple of more things:

First, please allow me to explain that image quality matters a lot to me - but I am not prepared to spend much effort to achieve the last touch of scientific perfection in my photography. I am happy shooting jpeg. (This is because of me, not because there would be anything wrong with the solid and indisputable arguments for shooting RAW!)

1) So, if the starting point is that I want to shoot jpg with, say, my D600, how much dynamic range (DR) do I lose?

The RAW DR at base ISO for the D610 is 14.4 stops, according to DxO.

The jpeg DR at base ISO for the D600, is something like 11.9 stops, according to PpR's graph.

using Active D-Lighting (ADL) set to extra high. (This is the setting I would use in a high-contrast scene in order to capture as great range as possible)

By opting for jpeg instead of RAW, is this the actual difference? Am I losing approx. 2.5 stops?

2) Next, if I am prepared to accept very slight artefacts (invisible to the untrained eye, perhaps ;-)), can I make use of the dynamic range in a jpeg much the same as I can in a RAW? Or will I have to discard some of the information in the jpeg not to create posterization

3) In such case, is it reasonable to compare jpeg DR from one camera with RAW from another?

I have been interested in trying to understand this. The best Canon sensor in this class is the 6D.

The RAW DR at base ISO for the 6D is 12.1 stops, according to DxO

The RAW DR at base ISO for the Df is 13.1 stops, according to DxO. Screen print from DxO:

Again, this the RAW. Base ISO is to the left; the points at approx 12, 13 and 14 Ev.

The base performance in jpeg would look like this, according to DpR. They are presented in the shape of tone curves, and below that the limit measurements:

In order not to compare too much of apples and pears, I chose the ADL setting for the Df which gives a tone curve resembling the D600 Extra High - for Df, it is called Extra High 1.

For the Canon 6D, I chose the setting which offers the greatest dynamic range. You can see that the 6D's tone curve has a different look - I guess this is what Canon shooters like about their jpgs, it creates a little more punch straight out of the camera.

So, my questions lead up to this (as you may have guessed):

The dynamic range of a Nikon D600 jpeg image seems more or less on par with a Canon 6D RAW image. 12 stops, more or less.

Good to see that the Nikon Df sits in between, at base ISO. It does seem better at RAW than jpeg though. Then of course it wins at high ISOs, we know that.

I don't know if it is correct to claim such a thing. If it is, it would be an interesting way to classify cameras.

I am sure that the D600 / D610 and the D800 jpegs, in particular, have a better dynamic range than the raw images from a number of older cameras.

Gabriel

 

[Iliah Borg]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

What is the vertical dimension of CAM module in D7100? Are you saying Df is created not to compete with D7100?

 

[brightcolours]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

What is the vertical dimension of CAM module in D7100? Are you saying Df is created not to compete with D7100?

--
http://www.libraw.org/

No that is NOT what he is saying at all

Thank you

from the pictures and the fact it is in the d7100 and the d300, that logic seems to be false

CAM DX and CAM FX do have the same size?

Yes. Same mirror box dimensions (height wise), same AF sensor size, same AF point spread over the same area, same optics for the AF sensor.

--
http://www.libraw.org/

 

[Iliah Borg]

1) So, if the starting point is that I want to shoot jpg with, say, my D600, how much dynamic range (DR) do I lose?

The RAW DR at base ISO for the D610 is 14.4 stops, according to DxO.

The jpeg DR at base ISO for the D600, is something like 11.9 stops, according to PpR's graph.

No, you can't use those numbers directly.

Do you have a printer? If you do, print an A4 page with vey dark grey L=25 on dark grey background L=35, 12 points text, serif. Shoot raw + jpeg, progressively underexposed. See how much underexposure each tolerates. Try at different ISO settings.

 

[Iliah Borg]

CAM DX and CAM FX do have the same size?

Yes. Same mirror box dimensions (height wise), same AF sensor size, same AF point spread over the same area, same optics for the AF sensor.

Dimensions, please.

 

[gabriel foto]

my only point is that the "size" of the 51 pt module (what ever it is) doesn't preclude it's use in a "compact" body, but just given the fact that that module IS in the 7100, one can only presume it's lack of use in the Df has little to do with cost of actual physical size, that leaves product segmentation, no ?

What is the vertical dimension of CAM module in D7100? Are you saying Df is created not to compete with D7100?

--
http://www.libraw.org/

No that is NOT what he is saying at all

what he is saying is that the op said that the multicam 3500 sensor would have made the df larger and hence nikon did not use it

from the pictures and the fact it is in the d7100 and the d300, that logic seems to be false

Hi David (and others),

In fact Iliah has understood me well, but I have failed at explaining this very clearly.

I made two observations:

First, the 51-point CAM module, as seen in pictures, is higher than the 39-point module (apart from being wider, of course)

Second, there is a little more space under the floor of the mirror housing and the bottom plate in the D300 / D7100, on one hand, and the D600 / D610 / Df on the other hand.

There are obviously other things packed in this part of the camera, but the outside difference is about 5mm - not more, but clearly visible:

Cropped picture from Camera Size. D7100 to the left, Df to the right.

My conclusion is this:

* We have seen the same decisions in the Canon 6D vs the 5D MkIII - a smaller CAM module in a smaller body. This is not unique for Nikon. Incidentally, perhaps you could say that the 6D's module is even more limited in terms of outer focus points.

* I believe both Canon and Nikon have concluded that the AF performance of these cameras is entirely satisfactory, if the intended customer is not a sports shooter.

* I do not think that Nikon wanted to save 40 or 60 dollars (no, I don't know the actual cost difference) by specifying the 4800FX CAM module in the Df. Believe what you like, but Nikon is not stupid. If they can offer a higher-specified camera for a similar price, they will of course do so. Yes, I am very familiar with the term cannibalizing.

* I beleive they were strictly focused on making the most compact DSLR in the world - something that DpR, ironically, failed to acknowledge.

To one person asking: No, these are my conclusions only. I have not taken these cameras apart.

* In some respects, in this particular review, I had a strong impression that DpR chose only to see the back side of the coin. That bothered me, and that is one of the reason I opened this thread.

Gabriel

 

[David314]

CAM DX and CAM FX do have the same size?

Yes. Same mirror box dimensions (height wise), same AF sensor size, same AF point spread over the same area, same optics for the AF sensor.

Dimensions, please.

--
http://www.libraw.org/

 

[Ruud Wilschut]

It's a pity why it isn't possible to let the camera apply (at least) the same basic lens based in-camera corrections such as vignetting, distortion and CA, to NEF's in the same way as it does with jpegs. When this is also possible with NEF's it can save a lot of PP time.

Modern Nikon camera's know which lens is mounted, so why not using this data for the correction of NEF's too?

So far I know AWB setting is applied to both jpegs and NEF's, so why not also with the other settings?

Nikon should make this a setting which can be turned on or off at will to please everyone.

Nikon does what you are wishing for. You just need to use their post processing software.

(PS: white balance is not applied to raw files. It's just that some non-Nikon processing software is able to read the white balance setting and apply it in post.)

Oke, but you still have to do PP with specified software (I assume you'r referring to Nikon Capture NX). Or do you mean that the software reads al the data from the camera embedded in the NEF's? If so, why doesn't Nikon make it possible applying this in-camera if the data is already there?

 

[Iliah Borg]

D7100

--

About LibRaw | LibRaw

www.libraw.org

 

[Iliah Borg]

why doesn't Nikon make it possible applying this in-camera if the data is already there?

Because the photographer or pre-press may decide differently. Because forgetting to switch to the right white balance will ruin a nef like it was on D1 original.