The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

E versus non E. Splitting hairs

glenerrolrd

Workshop Member
Jack

Sorry to be a pain in the *** ..but this is important and I can t agree with your explanation as I understand it .

Just to clarify .....I completely understand that the raw conversion software is applying its own presets and calibration in the RC software.

Lets just focus on what happens in the camera ....the cameras processor takes the raw sensor data and creates the initial raw file (along with an embedded profile ) . How the camera manufacturer decides to map that data ..does have an impact on the raw file . Lets just look at the Sony sensor used in both the Sony A900 and D3X (or use any over example you want ) . We know for sure that the D800 and D800E have the same sensor. Nikon decides exactly how to map the tones and colors into the raw file and they have a different approach than Sony . It is how they do this that is the key to some of their high ISO performance . The “raw file “ represents processed sensor data and each manufacture has its own signature .

I understand the problem Guy is speaking to is really about the fact that neither C1 or Lr appear to have a D800E profile and this is the source of his “need to change “ . However I would tell you that the same exact problem occurred with the D3s and D4 ..where they have had time to get the profile correct .

This only matters because I have found (starting with the s2) that some raw files just never quite get there until the raw conversion software is refined enough to handle the new raw file ....even though you would think that you can adjust all the parameters in post processing . It also explains why no one seems to be able to match cameras from different manufacturers (if its not mapped into the raw file you aren t getting it in post ).

I realize this is beyond the discussion that Guy initiated .
 

Guy Mancuso

Administrator, Instructor
Roger I think your getting into a area also that is wrapping the instructions ( camera profile that is) into a file extension. Aka NEF, DNG, Tif and others that certain raw processors see differently. Not all programs like C1 for instance likes a DNG file, it sees it as a generic file and will dump some or all of the instructions from the manufacter and use its own engine to interpret it.
I think this is what your getting at. Regardless of the OEMs instructions the raw processor may interpret it differently or ignore those instructions.
 

Guy Mancuso

Administrator, Instructor
Example here Roger the D800 has instructions for distortion correction built into the wrapper of the file extension. C1 does not see that but LR and ACR does.
 

Steen

Senior Subscriber Member
E versus non E - Splitting hairs


So here is what I want to know. How close is the D800E to the look of the DMR?

Not an easy question to answer with any adequate accuracy.
Personally I have a feeling there will always be a rendition difference between CCD sensors and CMOS sensors but I wouldn't know exactly how to describe it.
I would rather offer you a D800E RAW file to develop to your own taste with your own favorite RAW processor so that you can try to judge the D800E CMOS sensor rendition for yourself.

Here's a link to a D800E RAW file with a really good Nikon optic (85mm f/1.4 G), a picture captured in dull light on an overcast day.
It would be interesting to see some different conversion results from different RAW converters, with links to full size of course.

801_0127_D800E_85mm_f1.4_G_at_f5.6.NEF


Here's a Lightroom 4.1 conversion - click for native size (7.4 Mb)


Nikon D800E • AF-S Nikkor 1.4/85mm G • 1/250 sec. at f/5.6 ISO 100 • Lightroom 4.1
 

glenerrolrd

Workshop Member
Guy

I have deviated from your original intent ..which was to speak to the differences in best practices using C1 for conversions of D800 and D800E files.

Given a .NEF and the adjustments you and Jack have found to be most useful ..directly implies that the D800E file has more resolution ,higher micro contrast ...and what else ? Of course if C1 had a specific D800E profile established the raw conversions might look much closer . This makes complete sense and your thoughts on normalizing the different files is precisely in line with my thinking .

You simply want good raw conversions that work for your personal and professional work .

My point was intended to take this farther and to do that , we need to understand the next level down . The camera maker creates the raw file using the sensor data as interpreted by the firmware . Its not right off the sensor its processed . This is why improvements in noise reduction often come at the expense of the mid tone s . It is also why Nikon files don t have the same color as Sony . I believe Jack indicated a different point of view ..he disagrees . Or maybe I need yet another expresso .

You are also correct in that the raw files often contain embedded profiles (like the old DMR files ) which can be used by the adobe products . C1 has always done a great job of creating their own profiles for most raw files ....they are very very good at creating great color ...not necessarily at matching those pesky color charts but most photographers are happy . But they do this on their own and I would be surprised if their NEF conversion looks even close to a Leica DNG . It never did when I was comparing C1 and LR .

Whats the point ..its that the camera manufacturers play a key role in establishing that initial raw file and the look carries thru even with great post processing .

I also believe the the camera firmware and the raw conversion software have to be tuned together ..which requires great cooperation . Take a .NEF file from your d800e ...hold your nose and just view it in NX2 . Thats how the file was intended to look . You may hate it ,think its a waste of time but it provides a reference point . I found the conversions to be subtle , with balanced color (using auto WB in the camera ) and fine detail without sharpening . Plenty of things wrong with Nx2 ..but I found it revealing in how the file should look .

Sorry for going off topic .
 

Jack

Sr. Administrator
Staff member
....the cameras processor takes the raw sensor data and creates the initial raw file
That's the incorrect assumption creating the confusion. The sensor itself creates the raw file and the camera's processor has nothing to do with it. The camera's processor may create a jpeg out of that data, and it may create sidecar files that add processing instruction sets for certain dedicated raw processors, but it does not directly alter the raw data streaming off the sensor. That's why I say the raw data is raw data period. Then the methodology by which each raw processor renders that raw data into usable image data -- their proprietary debayering and processing algorithms -- explains why one raw processor can create "better color" or "sharper images" or "a more natural looking" image all from the same raw file.
 

Paratom

Well-known member
Do you guys think that less need for sharpening would be good for higher ISO because more sharpening would probably also mean that noise would become more visible?
 

Guy Mancuso

Administrator, Instructor
This is my thinking with the E version. At least it is a assumption that less sharpening would hold noise down going into the higher ISOs. One reason I wanted to try the E version. Now there is a catch to this a AA filter also may help diffuse the noise per say with higher ISOs but again you still have to sharpen it more to get to the same level as the E. now this maybe the real splitting hairs situation. I could see a argument either way on this and make some sense. This is a test I would like to see or do. My D800 is coming back to me for a day before it goes in to repair for a new buyer and I want to try a quick test of this real quick. My gut says the E maybe better at the same level of sharpness of both cams that the noise would be less. It's a question or test that no one has done that I have seen yet.
 

glenerrolrd

Workshop Member
jack

Let me concede the point on the in camera processing because I believe its not central to the logic. The point I was trying to make (and not doing so well) was that the camera mfg do tune the sensors to create a specific rendering . (it does not matter where or how this is done). Characteristics of the raw file are established in the camera .

Sony sensors are used all over the market and have very different characteristics when used by different makers . I know Leica determines the color pallet for the output of their raw files ...I spoke to the guys that do it at lunch in Solms .

This may or may not be relevant to the discussion of the d800/d800e because we know the sensors (as reported by Nikon are identical) and we know that the files you and Guy are looking at in C1 are different . In this case the differences maybe 100% do to the differences in filtration and it appears that only one camera profile is being used in C1 .

My point is that the initial raw file as established in the camera (doesn t matter where) has a set of characteristics that define a range of possible outcomes in post processing . (yes different processing algorithms can yield distinctly different results out of post and in most cases it only matters that the photographer obtains results they like ) .

As I noted earlier ..if you are comparing the D800/d800e ...sensor discussions are probably not relevant .

The implication in these discussions is that everything is established in post processing ...which is only true in situations where the camera generated raw file is identical . This is the important aspect of my response .

Apologize for taking this off topic but you probably remember form other posts that I have been trying to develop a consistent work flow that handles multiple brands ....Leica and Nikon . I can get close in post especially if I have reference image to match but finding a repeatable solution has proven elusive . (here is an example at the US Open the wind shades on the courts are deep blue ...when I shoot Nikon and Leica they don t match ..the sensors render them differently ....so I develop a reference image and then build a preset for the Nikon to match the Leica rendering . You can’t show different shading in the same collection of images ).

Consider two other attributes often discussed ...DR and Noise . Nikon and Sony appear to use the same sensor for the A900 and D3x . From the look of the files ..Sony appears to have favored native and low ISO performance while Nikon went for a wide DR and high ISO performance . How did they get such different performance characteristics from the same sensors ? Nikon compressed the mid tones providing more head room in the highlights and pull back in the shadows . The also kick in more in camera noise suppression ..so the raw files come out different . And its almost impossible to get the Nikon files to look like the Sony files . Are they different sensors ..I guess so after the manufacturer tunes them to a spec . But the differences are established at creation in the camera ..not by the algorithms in post .

Ok I beat this one to death and from the OP this is off topic so I will quit while I am behind .
 

eleanorbrown

New member
Yes reduction in clarity and sharpening is vital to smooth out the e files. Makes all the difference in the world. On another note, I had converted some of my 800e files to DNG and C1 can't handle them... They look awful. Does Phase have any plans to handle these in the in the future? Eleanor

We need to get Phase to make a specfice profile for the E version. Seems to me a lot of folks may not figure this out. C1 is usually very good at these profiles and I'm almost proof positive it was made for the D800 and not the E version. Like many of these systems there are growing pains. Nature of the game. I'll do some more testing this week and see what else I can figure out. The nice part for the D800 is the profile is very good in regards to sharpening and all that. Try those settings Eleanor see if your seeing what I am. The clarity one is the one that really smooths it down.
 

k-hawinkler

Well-known member
Yes reduction in clarity and sharpening is vital to smooth out the e files. Makes all the difference in the world. On another note, I had converted some of my 800e files to DNG and C1 can't handle them... They look awful. Does Phase have any plans to handle these in the in the future? Eleanor

Isn't that how Adobe locks you in so that those files only work with their software? Do you still have the original NEF files? They would work. IIRC that has been extensively discussed on the LUF forum.

Regards, K-H.
 

eleanorbrown

New member
I'm not converting my NEF's to DNG anymore but i don't think this is Adobe...isn't this Phase One's refusal to open up their software to DNG?? I could be wrong, but personally I like to have all my files in DNG because I can save my changes to the file and not have to rely on a sidecar file or Lightroom catalog or C1 for that matter. Correct me if I'm wrong. Eleanor

Isn't that how Adobe locks you in so that those files only work with their software? Do you still have the original NEF files? They would work. IIRC that has been extensively discussed on the LUF forum.

Regards, K-H.
 

Shashin

Well-known member
...and not have to rely on a sidecar file...
And that is the reason I like RAW--the adjustment is never applied to the original data, it is just recorded to metadata--the sidecar file. So none of my changes to the file are destructive--I can always get back to the starting point. From what I understand, dng is no different as it uses metadata as well.
 

Uaiomex

Member
Eleanor and Guy:
Very interesting thread to me despite I don't even own one single Nikon piece of equipment. However, this post in answer to Guy's suggestion to diminish clarity and sharpening, touched me in some way to chime in.

If sharpening works by creating extra separation at pixel level, it is understandable the image will get brittle if overdone. (it does when I over do it with my Canon equipment) but if Clarity is some sort of Local Contrast enhancement as suggested by Thomas Knoll a few years ago for PS, this shouldn't produce the same effect but in some way just the opposite as the radius is gigantic.

I may be entirely wrong here but thinking about the D800E and how to avoid brittleness, it occurs (ed) to me that a way would be to back the sharpening tool and use the Clarity to the right to regain sharpening but this time not at the pixel level, adding at the same time some pop and vibrancy which btw, many Nikon D800 detractors say it's missing.

Local Enhancement is used by many dslr users to gain some pop trying to emulate a bit the medium format look.

If it doesn't work like this for the D800 or any other camera, I'd love to understand why.
Thanks
Eduardo

Guy I'm glad you have the same observations of the e files as I do... I have been working in post processing to get the crunchy feel and contrast softened. I will also try your suggestions of less clarity and sharpening. I have also been shooting long exposure water images at iso 64 instead of 100 and that smoothes the tonal transitions a bit in the water. Thanks, Eleanor
 

eleanorbrown

New member
What I understand is that I can save to the dng file but this is stored as metadata separate and does not in any way, alter the RAW data. I can "reset" the RAW any time I want and get to the original RAW file info. eleanor


And that is the reason I like RAW--the adjustment is never applied to the original data, it is just recorded to metadata--the sidecar file. So none of my changes to the file are destructive--I can always get back to the starting point. From what I understand, dng is no different as it uses metadata as well.
 

Alon

Not Available
This the London Eye RAW file as processed in C1.

The first image is C1 Standard settings.
The second image is C1 Auto settings.
The third image is C1 Manual PP.




1- C1 Standard Settings






2- C1 Auto Settings






3- C1 Manual PP
 

Guy Mancuso

Administrator, Instructor
Eleanor and Guy:
Very interesting thread to me despite I don't even own one single Nikon piece of equipment. However, this post in answer to Guy's suggestion to diminish clarity and sharpening, touched me in some way to chime in.

If sharpening works by creating extra separation at pixel level, it is understandable the image will get brittle if overdone. (it does when I over do it with my Canon equipment) but if Clarity is some sort of Local Contrast enhancement as suggested by Thomas Knoll a few years ago for PS, this shouldn't produce the same effect but in some way just the opposite as the radius is gigantic.

I may be entirely wrong here but thinking about the D800E and how to avoid brittleness, it occurs (ed) to me that a way would be to back the sharpening tool and use the Clarity to the right to regain sharpening but this time not at the pixel level, adding at the same time some pop and vibrancy which btw, many Nikon D800 detractors say it's missing.

Local Enhancement is used by many dslr users to gain some pop trying to emulate a bit the medium format look.

If it doesn't work like this for the D800 or any other camera, I'd love to understand why.
Thanks
Eduardo
The issue here is the local contrast is too high at the zero default otherwise you are correct. What it does is knock back some contrast on a global level. The D800 needs about 10 to 15 to get the snap the E just has slightly too much. I also need to play with this more too since I am going by very early outdoor contrasts images. So maybe more flatter cloud cover images will be okay at the zero setting.
 

Alon

Not Available
While the Jpg files from the D800 and the D800E do show a difference. I am not sure such difference is so acute in RAW files.

My experience with the D800 and the D800E files working on C1 is that both need extra steps and specifically in the Clarity and Sharpening areas.

To say that you have to dial back on the D800E and push up on the D800 files is not the case I experience on different files.

However, C1 Standard opening of D800 files does seem better than D800E files....but I did not test the same image shot with a D800 and a D800E. This needs to be ascertained as it is only a subjective judgement so far.
 

Jan Brittenson

Senior Subscriber Member
Sony sensors are used all over the market and have very different characteristics when used by different makers . I know Leica determines the color pallet for the output of their raw files ...I spoke to the guys that do it at lunch in Solms .
Sony or Kodak, Dalsa, or anyone else will put whatever spectral response RGB filters you want on it. It's just a matter of price and preference.

Beyond the filter, a photon is a photon is a photon. If it's let through it makes no difference whatsoever whether it's captured by a CMOS or CCD sensor.

The sensor produces an analog voltage which is then quantized. CMOS devices have internal quantization; CCDs typically employ external ADCs. The quality of the ADCs matters, they're not all equal and differ in aspects like linearity and phase. (In imaging, phase response matters as much if not more than frequency response, unlike audio where phase is relatively unimportant.)
 

Jan Brittenson

Senior Subscriber Member
That's the incorrect assumption creating the confusion. The sensor itself creates the raw file and the camera's processor has nothing to do with it.
Many raw formats are resampled or stored in Lxy format... Canon's reduced size raws for instance. Of course, one can debate whether they're really raw at all, and whether some cameras can produce truly raw files at all. Some raw files are a lot more raw than others. :)
 
Top