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!

D800 v D800E, chapter 6

johnnygoesdigital

New member
Are you talking the crops Rob posted where he processed identically, or are you talking crops that you processed to each sensor's ideal? I assume the former, because I think if you actually take the time to do the latter, you won't see the differences like in the former...

:chug:
Jack, you're right! Thanks.
 

tashley

Subscriber Member
I've thought long and hard about chiming in here (apart from the little joke further up the thread) but having mulled this over in conjunction with the thread on Testing Protocols, here goes with my 2 cents.

I've got both cameras as many of you know. I've done and posted here one test, which was itself susceptible to all the methodological gripes one would expect. The light changed a little, the point of focus could be quibbled (though I don't think that was a relevant factor in my test) and so on and so on.

The problem with a 2D target for this sort of test is that it doesn't reveal whether there is any focus difference between the two setups. The problem with a 3D target is that it does, and therefore no one believes the results unless they want to :D

The first problem inherent in the test is, therefore, focus. Vibrations and motion blur can be exed out with MUPs and tripods and remotes but focus will always be the point of dispute. However carefully you LV the focus and whatever DOF you allow, someone will always perceive a difference in POF.

That person drops out of the 'belief group' at that moment.

The remaining consumers of the test might find a host of other issues to question: changing light (and therefore exposure, WB and contrast), issues of the same lens having different diffraction characteristics on different sensor setups, etc etc.

So let's look at the steps.

Step One: Take the shots, see how many people agree that they are a fair comparison. Lose some believers. For the rest who believe we have two files taken with as close to identical light as possible, of a complex, moderately distant target, with exactly the same POF, we move on to...

Step Two: Now we have PP.

Identical PP doesn't let you achieve, for each setup, the best it might achieve.

Ideal PP (the best for each file) instantly un-levels the playing field and opens up the issues of personal preference, personal bias, quid-pro-quo of trade-off decisions etc.

In the case of this thread, my problem with sharpening the D800 file to match the apparent resolution of the D800E file is that sharpening has costs. Artefacts. Enhanced grain, more or less evident at different ISOs and in highlights and shadows. Changes to tonality gradations. Clarity adjustments do the same thing. So now we lose more of the belief group; me for example. Moving on...

Step Three: printing. Everyone has their own secret juice. Each recipe has the potential to favour one kind of a processed file over another. The belief group shrinks further.

Step Four: the perception of the print. There was a well-known article at Lula a while back where Mr Reichmann showed a print made on a Canon P&S to a batch of other experts and pros and some of them mistook it for a MF file. This was, he thought, due to the enhanced appearance of clarity in the file caused by the fact that the enlargement was not too big, the sensor was small and therefore had great DOF, and the ISO was low enough not to show up the noise limitations of the sensor at that print size.

So now we've got to the print (and most of us will never even see each others' prints so we have to take each others' word for it) and the belief group is probably down to... about one. The tester.

For me, therefore, and this is very personal and not meant as an attempt at changing anyone else's minds or methods, the only useful resolution test is to get to the end of Step One.

If at the end of Step One, we have files that the great majority of the consumers of the test believe are perfectly comparable, then that's the place to compare them IMHO. In this test, for example, many (most?) people appear to accept that there are file sets out there that allow for a reasonable comparison of the resolution of the D800 and D800E sensors. And the majority of those tests show that the D800E has better resolution, as you would expect.

It is true that the difference in resolution is less than one expected from, for example, working with a 5DII and a Leica M9. As Jack has pointed out, Nikon hit the ball out of the park here with the design of the AA filter. And the technique of 'cancelling' the filter is probably less effective than removing it. But regardless, there is more resolution in the E files.

Whether or not that extra resolution is enough to make a difference in print is, for me, entirely moot, because there are too many extra steps for any group of people to agree on. The most you can say is that if you regularly need to show 100% crops on screen, and don't mind moire, then the D800E is a better choice. Beyond that it gets personal, and depends on the rest of the recipe used and how much one is influenced by the placebo effect :D .... I know I am!

So, to the personal, and this is where I can share an opinion that I think might be useful to others but YMMV.

I have shot well over a thousand frames, about 1/3rd to 2/3rds D800 to D800E and I have only shot ONE direct comparison. So my opinion is purely intuitive, fuzzy logic, but my opinion is that the D800E files are the ones I would rather work with. Significantly more often with it, I get that 'MF wow' factor that I get from Phase files and from many M9 files. The instant bang of clarity. Clean, subtle, real. Convincing.

I sometimes get that from D800 files but as you can see from the above ratios, the one I am reaching for is the D800E.

I will never be able to prove this... it's personal, subjective, co-factored, statistically and methodologically risible. But in the end we all have to reach for one camera and not another...
 

Amin

Active member
ended up setting the E at 120/0.6/2.0 which appears to be right at the edge of being oversharp. I then adjusted settings on the regular file to a similar look, which took sharpening up to 240/0.6/2.0 and required a clarity bump to 25 to get teh micro contrast similar
Jack, thanks for the excellent analysis. Depending on the subject matter, lighting conditions, and ISO settings, I'm guessing that the extra sharpening needed for the regular version will emphasize any noise in the file relative to the E file since the AA filter blur detail without affecting noise. Either way, it's probably going to be an on-screen at 100% phenomenon.
 

D&A

Well-known member
Tim, your well thought out beliefs regarding both the methdology and ultimate pitfalls of trying to demonstrate the differences in image output from these two fine cameras (D800/D800E) is in itself open to personal interetation and would probably make for a lively and lengthy discussion...this aside from debating the actual differences each one of us "sees" in such test files pairs.

Personally I am still of the strong opinion that the glass sandwitch technique Nikon employed to produce the cancelling effect of the AA filter, somehow mitigates to a degree the increase in sharpness we would normally associate (see) if the AA filter was simply removed altogther from a regular D800 body and replaced with s simple single layer protective cover glass. I am even thinking that possibly the glass sandwitch Nikon employed might even be "tunable", much the way a the two glass surfaces in a polarizer filter can be rotated to increase or decrease the strength of polarizing the incoming rays of light.

This way, Nikon found a compromise orientation of the cancelling effect that upon close examination, did improve sharpness slightly, but at the same time, lessened to a degree the potential for moire'. If my assumption is true, they could orientate the cancelling glass of the D800E to produce greater sharpness from the D800E files relative to the regular D800, but the potential for increased incidence of moire" might have had them extremely worried about lashback from those that weren't prepared for dealing with it to such an extreme (or at least seemed extreme to those who hadn't had previosu experience with using AA filterless DSLRs.). If this is partially true (and it's a big assumption on my part at this time that it is), that does open the door for future DSLR models whereby the tuning of the cancelling filter could be orientation in such a way, to have a much stonger cancelling effect, resulting in greater differences seen vs. the AA filter model of the same camera.

Dave (D&A)
 
Last edited:

ustein

Contributing Editor
>So now we've got to the print

If I think of that I feel that I would be bankrupt (paper and ink) to have a conclusive finding.

>Nikon found a compromise orientation of the cancelling effect that upon close examination

I think this is not a camera with just "no" AA filter. More like a camera with a lower AA filter effect.

Conclusion for me:

Would I buy now I get the D800E. But having the D800 lets me concentrate on more important stuff like what to photograph. Lots to work on. Also good that I don't have a MF back and cannot compare. I hardly ever use the term "what a good file" but use "good image". It can come from any camera I use. But I love my 36MP :).
 

D&A

Well-known member
I think this is not a camera with just "no" AA filter. More like a camera with a lower AA filter effect.
Exactly Uwe! That was precisely my feeling as illustrated in my post above, when I described what I thought might be behind Nikon's implementation of this (their) unique AA cancelling filter in the D800E (besides standardizing production assembly of both cameras). The net result is sort of a compromise of partial removal of AA filtering and along with this, the modulation of possible incidence of moire'. Some proof of this will be when someone has aftermarket removal of the AA filter on their regular D800 and image output is compared to both a regular production D800 and D800E.

Dave (D&A)
 
I think this is not a camera with just "no" AA filter. More like a camera with a lower AA filter effect.
Maybe. Whether we'd ever be able to see a difference between the 800e and a hypothetical filterless version remains a matter of pure conjecture.
 

Jack

Sr. Administrator
Staff member
Ideal PP (the best for each file) instantly un-levels the playing field and opens up the issues of personal preference, personal bias, quid-pro-quo of trade-off decisions etc. In the case of this thread, my problem with sharpening the D800 file to match the apparent resolution of the D800E file is that sharpening has costs. Artefacts. Enhanced grain, more or less evident at different ISOs and in highlights and shadows. Changes to tonality gradations. Clarity adjustments do the same thing. So now we lose more of the belief group; me for example.
Tim,

Normally I'd agree with you -- but in the case of the files above, my sharpening amounts as stated were right at the edge of adding artifacts while just shy of doing so, and did not accentuate noise (why threshold 2 instead of the default 1). If you have not done so yet, I suggest you download them and process them in C1 with the settings I gave, then let me know what you think -- you might not get lost :)

That said, I respect anybody's right to their own opinion for their needs -- that was the whole point I was trying to make by encouraging folks to work these files for themselves.

~~~

The surprise for me was clarity slider difference. Normally clarity REDUCES perceived DR -- however in the case here, the E file appears to start with less perceived DR, and so when processed per above the end result is still pretty identical -- at least close enough I do not perceive a DR difference onscreen or in print. (We can assume that the native sensor DR is identical in both cameras, so what I am referring to here is DR as visible in the processed files.) Moreover, and again to my eyes, the E file is too hot on micro contrast at the base setting of 0 clarity. I have since reworked these files and find that using a -10 clarity as base for the E file and then bumping the regular D800 to +15 makes both look "better but still equal" than my initial settings of 0 and +25. So were I to get an E, my base style would include a -10 or even a -12 clarity as standard.

Again, I specifically did not post screenshots from my crops as I wanted to encourage each to process for themselves. So I'll repeat: I heartily encourage EVERYBODY interested in either one of these cameras to download these raw files and work with them yourselves -- in this way you will see firsthand the difference, and can decide if it's really relevant/significant/meaningful for your needs.

Cheers,
 

glenerrolrd

Workshop Member
Two issues I have raised ..I still believe are important.

1. FOCUS...pick any target you like but give me have a chance to see if you missed it. Without anything in the near foreground and background ..the reviewer has no idea if the focus point was missed . Pretty much kills any discussion of resolution differences . This is obviously much easier if the lens used has a modest DOF and that may not be what you are testing for.

2. Use the native ISO and process to the “ideal” . Not sure about others but it can take a few weeks to establish a set of presets . You could argue that right out of the camera the D800E blows away the D800 and I am sure I can t yet make them the same . But if Jack can then its just a matter of sharing “best practices” ..which is where I hope these threads go to.

The 2nd point I think is a good debate topic ..I favor ideal over identical but I can see the other view .
 

tashley

Subscriber Member
Tim,

Normally I'd agree with you -- but in the case of the files above, my sharpening amounts as stated were right at the edge of adding artifacts while just shy of doing so, and did not accentuate noise (why threshold 2 instead of the default 1). If you have not done so yet, I suggest you download them and process them in C1 with the settings I gave, then let me know what you think -- you might not get lost :)

That said, I respect anybody's right to their own opinion for their needs -- that was the whole point I was trying to make by encouraging folks to work these files for themselves.

~~~

The surprise for me was clarity slider difference. Normally clarity REDUCES perceived DR -- however in the case here, the E file appears to start with less perceived DR, and so when processed per above the end result is still pretty identical -- at least close enough I do not perceive a DR difference onscreen or in print. (We can assume that the native sensor DR is identical in both cameras, so what I am referring to here is DR as visible in the processed files.) Moreover, and again to my eyes, the E file is too hot on micro contrast at the base setting of 0 clarity. I have since reworked these files and find that using a -10 clarity as base for the E file and then bumping the regular D800 to +15 makes both look "better but still equal" than my initial settings of 0 and +25. So were I to get an E, my base style would include a -10 or even a -12 clarity as standard.

Again, I specifically did not post screenshots from my crops as I wanted to encourage each to process for themselves. So I'll repeat: I heartily encourage EVERYBODY interested in either one of these cameras to download these raw files and work with them yourselves -- in this way you will see firsthand the difference, and can decide if it's really relevant/significant/meaningful for your needs.

Cheers,
Thanks for that Jack - actually I agree that with these files there's not a significant difference. It's more in marginal files at higher ISOs or with shadows you want to bring out a lot that I think the small things might show. All other things being equal I think the more base detail the better, even if it's only a small improvement, but I am at least for now keeping the non-E alongside the E until I've done a lot more fabric shooting. If people have a broad based practice covering many different genres, I think what you have done is show that the 800 is probably the better choice. For landscape, I want as much detail in distant leaves and grasses as I can get but if I end up doing a friend's wedding I'll be reaching for the non-E. I will also probably use the non-E for any portrait stuff that comes along.
 
Last edited:

ustein

Contributing Editor
>I want as much detail in distant leaves

I understand. But also to be sure: Often aliasing shows as detail but actually is fake. Very obvious with video.
 

Jack

Sr. Administrator
Staff member
For landscape, I want as much detail in distant leaves and grasses as I can get but if I end up doing a friend's wedding I'll be reaching for the E. I will also probably use the E for any portrait stuff that comes along.
I think you meant you'll be grabbing the E (no AA) for landscape and the non-E (w/AA) for weddings and portraits? ;)
 

Jack

Sr. Administrator
Staff member
>I want as much detail in distant leaves

I understand. But also to be sure: Often aliasing shows as detail but actually is fake. Very obvious with video.
Uwe raises another good point: In addition to the perception of "fake" detail, like adding grain can sometimes do, with my Phase I frequently find moire in tiny leaves and grasses. Fortunately, it is usually of such a small level it never shows in a print, but sometimes is significant enough that it needs to be blurred or processed out. Either way, you do find it surprisingly often in "nature" shots.
 

tashley

Subscriber Member
I think you meant you'll be grabbing the E (no AA) for landscape and the non-E (w/AA) for weddings and portraits? ;)
You think I was confused there? I have the D800 strap on the E version :ROTFL:

I edited it - thanks for pointing it out - that was indeed what I meant!
 
Here's an idea to play with. A friend suggests, based on the design of the lowpass filter array, that a simple linear polarizing filter on the lens, oriented either vertically or horizontally, will cancel the effects of one of the filter elements on these cameras. If I understand the reasoning correctly, this would have the effect of slightly increasing the resolution of the 800, or decreasing the resolution of the 800e (bringing them both to the same level).

I've asked if he has a source or if he's tried it, and am still waiting for a replay. It would certainly take no more than wasted afternoon to test the theory.

edited to add:
If this works, it could open up some possibilities for geeky tweaking... like if you have a vertical pattern that causes moiré, you could get some horizontal blur from the AA filter, but cancel the vertical blur, for an infinitessimal gain in sharpness. I'm guessing the advantages would be purely academic in most cases, but for those with an 800e, you might be able to get a bit of moiré control in some settings.
 

D&A

Well-known member
Jack, although I"m in almost total agreement with your assessment that files from both cameras are near identical in detail, for all practical purposes....and have also come to this conclision by printing large format crops from test files....I wonder if this will still hold true when higher ISO comparative files are examined from both cameras. With additional sharpening of the D800 files required in part to emulate and match those from the D800E, I suspect that producing near identical output from both cameras will be more problematic/difficult as the ISO settings climb. Has anyone done comparions of near identical files adjusted optimally and individually for both cameras at say ISO 1200 and above?

Dave (D&A)
 

Jack

Sr. Administrator
Staff member
Here's an idea to play with. A friend suggests, based on the design of the lowpass filter array, that a simple linear polarizing filter on the lens, oriented either vertically or horizontally, will cancel the effects of one of the filter elements on these cameras.
I am not so sure. Assume polarized light hits the first splitter (call it vertical), so a single point source turns into 2 points even though it's polarized. Now it hits the 1/4-wave plate, and since it's already split the split pair gets rotated 90 degrees regardless of polarization. Now the beam pair hit the second splitter (call this one horizontal) and it still gets split into 4 points. I may be wrong in my understanding though...
 

Jack

Sr. Administrator
Staff member
Re ISO...

We won't know for certain until we play with good comparative files from both cameras. HOWEVER, if past history is any indication, an AA filter can actually help attenuate the effects of noise. This new design Nikon filter may behave differently however...
 
Top