D800 - the best affordable digital landscape camera yet?

[Range Loser]

I agree with Shadowfox that the D800 may not look "similar" to film, but it is not my aim to start a film v digital post with this thread, I only referred to medium and large format film as a comparison of a highly detailed media.
I also agree that both film and digital have hassles in one way or another, and some prefer one way of working to another. Looking at my film scans, I can easily see where my lack of skills/decent equipment lies. Personally I find it much easier to achieve a level of proficiency I am happy with using the digital route.

I was at Focus on Imaging today and was only inches from a D800, but as there was only one camera on the Nikon stand, and a large queue of people eager to have a hands on, I decided I could wait until they appear in the shops.

 

[sojournerphoto]

Somewhat. But again, these are not tiny pixels we're talking about. 4.7 µm pixels = 212 lines per mm. Cell phone sensors are routinely using 1.5 µm pixel pitch!

Even moderately good lenses readily resolve 100+ line pairs per mm. That's roughly what you get with Fuji Velvia 100 (100 lp/mm @ MTF 30) or even Superia 1600 (@MTF25). TMAX100 and ACROS resolve 100 lp/mm at MTF70 and at lower contrast TMAX and ACROS resolve up to 200 lp/mm.

The D800E may show fine detail with higher contrast (at least on center; the corners will be another story), but if the lens and technique are good enough to generate that sort of detail, you'll also get aliasing of fine structures. I will go further: when people use M9s and claim that they don't see aliasing, that's simply an indication that their technique (focus, camera support) is not sufficient to see it.

If you want to shoot wide open, the lens will matter more, of course. But what these sensors (or TMAX100, or Velvia) will really reveal is poor technique. Holding the camera stable enough to actually resolve 5000 vertical lines (2500 lp) will demand a stable support, no wind, mirror up, and remote release.

For landscapes there are further limitations: the need for depth of field and (for distance shots) atmospheric haze. Stop down to f/11 or f/16 to keep the foreground in focus and you're well into diffraction-limited territory. Now that fancy lens's fancy rendering wide-open doesn't matter so much, eh? This camera is going to convince a lot of (serious) landscape photographers that what they really need is a T/S lens or two.

Sharpness is a bourgeois concept, anyway. I know this because a bunch of people on RFF told me so.

Actually, I'm not sure your technique needs to be that superlative to see aliasing on M9 files. I routinely find areas of colour moire in M9 images shot hand held and from wide open to f8 using a ZM 2/35 Biogon (or any other ZM lens apart from the C-Sonnar wide open!). Sometimes it's not important and other files are really quite spoiled for printing at anything bigger than small sizes. There is also the issue of 'funny edges' that I read to be aliasing or sampling artifacts where there isn't sufficient repetition to create full blown moire patterns. I presume this will apply to the D800e, albeit at a finer scale of detail.

The other issue is that it will reveal camera shake more quickly than a lower resolution camera - somehow this carries throgh into print once it's recorded - which is one reason I don't really think the M9 needs any more resolution as it is primarily a handheld camera I think.

Having said all that, I think the D800e will be a lovely piece of kit, but make sure you budget for an L plate and decent tripod.

Perhaps strangely, I've sort of concluded that the 5d3 may be a better generalist camera - but I've got some Canon lenses and bits lying around and no slr body these days.

Mike

 

[JRG]

Don't forget developing costs and $2000 for a Coolscan 9000.

You know a place where a CoolScan 9000 sells for $2000?

 

[semilog]

Actually, I'm not sure your technique needs to be that superlative to see aliasing on M9 files. I routinely find areas of colour moire in M9 images shot hand held and from wide open to f8 using a ZM 2/35 Biogon (or any other ZM lens apart from the C-Sonnar wide open!). Sometimes it's not important and other files are really quite spoiled for printing at anything bigger than small sizes.

Maybe your technique is better than you realize : D

 

[Chriscrawfordphoto]

You know a place where a CoolScan 9000 sells for $2000?

It sold for that new, but people now are demanding stupid prices for new ones. Often double the new price.

 

[MCTuomey]

Perhaps strangely, I've sort of concluded that the 5d3 may be a better generalist camera - but I've got some Canon lenses and bits lying around and no slr body these days.

Mike

+1 Agreed, I look forward to trying one. Canon TS-E 17/24/90, MP-E 65, great telephotos, portrait and street primes like the 35L, 50L, 85L, 135L - fine flexible system indeed.

 

[rpsawin]

Somewhat. But again, these are not tiny pixels we're talking about. 4.7 µm pixels = 212 lines per mm. Cell phone sensors are routinely using 1.5 µm pixel pitch!

What can a really good lens do? Using a Tech Pan-like emulsion and Spur developer, Zeiss has shown that under laboratory conditions the ZM 25/2.8 Biogon can hit 400 lp/mm! A perfect optical system and detector operating at that resolution would, in full frame, require 800 lines/mm -- on FF, a 500 (five hundred) Mpix sensor...

Even moderately good lenses readily resolve 100+ line pairs per mm. That's roughly what you get with Fuji Velvia 100 (100 lp/mm @ MTF 30) or even Superia 1600 (@MTF25). TMAX100 and ACROS resolve 100 lp/mm at MTF70 and at lower contrast TMAX and ACROS resolve up to 200 lp/mm.

The D800E may show fine detail with higher contrast (at least on center; the corners will be another story), but if the lens and technique are good enough to generate that sort of detail, you'll also get aliasing of fine structures. I will go further: when people use M9s and claim that they don't see aliasing, that's simply an indication that their technique (focus, camera support) is not sufficient to see it.

If you want to shoot wide open, the lens will matter more, of course. But what these sensors (or TMAX100, or Velvia) will really reveal is poor technique. Holding the camera stable enough to actually resolve 5000 vertical lines (2500 lp) will demand a stable support, no wind, mirror up, and remote release.

For landscapes there are further limitations: the need for depth of field and (for distance shots) atmospheric haze. Stop down to f/11 or f/16 to keep the foreground in focus and you're well into diffraction-limited territory. Now that fancy lens's fancy rendering wide-open doesn't matter so much, eh? This camera is going to convince a lot of (serious) landscape photographers that what they really need is a T/S lens or two.

Sharpness is a bourgeois concept, anyway. I know this because a bunch of people on RFF told me so.

From the Nikon D800 Manual:

#1 item from D800 D800E user manual

"While its high pixel count of 36 megapixels gives the D800/
D800E resolution unrivalled by previous digital SLR cameras,
a side ef ect is that bokeh and blur are made that much more
obvious. Realizing the full potential of a camera with over
30 million pixels involves a thorough appreciation of bokeh
and blur, careful selection of settings and of tools (such as
lenses and tripods), and working with the best possible subjects."

I appreciate your response and admit that the deep drill on spec's is a bit beyond my interest. My point is that, from a practical perspective, it pays to be thoughtful about gear selection and workflow in the field.

Best regards,

Bob

 

[sojournerphoto]

Following the discussion on aliasing and colur moire I thoght I'd shar ethis picture from this afternoons stoll with family

Mike

M9, Zeiss ZM 1.5/50 C-Sonnar f8 1/60 sec handheld

Full frame

Crop at 300% - moire and detail has gone wrong in raw conversion

 

[ChrisC]

.....M9, Zeiss ZM 1.5/50 C-Sonnar f8 1/60 sec handheld

Crop at 300% - moire and detail has gone wrong in raw conversion...

It's not only the lens/camera/subject/aperture/distance you are exhibiting, but also how your chosen Raw Developer has handled the file. Other Raw Developers may well handle the development differently, and possibly without the artifacts. Some LUF members reckon on Capture One for problem files in particular.

.......... Chris

 

[gavinlg]

Haha apparently the d800 isn't fully compatible with the nikkor 24mm PCE tilt/shift lens. Nikon mentioned it somewhere in some technical report. So much for F mount compatibility...

 

[Jay Decker]

Mounted a Schneider Xenotar MF 80/2.8 (a Pentacon 6 mount lens) on the D800 with an adapter and shot a couple photos of our roses this evening in full evening shade. The photography won't win an prizes, but this lenses color rendering and bokeh is causing me to think twice about the "quality" of Nikon lens... think I need to try the Zeiss Sonnar 180/2.8 (a Pentacon 6 mount lens) on the D800.