dbreview on the D800 and switching from Canon to Nikon [1/5]

Started Apr 4, 2012 | Discussions thread
Flat view
Daniel Browning Senior Member • Posts: 1,058
dbreview on the D800 and switching from Canon to Nikon [1/5]

Of course, dbreview stands for Daniel Browning's Review.

I've been shooting Canon since I took an interest in photography seven years ago. Last week I decided to switch to Nikon, so I waltzed into a local store, grabbed a D800 off the shelf, and proceded to sell off all my Canon gear. So far I am very pleased with the D800.

I switched camera systems for three main reasons: dynamic range, resolution, and video. The first of these is that Canon's dynamic range is limited primarily by pattern noise (AKA "banding"). A large number of dynamic range tests that ignore pattern noise show Canon to be trailing the competition by a significant margin (e.g. DxOMark), but when you factor in the effect of pattern noise, the disparity grows even wider, to an order of magnitude difference.

So why did I switch now? Why not wait and see if Canon addresses at least the dynamic range problems in a future camera? Mostly because I'm tired of waiting. Unlike their competitors, Canon hasn't improved it much over the last 7 years. I looked at some 5D3 raw files to see if Canon finally fixed the problem and found they had not:

For example, just opening this ISO 125 raw file ( http://www.kleptography.com/dl/5diii/raw/02.cr2 ) with default settings in LR4 shows slight (to me) line noise in the dark tonal levels. Enable lens corrections and add a 2/3-stop push and it becomes very obvious. That is a woefully inadequate performance compared to the competition.

The degree to which pattern noise occurs on my 5D2 varies with color temperature, ISO tweeners, lens & SAOIR correction, etc. With other DLSR cameras, it becomes possible to photograph subjects that have a greater range between highlights and shadows, to emulate a film-like highlight rolloff, or even to keep an image that had the the optimal exposure but a sub-optimal gain (AKA ISO) setting -- a great heresy among the in-crowd.

The increased pixel count also factored into my considerations. A lot of photographers are characterizing this jump to 36 MP as a sort of tremendous, earth-shattering, never-seen-before leap in pixel count. But it's only a 72% increase (32% resolution) over the 5D2. Compared to the D3X, A850, and NEX cameras, it's only 50% more pixels (23% more resolution). When you consider it from the perspective of a relative increase, it's the same ballpark as going from a 30D to 450D (50%) or 40D to 50D (50%). It's still a noticeable increase of course, which made it a factor in switching.

It seems to me that most photographers only concern themselves with attaining the bare minimum resolution necessary for a good looking print. I look at it from a different perspective: what is the highest possible resolution that is needed before further increases make no difference in the print. Some photographers say that if you are only printing an 8x10, there is no benefit in the increased resolution of the 5D2 (21 MP) over the 5D (13 MP). But in my tests I found that to be untrue. Even in a print as small as an 8x10, 13 MP is visibly worse than 21 MP due to lower contrast from the OLPF, post processing (distortion, horizon, and other corrections), a small amount of cropping (e.g. 5% from each side), and other factors. See http://photography-on-the.net/forum/showthread.php?t=747749

So while I do consider 21 MP to be capable of delivering the maximum possible detail and contrast in an 8x10 (if there is very little cropping), I'd prefer to be able to do the same with larger print sizes such as 12x18, 16x20, 24x30, etc., and with non-trivial degrees of cropping. Compared to the 5D2, the D800 should be capable of delivering the maximum possible contrast in a print that is 31% wider (e.g. 8x12 -> 10x15), but only if all other sources of diminishing returns are avoided.

Video was another significant factor. Perhaps about a third of my time with a DSLR is shooting video (99.9% going by disk space used). While I do also have an XH-A1 video camera and sometimes bring it along on my shoots, the video feature on my HD-DSLR has a number of significant advantages over it: superior low light capability, more control over DOF, takes up 0% of the space in my camera bag, weighs zero pounds, and uses all my existing still camera lenses. That is why video functionality factors into the decision so heavily.

With the clean HDMI feature on the D800, it is now finally possible to shoot for longer than 10 minutes at a time and with a huge increase in compression quality. Before now, my only option was the XHA1 with 1080p over component. Sometimes I setup to record from uncompressed 1080p directly to a high quality intermediate format using my BMD Intensity Shuttle. That way I can skip over low-quality compression (e.g. 25 Mbps HDV or 50 Mbps AVC-Intra) and go right for the good stuff (500 Mbps wavlet such as Cineform). Plus there is no time limit (not even tape changes every hour). I find it baffling that Canon did not provide clean HDMI out.

[Part 1/5]

Flat view
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow