Nikon's D4 has been officially announced. Headline specs, in case you missed them, include the following:
- 16mp FX sensor with EXPEED 3 processor
- 10fps with full AF and AE; 11fps without
- tweaked 51 point AF system with more sensitivity
- ISO100 - 12800 native; expandable to ISO50 - 204800
- 91,000 pixel RGB sensor to replace the previous 1005 pixel version
- lots of ergonomic tweaks including illuminated buttons and two extra joystick controllers
So far, so evolutionary. The D3/s have a devoted (and deservedly so) following. They're accurate, reliable and capable of turning out incredibly clean files in all sorts of challenging situations. Nikon would've messed with that succesful formula at its peril. Still, taken on their own, there's not really enough here to persuade a happy 3-series owner to upgrade. A few extra pixels, a tiny bit more speed... at the price Nikon's asking ($6000 or £4800) those incremental extras are looking expensive.
Oh, and inevitably the forums are already full of doom and gloom about the fact that Canon's forthcoming competitor is, apparently, better (summary: 'OMG the Nikon D4 is 2 smaller and 2 slower. WTF?'). Can I be the first to point out that this isn't remotely like the D2H vs the 1DII? (in case you weren't there, Nikon's 8fps 4mp D2D was launched weeks ahead of Canon's markedly superior 8fps 8mp 1DII, although it wasn't the pixel count that killed the D2H. But that's another story).
The real story is away from the internet sideshow and the D4's basic specs. The real story is video. Because Nikon has been working hard to build as much video goodness into the D4 as it possibly can:
- a full set of 1080p and 720p output options inluding 30, 25 and 24fps (60fps only at 720p, though)
- reduced rolling shutter effect thanks to faster data processing
- B-frame compression to improve H.264 output
- uncompressed video output available via HDMI port
- stereo headphone jack
- stereo mic jack with manually adjustable level
- video-specific controls
- full manual exposure via shutter speed, aperture and ISO
- full-time AF with face detection
- full HD integration with crop modes (1.5x and 2.7x)
It's an impressive feature set, for a stills camera at least. If you need to produce video as well as stills, it's possibly worth a look. But (unfashionably) I'm still not convinced by video in dSLRs. The biggest problem, as I see it, is that dSLR ergonomics are just plain wrong for video shooting. The D4, as a stills camera, is designed for fast, portable operation. Decades of design experience have gone into the user interface (I can trace the basic design of many of the D4's controls back to my 1989 F801). It's a camera that'll work very well indeed on the fly, off a tripod, with a photographer who's not afraid to follow the action. I can say that with confidence because, in all important aspects of design, it's barely different from my D3.
Switch to video and it's a different story. A video-enabled dSLR really needs to be tethered to a tripod - and a solid one, at that. While you can certainly make the case that that's good practice for any video shooting, dedicated video cameras are at least useable handheld. A dSLR? Not so much. The ergonomics for video are just plain terrible, and that's not really surprising. I'm as impressed with video in a dSLR as I would be with a stills function built into a video camera.
Much of the original impetus for incorporating video into a dSLR came from big news organisations, who were clamouring a few years ago for multimedia content. But the revolution never really happened, partly because (surprise, surprise) the budget wasn't there. It may sound attractive to a commissioning editor to ask for video and stills from a single photographer, but have you ever tried shooting both at a single event? It's not easy. And, contrary to popular belief, simply having a video mode on your stills camera doesn't actually make it any easier.
So I'm left with the feeling that video is an extra that relatively few photographers need, let alone know how to make effective use of. If you strip away the big sensor / shallow depth of field thing (which has become so over-used in the past few years that it's now a visual cliche), the only remaining 'benefit' of incorporating video into dSLRs is the 'two into one' argument. And, as I've just argued, I'm not sure that's compelling either commercially or in ergonomic terms.
None of which would matter much, if it weren't for the fact that video is clearly eating up more R&D resources and almost certainly adding to production costs.
Will I buy a D4? No. Would I buy a D700 with the D4's sensor (and no video) in it? Probably.