What image quality really means – Part 1

March 03, 2012  •  Leave a Comment

I’ve spent much of yesterday and today prepping images for submission to Alamy. Since Photolibrary was swallowed by Getty, who then presented me with an unacceptable contract, Alamy has been my main library outlet. 

Mills at Tampere, Finland

In some senses that’s neither here nor there; the point is that preparing images for any professional library means giving them some very close scrutiny. This includes examining the entire image at 100% magnification to check for any dust spots or other issues. This is boring but necessary – it’s certainly better than having images rejected by Alamy’s quality control.

And during this rather mechanical process there’s both time and fuel for reflection on what image quality really means. Which all relates to my thoughts about the headline-grabbing 36 megapixels of Nikon’s new D800/D800E. It doesn’t look like I’ll get my hands on either of these until around the end of March, but I’ve seen a number of sample images and studied reports from those who've had access to pre-production cameras.
Glasson Dock, Lancashire
Meanwhile, I’ve been working mostly with images from my Nikon D700 (a “lowly” 12 megapixels) and D7000 (16mp). Here’s one from each camera. Without reading ahead, can you tell which is from the 12mp camera and which is from the 16mp one? (No cheating, now.)
The first image is from Tampere in Finland; the second is from much closer to home, Glasson Dock in Lancashire. I’ve picked these shots because they’re both full of fine detail. 
When I view the Tampere shot at 100% on my 24” screen, it’s the equivalent of a print about 50”/128cm wide. I can’t just see every brick – that’s easy. I can see variations in texture form one brick to the next. See those windows on the shadowed wall facing the camera? I can clearly see every slat in the blinds.
Same thing with the Glasson image – except that (because it’s from the D700) it’s “only” the equivalent of a 44”/110cm print. But if anything, the resolution of fine detail seems even crisper. Every rigging line is clearly resolved and I have no difficulty reading the names of every boat. 
This being so, the obvious question is: just how much more resolution do you actually need? I’ll return to this shortly, but first: I said that the Glasson (D700) image looks crisper. The most likely reasons for this are nothing to do with any difference between the two cameras. They’re far more likely to be due either to the different lenses used or to the fact that the Glasson shot used a tripod while the Tampere image was hand-held (albeit with elbows braced on the parapet of a bridge). The lenses are, respectively, a Nikon 80-200mm f/2.8 and a Nikon 18–70mm f/3.5/4.5 ‘kit’ lens.
How much is enough?
Two points. 
First, even the smaller file bears scrutiny at a magnification equivalent to a 44” print. I can count on one hand the number of occasions I’ve needed to make prints larger than that. Of course people will say, “but you need to print at 300 ppi and a 4256 pixel-wide file will only make a 14” print”. This is little better than nonsense. 300ppi is the default standard for book- and magazine-printing, so this image file would not quite make a double page spread without a bit of scaling up. However, because the detail in the original is so crisp and clean, this would not be a problem. For a larger print to go on the wall there’s no need to print at 300 ppi.
But, more to the point, how often do most people supply images for double-page spreads or 50” prints? If they print at all, they’re usually printing at A4 size on a domestic inkjet. The largest they’ll normally see their images is on a computer screen or an HD TV. And yes, your HD TV might be 50” wide – but it sure as hell doesn’t have 300ppi resolution. I’ve noted before that ‘Full HD’ resolution is 1920 x 1080 pixels.
At some point, but probably not for at least five years, there will be Ultra-High Definition TV (UHDTV) sets available; one Japanese company has already demonstrated an 85" LCD display capable of 7680 x 4320 pixels. Which may be of interest for those with homes big enough to accommodate a TV as big as the door into the room, but may prove more relevant to big-screen public displays (they’ll be trialling some during the London Olympics). And, yes, that’s equivalent to about 33 megapixels – so the D800 appears future-proof in that respect.
However, it’s a fallacy to think that the existence of 33mp displays means that every image capture will have to match that standard. In fact, they probably never will. Devices with smaller sensors, like compact cameras and mobile phones, are butting up against the physical limits already. One of the key issues is diffraction. Smaller sensors take smaller lenses and therefore the lens aperture is physically small. Already we see that most compact cameras won’t stop down to apertures smaller than f/8 because diffraction would just make the image appear softer. Trying to cram more pixels onto a sensor of the same size achieves nothing if the actual resolution is limited by diffraction. There’s some information here (albeit referencing cameras that are now out of date).
Which leads on to my second point.
Yes, the D800 (and possibly even more, the D800E) are capable of capturing extraordinary levels of fine detail. For some photographers, this is really worth having. However, their true potential is only achieved if every other aspect of the photographic process is also up to scratch. 
I’ve already mentioned that the differences between my sample images are attributable to differences between lenses, and/or the use of a tripod. The astounding resolution of the D800 will only be fully realised by photographers using the very best lenses and impeccable technique. And when you start ‘pixel peeping’ these images at 100% depth of field is going to appear to have contracted, too.
I am really looking forward to getting hold of a D800, but I’m also nervous that it will show up the shortcomings of some of my lenses – and will it also make my feel that my long-trusted carbon-fibre tripod is’t quite solid enough after all. The cost of the camera body could be just the start… maybe it’s a good thing I’ll only be getting it on loan!
Coming up in Part 2 – what else does image quality mean?


No comments posted.
January February (2) March (4) April May (3) June July (1) August (1) September October (1) November (3) December
January February March April (1) May (2) June (2) July August September (1) October (2) November (3) December
January (3) February March April May June July (1) August September October (1) November (1) December (1)
January February (2) March (1) April (1) May June July August September (1) October November December
January February March April May June July August September October November December
January February March April May June July August September October November December