Megapixels are a mug’s game – most of the time. If you’re producing close-ups of your new-born which will be distributed via cell-phone or web, you don’t need a ton of megapixels. For such photos, 3 megapixels would be plenty, even if you wanted to print the occasional 5x7 or 8x10. You don’t need a $5,000 rifle to shoot woodchucks (and yes, you can buy such a varmint rifle).
So what’s the big deal about 6, 8, 12, 22 more megapixels? It depends on what you’re trying to do and how the images are intended to be used.
Super resolution only matters when you’re super fussy and are planning on making super-large prints, OR when you’re planning on cropping to just a small part of your frame and then blowing it up to even normal sizes like 5x7 or 8x10.
Notice, I said the P-word, as in PRINTS. There isn’t a monitor made that can portray the full dynamic range and intricate detail of a finely-crafted print. That’s not an opinion – it’s a simple matter of fact. The technology of monitor display spaces color dots too widely and too imprecisely to really capture silk-smooth texture or subtle tonal shifts. You need a high-quality print to do that.
There are a number of components affecting resolution. The most important are glass, cleanliness, pixel count, pixel size, and the camera’s internal data-capturing algorhythms.
For now, let’s leave the camera’s picture-taking logic out of it and assume you have a competent camera – today, most digital cameras are better than competent, they’re really rather good at finding exposures, autofocus, and the like – all of which affect the final resolution (clarity) of the image.
Glass
Lenses get very expensive very quickly. That is because one piece of curved glass is never going to do the whole job. So lenses are made of many clusters of glass (elements), each with its own coatings, curvature, placement in the barrel, transmisivity (ability to transmit light) and color characteristics. The greater the transmisivity (read ability to shoot in low light), the greater the cost of the lens. That’s why a 300 millimiter f2.8 lens costs about 4 times as much as a 300 millimeter f5.6 lens.
The quality of your glass (lenses) really matters. Imagine that you smeared Vaseline all over the front of your lens – no way would a picture shot through than lens ever be sharp. Ugh – yet in fact, smearing Vaseline on a lens was an early way to achieve soft-focus and vignetting ON PURPOSE.
Dirt
Even a thin film of smoke, chemical deposits (as from cheapo lens cleaning sprays) will interfere with the lens’s ability to do it’s two main jobs: transmit light and focus light without reflection, fringing (chromatic aberration), color distortion, etc. Dust is your enemy – and I know this to my cost from extended shooting in Canada Alaska, which are permanently and inescapably dusty.
You can do things to reduce dust, but you can’t eliminate it permanently. It WILL build up and haunt you. You have to go after it carefully, frequently, and thoroughly. And that includes cleaning your sensor, not just your lenses.
So if you need really high resolution, get the best glass for the job you want to do. And keep it clean, but don’t scratch it while cleaning it! And by the way – don’t apply a cloth to the mirror in your camera – it’s not coated and scratches very easily.
Megapixels -- How Many?
Now, at last, we get to those megapixels. Here’s the deal. Let’s imagine you are a landscape, architectural, or product photographer. Let’s further say you have a 6 megapixel camera. You take a picture of a distant tree with an eagle sitting on top of it. The tree is standing in a rocky field. The tree and eagle take up about 20% of your shot – the rest is field and sky. So only 1/5 of your pixels are available for the important part of your subject! In this example, that means you have at most 1.2 megapixels out of 6 dedicated to tree and eagle. If you wanted to crop the image to show just the branch and the eagle sitting on it, you’re down to about 5% of your pixels, of about 300,000 pixels.
You’re not going to get any kind of sharpness and detail from that number of pixels.
Let me make this simpler. The more of your pixels are devoted to your subject, the better your print will look. If you have only 20 pixels devoted to your eagle, you’re doomed. 20 million pixels – now you can make a 30x40 and it fly jump out of the tree at you.
That’s why resolution matters much less for close-ups than it does for landscape shots. You don’t crop close-ups much, so you don’t give away pixels to useless parts of the image.. On the other hand, if you want every detail in a landscape to be clear and sharp, you need great glass, correct exposure, and all the pixels you can get. Here are some examples:
The images differ some in color, but that’s not the point here. The Mamiya camera / lens combo is FAR sharper than the Nikon because 1. The glass is better and 2. There are more pixels devoted to the same portion of the subject.
Here's another example of where you need pixels to capture fine detail in a landscape. This picture of Yellowstone Falls was taken from more than 1/2 mile away. The actual-pixel-size insert shows just how fine the detail can be when you have enough pixels. Pity there's just a hint of motion blur, or this would be even better!
Here's another example of where you need pixels to capture fine detail in a landscape. This picture of Yellowstone Falls was taken from more than 1/2 mile away. The actual-pixel-size insert shows just how fine the detail can be when you have enough pixels. Pity there's just a hint of motion blur, or this would be even better!
Pixels -- How Big?
Sorry, folks, size does count. Bigger pixels do a better job of capturing light than small ones. Imagine you are playing a carnival game in which you’re tossing dimes at a tray crammed with red, green, and blue shot-glasses. Some will go into the shot glass and win you a stuffed bear. But many will bounce off and either miss the shot glasses altogether, or land in the wrong one. This is called "scatter."
The pixels on your sensor are exactly like the shot glasses – only the photons striking the bottom of the glass are captured. Bounce-offs don't count at all! If the pixels are small and very tightly packed, many will scatter instead of going to the active part of the sensor and being recorded.
Now imagine that you’re still tossing dimes, but this time the shot glasses have been replaced with coffee cups (no handles, so they’re still tightly packed). This time most of the nickels will find a home (and the correct home at that). You win the teddy bear!
When you have a DX sensor (smaller than 35mm), and give it 10 megapixels, those pixels are mighty small and very tightly packed. That means that there’s more light scatter, and the effect of that is less dynamic range. Bad. But if you take the same number of pixels and put them on an FX sensor, and if they’re still tightly-packed, each pixel will be bigger and your dynamic range goes up.
On the Phase One P25 sensor (now obsolete) the pixels are 9 microns across. And there are 22 million of them spread over a sensor the size of a 645 film negative (that's much bigger than a 35mm). On a Nikon D300 the pixels are 5.45 microns across. On the D700, they’re 8.45 microns across and there are 12.8 million of them on a 35 mm (FX) size sensor. Pretty good, but as the Grand Teton picture above illustrates, not AS good for landscapes -- but terrific for portraits and macros.
So go back to the dime-tossing analogy. Bigger sensor+bigger pixels = more light captured.
Pixels -- Bit Depth
One more factor affects dynamic range, and with it, effective resolution (the amount of detail in shadow areas, for instance). This is bit depth. Bit depth is defined as the number of bits (binary digits) used to record the illumination level of a pixel. More is better. The D700 carries up to 14 bits of color information about each photon of light striking the sensor, while the P25 carries 16. This may not sound like much of a difference, but it’s a very big deal. RAW image files from the Nikon are about 12 megabytes, from the P25 about 24. megs. Better have plenty of memory cards! Big files take more processing horsepower, and hence allow fewer shots-per-second. The P25 is dead slow (0.7 frames per second), while the D700 shoots up to 7 comfortably. Frames-per-second is irrelevant for landscape work, but vital for shooting sports or birds. Buy for what you need to shoot!
Here's one more full-image / detail example (Mamiya / P25). In this case the bit depth is important; without it you wouldn't see the red in the boat, small as it is. Believe it or not, that little touch of color is what sold an 18" wide print of this scene.
Summary
The point of this article is not to compare specific cameras. It’s to clarify when you need lots of pixels for maximum resolution and when you don’t.
You need lots of pixels when you want to make razor-sharp wall-size pictures of landscapes OR when you want to make large prints of deeply-cropped subjects.
You don’t need lots of pixels to shoot close-ups, or when you don’t plan on printing large, or when you don’t need to crop then print.
Most of us carry cameras with more megapixels than we need. But as it usually is in life, we make choices because some marketer tells us simplistic stuff like “More pixels is BETTER.”
More may be better – but it depends on the jobs you want to do. Define what you’re trying to do, then match the megapixels to the task – or contentedly waste your money for something you don’t need and will probably never take full advantage of.
