by Guy Lerner
Most people buying digital cameras for the first time are lured by the numbers game. For them, the more megapixels, the better the camera. But not all pixels are created equal.
Unlike most things digital, when you’re talking pixels, smaller is not necessarily better. Pixels used by imaging sensors (the light-recording components found in most digital cameras) vary in size from one manufacturer to another and from camera to camera. I’m no scientist, but it makes sense that pixels used to convert light information coming from a lens would perform better if they had a larger surface area to capture as much light as possible. Put another way, given the same number of pixels on an imaging sensor, the sensor with the larger individual pixels will record more light information – and be able to produce higher-quality digital images – than the one with the smaller pixels.
Miniaturization has allowed scientists to squeeze more and more pixels onto imaging sensors by making them smaller. That means the same number of smaller pixels can be squeezed onto a sensor the size of a fingernail as the number of larger pixels on a sensor the size of a postage stamp. However efficient the smaller sensor may be, it will not necessarily produce better images than the larger sensor with larger pixels.
This explains why digital SLR (single lens reflex) cameras usually produce higher quality images than compact and point-and-shoot cameras. Digital SLRs are equipped with imaging sensors approaching the size of a 35mm film frame (some, like the Canon 1Ds, boast a full 35mm frame-size sensor). Compact cameras – even high-end compacts costing upwards of US$1000 – use sensors a fraction the size of SLR sensors, typically about 1/1.8″ or 5.5mm x 4.1mm. The compact may have a 3-, 4- or even 5-megapixel sensor, but in most instances it can’t produce images comparable in quality to those made by a 3-megapixel digital SLR.
What’s in a pixel?
Camera manufacturers aren’t always upfront with their customers when it comes to pixel numbers. What is actually quite a simple nomenclature can be contorted by wily marketers just because confusion is rife and the technology, for consumers at least, is relatively new.
Sensors don’t always use all their pixels to record an image. A sensor may have 4.3 million physical pixels, but only 4-megapixels are used to record light information to form an image. This is typically described (buried in the technical specs on page 258 of your user’s manual) as “effective” pixels, and is the number you should always look at when determining a camera’s actual pixel count. A 3.3-megapixel camera with 3.1 effective megapixels is for all intents and purposes a 3.1-megapixel camera. Most companies are now required by law to promote their products’ effective pixel counts alongside total pixels, but some are not, so buyer beware.
As discussed above, however, the number of pixels – effective or otherwise – is not the only variable in the resolution equation. JR Geoffrion in his article “Understanding Digital Camera Resolution” sums it up superbly: “…several other factors affect resolution. Some of the most important factors include sensor design, interpolation algorithm, the lens itself, focal length (in the case of zoom lenses), focus distance, aperture, position in the image field, orientation (horizontal, vertical, and diagonal), scene contrast, and vibrations. Due to the number of degrees of freedom, comparing two camera’s megapixels rating would be absolutely pointless.”
While resolution measured in pixels can be used as a yardstick for the “amount” of information a sensor can capture, and therefore the amount it can output, it does not equate to the visual quality of the final image produced. For starters, the resolving power of more pixels is not relative; a 6-megapixel camera won’t output an image with twice the information of a 3-megapixel image, but rather that of a 1.5-megapixel camera (in a two-dimensional image, four times the number of pixels are required for every doubling of resolution).
Pixel density, in contrast, tells us more about the quality of an image. Knowing how closely pixels are packed together is one way of knowing how much detail can be captured. Pixels spaced too far apart won’t be able to resolve as much detail as the same number of pixels packed closer together, all else being equal. But pixels spaced too closely together, while potentially being able to resolve finer detail, also suffer from the effects of noise, where information from one pixel “leaks” onto another, and the circuitry in the camera is not sophisticated enough to compensate for erroneous information read back from the sensor.
Pixels, lines and inches
If we know the pixel number, size and density of a sensor, and understand that other factors affect resolution in different ways, can we make a better assessment of a camera’s resolution as it relates to image quality? Yes and no.
Camera resolution is related to image resolution only in units of measurement (pixels). While raw camera resolution is expressed in pixels, image resolution is expressed in pixels-per-inch (ppi) – the number of pixels used for every square inch of the image. A 300 ppi image is made up of 300 pixels for every square inch; the larger the image, the higher the number of pixels at any given ppi resolution (a 4″x6″ 300ppi image is made up of fewer pixels than a 8″x10″ 300ppi image, for example). Since same-sized sensors can have different numbers of pixels (remember pixel size and density), measuring a sensor on a ppi scale is impossible.
Another curveball: printer resolution is measured in dots-per-inch (dpi). Printers don’t use pixels to make up an image, they use dots of ink. Different printers use different numbers of dots to render each pixel in an image. For example, a printer that uses six dots for every pixel with a resolution of 1440×1440 dpi can render an image at 240 ppi (1440 divided by 6). The higher the number of dots and the higher the dpi resolution, the better the image quality, because each pixel is made up of various color, contrast and brightness values so the more dots of ink of different colors and densities used to express these values the better.
It is generally understood that an image resolution of between 240 ppi and 360 ppi renders a photographic quality image from a normal viewing distance, so a printer capable of rendering an image at this resolution on paper is required for producing photo quality prints.
The output resolution of a camera may or may not be hampered by the compromises it makes in translating the raw information from the image sensor to a workable image on computer or printer.
Storage is one such compromise. Digital images, especially those from high megapixel cameras, consume significant amounts of storage space. Given finite storage capacity, they have to be managed properly to ensure that enough images can be taken and stored before they have to be taken off camera. To achieve this, cameras use various levels of compression to size-down the captured images and transfer them to memory. The most popular method of doing so is by using JPEG compression – a widely-used image standard that ‘discards’ information from the image which, in general, can’t be seen by the human eye or resolved on a printer. The higher the level of JPEG compression, the more information is discarded. At very high compression settings, visible ‘blocks’ start appearing in JPEG images that seriously degrade image quality, but the resulting file size may be very small.
For serious or professional images, where quality comes above any other consideration, JPEG compression should therefore be avoided. Most mid-to-high end digital cameras offer other forms of compression that don’t discard any information, but rather compress the file using sophisticated mathematical algorithms. These files are then saved in the popular TIFF format, or as RAW files containing only the information captured by the sensor without any in-camera processing. The disadvantage of this ‘lossless’ compression is larger file sizes, at the gain of more post-processing flexibility and usually higher image quality.
In-camera processing, in fact, is another compromise often made by cameras to achieve predefined results. Many cameras use ‘sharpening’ filters to restore the contrast lost in the edge details of an image when it passes through the camera’s various optical filters (these anti-aliasing filters are used to smooth out uneven or blocky edges between high-contrast parts of an image that are recorded by widely-spaced large pixels). The more an image is sharpened in-camera, the less it can be sharpened once it’s been downloaded to a computer because over-sharpening introduces its own harsh, unnatural edges (and in some cases serious color shifts as well).
Finally, the computer screens we use to view and edit our digital images are themselves a compromise. Monitors can display a maximum image resolution of 72 ppi (Macintosh) or 96 ppi (Windows), which is ideal for Web displays and general-purpose applications. Quality is inevitably compromised when working with images larger than the monitor’s set resolution, and careful attention needs to be paid to the output resolution from a printer rather than a monitor when working with high-resolution images (unless the images are only intended for Web display).
Camera resolution can be dissected in different ways, but the basics are easy to digest when you know how the pieces fit together in the bigger picture (pun definitely intended). The misconceptions I’ve tried to clear up in this article should give you a better understanding that:
The number of pixels on a sensor is only an indication of resolution, and is not a determinant of image quality.
Pixel size affects image quality, as does pixel density.
Effective pixels should be used to determine the number of pixels that make up an image, irrespective of the number of physical pixels on the sensor.
Printed resolution is related to, but separate from camera resolution, and again depends largely on other factors to determine quality.
Image quality is further compromised by the choice of in-camera processing done on images once they’ve been captured.
Still, with all this talk of pixels, resolution and quality, the deciding element in the ultimate enjoyment of an image is its content. So many aspiring photographers get caught up in the megapixel race they become self-confessed equipment junkies, unable to express themselves without the latest-and-greatest in hand. A technically perfect image captured at the highest possible resolution but with a poor choice of content will usually be frowned upon next to a less-than-perfect low resolution image captured by a skilful photographer with an eye for his subject.
While it’s true that, with digital cameras at least, equipment plays a larger role in the production of quality images than with film cameras (film is camera-independent, sensors are not), the photographer is still the key ingredient in the mix. Without him – his vision, technique and understanding of the elements that make up a successful photograph – we’d end up with millions of bytes of well exposed, high resolution, neatly printed garbage. Even if you disregard everything else, just consider this: think less about the camera you have and more about the way you use it. Your photography will be better for it.