Businesses such as estate agencies, insurance firms and newspaper publishers are finding that the ability to process an image in an instant, incorporate it into digital media and send it around the world in seconds are opening up new applications. Investigators can record accident scenes, house specifications can be put onto the Web in minutes, and trainers can incorporate images from a seminar into a presentation on the same day.
At the heart of the digital camera is a charge-coupled device (CCD - see box on p86). The resolution of this chip defines the detail the camera captures. For most digital cameras this runs into millions of pixels (megapixels). A photograph visually comparable with the output of a conventional film-based camera requires about 1.5 megapixels but professional models deliver up to four times this resolution.
It is tempting to go for the highest specification available but price follows pixels, making it more sensible to match camera to application. A bargain-basement camera on a petty-cash outlay is fine for a 320x240 pixel Web site image but a picture for publication in a magazine has to have a minimum resolution of 300 dots per inch, which means that for a reproduction size of 5x7in the image would have to be at least 2,100x1,500 pixels.
A second, near-essential feature that influences price is optical zoom, using a traditional zoom lens to control the area included in the shot. Most cameras also offer digital zoom, but this is of dubious benefit because the camera merely electronically magnifies part of the image. Unlike optical zoom, there is no increase in information captured, it is just like looking at the output with a magnifying glass.
Some cameras minimise the bad effects by adding interpolation or resampling to enhance the image. Here the processor sets the pixels between those captured at an intermediate value, smudging one into the next. This reduces jagged edges, but doesn't add useful information - the result looks more natural, but there's no increase in detail.
Once the image is captured it is held in flash memory. This non-volatile storage is usually supplied on small, removable cards a few centimetres in diameter ranging in capacity from 4Mbytes to 128Mbytes in Smartmedia or Compact Flash format - though a few cameras use laptop-style PC cards.
The number of pictures stored on these cards cannot be specified exactly. Typically, a mid-range 1,280x960 pixel image might take up 300Kbytes, fitting 50 images onto 16Mbytes, but compression varies with the nature of the image, so the file could be anything from 100Kbytes to 500Kbytes. The highest quality images currently fill more than 16Mbytes, which seriously limits the number that can be stored on memory cards.
However, IBM now produces a 340Mbyte hard disc that fits a Compact Flash slot, giving sensible capacity for those needing ultra-high quality output.
An alternative is transferring files to a laptop, which also helps when checking images on location - there is a limit to what can be seen on a tiny LCD. Most cameras link via a serial port, and many now include a faster USB connection. Flash memory cards can also be slipped into a converter and plugged into a PC card port, giving high speed, disc-like access.
Battery power is a second capacity issue. Digital cameras are power-hungry, making it possible for the batteries to give out before the memory space is filled. This will certainly happen with conventional alkaline batteries. Most cameras now use rechargeable cells:
Bear in mind that all rechargeables lose power even when not in use. A NiMH, for instance, will lose 5% of its charge in a day.
Compared to conventional photographs, today's digital technology has limitations, particularly in applications that require proof of authenticity. Digital images are notoriously easy to manipulate, whether to change the content of the image or to modify the date and time the exposure was made.
These changes can be undetectable, which means that digital images currently have to be used with care in court cases because their veracity can be easily disputed. Tamper-proof certification could be built into the image to prevent alterations being made but camera manufacturers have yet to implement such safeguards.
Even so, speed of turnround, easy transmission and reusable media make digital photography a superb business tool.
For more on peripherals see www.itnetwork.com
Adjusting to light and colour
Contrast and hue adjustments within an image are essential to ensure that the final picture accurately reflects the viewers' perceptions. The eye compensates for huge differences in lighting levels. Sunlight may be 100 times brighter than office lighting but we only really notice this when emerging from an artificially-lit room on a sunny day, after a while the eye and the brain acclimatise to the new condition.
In a similar way, we also even out colour differences between natural and artificial light. White looks white to us whether under yellow-tinted tungsten light or blue-tinted daylight but a camera has to identify what hue "white" has under the ambient lighting so that it can modify all colours in the image accordingly.
Most digital cameras do this automatically by making colour comparisons across the whole image but this technique can be fooled, for example, where one colour dominates. The best cameras provide a manual white balance reading by measuring the output from a standardised white card before the actual picture is taken.
How the CCD records the image
The surface of a charge-coupled device (CCD) is divided into a grid of pixel units. Each unit is a tiny photoelectric detector. As a photon of light hits the unit, a small electrical charge is stored in it. Subsequent photons increase the charge, registering the light level for that area of the picture, making no distinction between colours.
Colour information is compiled by filtering out all but one colour for each pixel. This is done by dyeing the surface of the chip in a chequerboard of the three primary additive colours - red, green and blue - so that each pixel records a specific colour. Because the eye is not evenly sensitive to the colours, there are usually significantly more green pixels than blue or red.
At the start of the photographic process, the CCD contains no electrical charge until it is exposed to light by the opening of a shutter. The system then moves row by row to read the charge level in each pixel unit. The charge levels produced are tiny and each analogue value has to be amplified before being passed to a digital converter. From there they pass through a digital signal processor to adjust contrast and hue and finally the image is compressed, typically using JPeg format. This process accounts for the marked delay before the captured image appears on the camera's LCD.
The resolution of a CCD is noticeably less than the size of the array of pixel units. A CCD with 2.1 megapixels produces a picture of about 1.9 megapixels, partly because the image is cropped to a 4:3 ratio, such as 1,600x1,200 pixels, and partly because a few rows and columns are dyed black to set a "black level" to compare the result of the active pixels against. Manufacturers refer to the smaller number as the 'effective pixels' value.
This was first published in November 2000