In Depth
Technology
Digital cameras
The complexity behind point-and-shoot simplicity
May 22, 2007
By Chad Sapieha, CBC News
An illustration of the way light passes through a digital camera lens. Note that the light moving through the digital camera lens strikes the image sensor at a nearly perpendicular angle.
Digital cameras are among the most ubiquitous of modern technologies, generating instant photos with point-and-shoot simplicity. But underneath their sleek exteriors, they're far from simple machines.
They're often — and accurately — described as computers with lenses. Most of the components inside digital cameras, from image capturing mechanisms to lens optics, are remarkably unlike what is found within traditional film cameras.
Avid photographers tend to be gearheads, taking pleasure in discussing highly technical subjects such as varying focal lengths and their influence on magnification factors. The snapshot-happy public, on the other hand, seems satisfied to remain blissfully unaware of what goes on underneath the hood of their high-tech picture poppers.
In explaining how a digital camera works, we're going to shoot for a happy medium between the two. Without getting too technical, we'll take quick peek at some of the primary elements at work, including the image sensor, image processor, lens, image compression methods, and LCD viewfinders.
The image sensor: A marvel of microscopic proportions
The primary difference between traditional cameras and digital cameras is the means by which they capture images.
An illustration of the way light passes through a traditional film-based camera lens.
Film cameras use strips of plastic covered in granular silver halide crystals that collect light. If you were to look closely at a greatly enlarged image captured by a film camera, you would see millions of tiny, irregular grains of colour or shadow. Each of these grains roughly represents an individual silver halide crystal.
Digital cameras replace the strip of crystal-coated plastic film with a high-tech image sensor. An image sensor is a seemingly flat surface composed of millions of photodiodes. These photodiodes are essentially substitutes for the individual silver halide grains found on film. Each diode collects light that will end up representing an individual pixel, or dot, in the final image.
The number of pixels necessary to approximate the amount of detail found in an image captured on standard 35mm film is up for debate. Experts in the industry suggest anywhere from four million to 20 million pixels (four megapixels to 20 megapixels) would do the trick.
Canon Canada's Neil Stephenson seemed to have the most definitive evidence to back up his claim. He said his company performed a test in which it compared the grains on an enlarged image captured with 35-mm film with the pixels of the same image shot with a 16-megapixel digital camera, and found that the two photographs were virtually indistinguishable from one another.
That said, a 16-megapixel camera is overkill for most consumers, unless they plan to blow the size of the image way up to make poster-size prints.
An illustration of the Bayer filter grid used with most digital camera image sensors.
The image sensor is paired with an important piece of hardware: a special filter.
Image sensors capture lots of tiny dots of light, but on their own they would simply register the intensity of the light, not its colour. To solve this problem, a colour filter is placed on top of the sensor.
While there are many kinds of colour filters in use in digital cameras, the most common is the Bayer filter. It's essentially a grid composed of squares representing the three primary colours — with twice as many green photosensors as red or blue, because the human eye is more sensitive to green light.
The filter assigns each pixel in the image a colour as light passes through.
Image processing: A vendor's 'value add'
With both light and colour captured by the sensor, the camera now has a set of data that can be assembled to create an image.
The information captured by an image sensor is digital — just a series of ones and zeros. This data is read and converted into a viewable picture via image processors, which consist of circuit boards and software.
Image processing not only decodes digital information to create a picture, it also manipulates the image through algorithms to achieve, for example, balanced colours and lighting. While many camera manufacturers purchase image sensors from other companies, image processors are largely unique to each vendor — their so-called "value add." Different image processors employ varying algorithms, which can deliver vastly different image results.
A comparison of successful de-mosaicing (the image with gently shifting shades of green on the left), and failed de-mosaicing (the lightly gridded image on the right).
As an example, most major camera manufacturers have their own unique "de-mosaicing" algorithm. Remember the colour filter that sits in front of the image sensor and provides a basic value for the light collected by each photodiode? Without de-mosaicing, the resulting picture would look like … well, like a mosaic — an image composed of tiny bits of basic colour.
A good de-mosaicing algorithm will look at those simple reds, blues, and greens and convert them into nuanced, lifelike shades to create a much more natural looking image. Poor de-mosaicing might result in a picture in which the colours seem slightly out of whack. In extreme cases a faint, grid-like pattern can emerge.
It's all about the perpendicular
In order to understand the difference between traditional and digital lenses, we need to dig a bit more deeply into the way both types of cameras collect light.
The silver halide crystals stuck to a strip of film create an uneven surface, and this three-dimensional quality allows them to gather light from many angles. Lenses designed for film cameras exploit this property by using concave, convex or aspheric elements to direct light towards the film at an angle.
A digital camera's image sensor, on the other hand, is essentially a flat surface and not well equipped to receive light at an angle. If light arrives at the sensor at a slant, it can result in the photo having a lack of clarity.
"Think of several shot glasses lined up in a grid," explains Olympus America's Sally Smith. "They create what appears to be a flat surface — you can set a piece of paper on top of them. But each glass has depth. The same is true of the photodiodes on an image sensor. Light needs to be guided directly into them."
That means the light passing through the primary lens of a digital camera needs to be redirected to hit the image sensor at a perpendicular angle, allowing it to funnel straight down into the photodiodes. This requirement has led to new lens optics designed specifically for digital cameras that direct light in a manner conducive to an image sensor's method of collection. Without getting into too much detail, a digital camera's lens system is made up of varying pieces of glass that redirect light to ensure it smacks the image sensor head-on.
Compression: a semi-necessary evil
Image compression is a procedure carried out by the camera's processor to reduce a picture's file size.
Flash memory cards used to store images have increased in capacity to such a degree that users can now store 1,000 or more photos on a card that costs less than $50.
Consequently, image compression's job these days isn't so much to store more photos as it is to create files that are diminutive enough to be easily managed. Smaller files can be transferred from one device to another more quickly and are easier to work with in photo editing applications.
A head-on, enlarged view of an eight-megapixel digital camera image sensor. (Courtesy Canon Canada)
Compression essentially deletes pixels that are of similar but not exactly the same colour and replaces them with a single colour, resulting in less information and smaller file sizes. Most cameras use JPEG compression, which, when properly applied, results in a more compact file with a relatively minor amount of lost visual detail.
However, there are varying levels of compression available within the JPEG format, hence the selectable file sizes in your camera's settings menu.
Kodak's Brian Fox says a simple way to understand varying levels of compression is to start by thinking of an image with a clear blue sky.
"If you were to choose a very high level of compression, all of the subtle hues of blue in the sky might simply be replaced by a single shade repeated over and over again," said Fox. A low level of compression, on the other hand — or what Olympus' Smith refers to as "intelligent compression" — would retain far more of the sky's blue nuances, replacing smaller groups of nearly identical pixels here and there with a single colour.
Still, some detail is always lost when an image is compressed.
And that's why some avid photographers prefer to shoot in uncompressed formats such as RAW and TIFF, which preserve every pixel of data captured by an image sensor. The files are larger, but image quality is left completely intact.
View screen vs. viewfinder
The LCD screens on the backs of digital cameras have fundamentally changed the way people take pictures.
A cutaway view of the internal design of of Canon’s EOS 1 Mark III digital SLR camera, showing the viewfinder mirror system, image sensor and circuitry. (Courtesy Canon Canada)
Rather than forcing photographers to position an optical viewfinder directly in front of their eyes, digital photographers can position the camera well away from their faces, which makes framing a picture above a crowd, for example, much easier.
And, thanks to the fact that the picture on the display reflects exactly what is seen by the image sensor, LCDs provide 100-per-cent frame accuracy. In other words, what you see before you press the shutter button is what you get in the final image.
By contrast, the viewfinders on most non-professional film cameras — which provide previews using either a simple window above the lens with a direct view of the subject or a mirror assembly that directs light from the lens up to a closed window — typically provide a frame accuracy of 95 per cent or less.
This could potentially result in, say, in a portrait with the top of the subject's head cut off.
However, many serious digital photographers still prefer viewfinders to LCDs. One reason is that a viewfinder image is optical as opposed to digital. An optical image delivers maximum detail, consequently giving the shooter an extremely clear look at the subject being framed, including tiny visual details that may not be viewable on the 200,000 or so pixels that make up most camera LCDs.
But in the end it could simply be a rare case of old-school camera technology being more comfortable and intuitive than digital. All of the sources interviewed for this story — each one an avid photographer — said that they prefer using a viewfinder to an LCD.
As Canon's Stephenson puts it, "It just feels like you have more control. It sounds funny, but when you bring the camera up to your face you become one with it. That feeling doesn't happen with an LCD."
Menu
- Main page: How it works
- Bluetooth
- Broadband
- Cell phones
- Digital cameras
- DNA
- LCDs
- Noise-cancelling headphones
- Ornithopter
- Surround sound
- Thermobaric weapons
- VoIP
- Wi-Fi hotspots