Larger Sensor, Better Photos

DSLR cameras are a natural choice for anyone who wants to take the highest-quality pictures because they offer so much more than digicams in both performance and functionality. This is why a DSLR is the primary choice for a professional photographer. But everyday photographers who care about picture quality can also benefit from the following features of a DSLR:
* Plenty of user-adjustable controls, especially for lens aperture settings;
* Interchangeable lenses that cover a wide range of focal lengths;
* Faster operation and response times;
* Superior low light performance at high ISO settings
* A larger, brighter and more accurate viewfinder.
All entry-level DSLR cameras (and some ‘pro-sumer’ models) include a fully-automatic shooting mode for novice photographers. Many also come with a selection of scene modes that make it easy to obtain correctly-exposed pictures with common subject types, such as portraits, landscape shots, night shots and sporting action. You also get a full range of user-adjustable controls on even the most basic DSLR camera.
However, there are a few general factors to take into account before selecting a camera for your own use. In this chapter we look at the influences sensor size and resolution, image noise and dust removal can have on a camera’s performance and suitability and outline some of the factors you should consider before choosing a DSLR camera.
DSLR Sensors
Although a high-end digicam may offer the same megapixel resolution as a DSLR camera, the individual light-capturing photosites on a DSLR’s sensor are usually four to six times larger than those in a digicam’s sensor. The DSLR will, therefore, have better imaging capabilities. The diagram below shows just how wide these differences can be.

Even among DSLR cameras, sensor sizes vary and there are three commonly-used sensor sizes, shown in the diagram below.

The larger the image sensor with respect to the number of photosites on it, the higher its potential light-capturing ability. More light gives the camera’s image processing system more information to work with. Consequently, the camera can record a wider range of tones and reproduce colours more accurately than a compact digicam. It will also produce sharper and less grainy-looking pictures in dim lighting.
Unlike digicams, which almost universally operate with CCD (charge-coupled device) imagers, the sensors used in DSLR cameras can be one of two types: CCD or CMOS (complementary metal oxide semiconductor). Both types can deliver good picture quality, although CMOS chips offer lower power consumption.
CMOS imagers are favoured by manufacturers of high-end professional cameras because they can be made with more ‘camera’ functionality on the actual sensor chip. This makes it easier to combine high resolution with superior light-capturing capabilities. These sensors can, therefore, record digital images with a wider range of tones (from highlights to shadows), smoother gradations of colour, more accurate hues and lower image noise.
Sensor Dust Removal
Almost every DSLR camera released in recent times has come with some kind of sensor dust removal system. This technology is essential because dust that falls on the low-band-pass filter in front of the sensor can produce unattractive blotches on digital images. An example is shown below, with the dust marks circled in red.

The most commonly-used dust removal systems vibrate the low-pass filter very rapidly to shake off the dust. Some manufacturers place a strip of adhesive material in the base of the camera body to trap the dust so it’s not redistributed each time the shutter is fired (which lifts then lowers the SLR mirror). Others create an airstream in the mirror box to carry dust away from the sensor.
Several manufacturers also coat the surface of the low-pass filter with anti-static material that repels dust and makes it less likely to lodge on the filter. They may also allow a little more distance between the filter and the sensor itself so any dust on the filter will be out-of-focus. This makes tiny dust specks less likely to be visible, while larger spots have softer edges and are easier to remove when the image is processed (see below).
Unfortunately, it’s almost inevitable that some dust will stick if lenses are changed in damp or humid conditions so a few manufacturers have a third strategy for dealing with ‘welded-on’ dust. A ‘dust deletion’ function ‘maps’ the location of dust particles and allows them to be removed digitally when the image is processed with the bundled (or, occasionally, recommended) software. A combination of these systems appears to provide the most effective way to ensure digital photos are free of dust spots.
Image Noise
Image noise, which is caused by random fluctuations in the digital signal, commonly appears as graininess in the picture. In most cases, noise can only be seen when the image is enlarged substantially – at least 200 times. There are two main types of noise: luminance (or brightness) noise and colour noise. Of the two, colour noise is more objectionable because it produces an un-natural appearance. Luminance noise is more like film ‘grain’.

The photograph above is a 25-second exposure taken with a DSLR camera using ISO 3200 sensitivity.

An enlargement of part of the above shot shows a good example of colour noise, which shows up as coloured blotches in the image structure. A single stuck pixel (white rectangle) can be seen near the left border.
Both types of noise are more visible in dark areas (shadows) than bright areas (highlights) in digital pictures because brighter regions are produced with a stronger signal, giving a higher signal-to-noise ratio. The relative amount of luminance and colour noise can vary from one camera to another. However, digicams generally have higher inherent noise levels than DSLRs because of their smaller image sensors.
Almost all cameras include noise-reduction processing, particularly for exposures longer than about one second and when high ISO settings are used. DSLR cameras usually provide separate high-ISO and long exposure noise-reduction adjustments for photographers. In most cases they can be switched on and off independently and some cameras allow the ‘strength’ of the noise reduction processing to be adjusted by the photographer.
Sometimes you may see tiny white or coloured dots, scattered randomly throughout the image. This is a type of pattern noise, which produces ‘hot’ or ‘stuck’ pixels. The pattern is repeated in all shots taken under the same conditions. Pattern noise is more common in very long exposures, particularly when the ambient temperature is high.
Most kinds of pattern noise are easily removed by a process known as ‘dark-frame subtraction’ in which the camera records two images, one with the shutter open (to record the picture) and the other with the shutter closed (to record the noise). The noise pattern is then subtracted mathematically from the image data, leaving the image noise-free. The process roughly doubles the total image recording time.

The photograph above is a 17-minute exposure taken with a DSLR camera using ISO 400 sensitivity.

An enlargement of part of the image shows the coloured dots that characterise pattern noise associated with long time exposures.
How Important is the Image Processor?
All digital photographs are created by processing the image data from the sensor. The degree of processing varies with the type of image files the camera creates. See File Formats for more information.
Camera manufacturers like to promote the image processors they fit in both their compact digicams and DSLR cameras, largely because the image processor has a major impact on the camera’s pictorial capabilities. Most manufacturers even have specific brand names, such as DiG!C, Bionz, True-Pic Turbo, Venus Engine, PRIME or Expeed for their processors.
At the heart of all image processing systems are processing algorithms. These mathematical operations convert the digital data from the sensor into coloured pixels in the digital image. Processing algorithms are constantly being improved and manufacturers often indicate when a processor is updated by adding numbers or letters to the processor name.
Each manufacturer develops processing algorithms to match the performance of the image sensors they use and details of processing algorithms are closely-guarded trade secrets. However, experienced camera users can identify the often subtle differences between one manufacturer’s processor and another’s. Certain manufacturers have a characteristic ‘look’ to their cameras’ shots that appeals to certain photographers.
Differences can be most easily seen in subjects with subtle tonalities, such as portraits and still life shots. If you like the ‘look’ of the images from one particular camera brand, it’s probably because you like the subtle adjustments that result from the image processor.
Image processors also determine the camera’s responsiveness and can influence the speed at which data is processed. In this area, some processors are quantifiably faster than others – and the processors in professional cameras are the speediest of all.
Lens Multiplier Factors
When a lens designed for a 35mm camera is used on a camera with a smaller sensor, its effective focal length changes by a ‘lens multiplier factor’ (LMF) or ‘crop factor’ that relates to the difference in the two sensor sizes. Because the smaller sensor covers a smaller area than a 35mm camera’s image circle, to find the effective focal length of any lens fitted to the camera you must multiply the 35mm focal length by the LMF.
For example, the crop factor for a camera with a 22.2 x 14.8 mm image sensor is 1.6 times. Consequently, when you fit a 50mm lens (in 35mm format) to this camera, its effective focal length changes to 50mm x 1.6, which is 80mm. Similarly, a zoom lens that would have a focal length range of 28-70mm on a 35mm camera changes its focal length to the equivalent of 44.8-112mm on the camera with the smaller sensor.
Some camera manufacturers use marginally larger sensors with a crop factor of 1.5 times. On their cameras, the 50mm lens would have a focal length equivalent to 75mm, while the 28-70mm lens would be equivalent to 42-105mm. The tiny difference in sensor size has no effect on image quality.
Manufacturers that build cameras to the specifications of the ‘Four Thirds System’ use even smaller sensors with a two times crop factor. On their DSLR bodies, a 50mm lens would be equivalent to 100mm, while a 28-70mm lens covers a focal length range equivalent to 56-140mm. Smaller sensors limit the wide-angle coverage of 35mm lenses, although photographers make some gains in effective focal length at the tele end of the zoom range.

The illustrations above simulate the influence of the focal length crop factor. The coverage of a 50mm lens on a camera with a 35mm-size image sensor is shown at the top, while the same shot taken wit a DSLR with a 1.6x crop factor is shown above.

Dynamic Range
A DSLR camera can record a wider range of tones than a digicam because its sensor has larger photosites that can capture more light. With correct exposure techniques, you should also be able to record a wider range of brightness levels in each shot and even exceed the dynamic range of film.
When a photosite receives more light than it can handle, the signal overflows into adjacent photosites, which may also saturate. This usually results in blown-out highlights (e.g. white skies or snow scenes with no detail). Some cameras may also block up shadows, especially when shots are taken in JPEG format.
To prevent over-exposure of highlights, most digital photographers set the exposure compensation (see Exposure Metering.) on their cameras to +0.3EV, raising it to +0.7EV or even +1.0EV for snow scenes with large areas of white snow. You can usually bring out shadow detail in editing software but if no detail has been recorded in the brightly-lit parts of the image, no amount of editing can put in detail that was not recorded in the first place.
The histogram display is a useful guide for setting exposures because its shape reflects the tonal distribution in the subject. When the graph comes down to zero at or near each end of the scale, the photograph has recorded the full subject tonal range and highlight and shadow areas should contain detail.

In a correctly-exposed photograph, the histogram is evenly distributed across the baseline of the graph.

When the shot is under-exposed, the histogram is pushed to the left side of the graph and shadow detail is lost.

With over-exposure, the histogram is pushed to the right side of the graph and highlight detail is lost.
To avoid blown-out highlights make sure the histogram does not touch the right hand end of the scale. Blocked-up shadows can be avoided by ensuring the graph is not biased towards the left side of the scale. Use the exposure compensation function (+/-EV) to make the required adjustments. Checking the histogram to see the extent to which the exposure should be changed.
Some DSLRs include dynamic range extension settings, which use additional processing to reduce contrast at the extremes of the brightness range, thereby pulling highlights and shadows back into the brightness range that can be recorded by the sensor.
Note: shooting raw files (see File Formats) allows you to adjust exposure levels selectively to restore highlight or shadow detail that may have been marginal at point of capture. However, you can never recover detail that was lost through over- or under-exposure.
USEFUL URLs
The following websites provide additional information on the topics covered in this article.
en.wikipedia.org/wiki/Image_sensor_format for general information on image sensors.
web.canon.jp/imaging/cmos/index-e.html for information on CMOS sensors and the advantages of ‘full-frame’ sensors.
www.four-thirds.org/ for more information on the Four Thirds System.
www.foveon.com for information about Foveon’s sensor technology

SPONSOR LINK

Canon. Advanced Simplicity. Visit canon.com.au for more details.