The Digital SLR

Jorge DaSilva explains the choices...

Jorge DaSilva by Robert Lansdale

Jorge has many years experience as a professional wedding photographer. Seven years ago, he joined Henry’s where he is currently Dean of their School of Imaging. When the digital imaging market began to form, Henry’s offered casual training in the new technologies. Two years ago, demand for digital training surged. and Henry’s responded with the School of Imaging which has a formal program of half day courses in both seminar and hands-on workshop formats. Jorge based tonight’s talk on a Henry's half-day seminar called the “Digital SLR & Scanning Primer”.

When deciding what camera or scanner you wish to purchase, understanding the claims made by digital camera manufacturers can be a dizzying experience. In the digital world, the specifications and models are changing rapidly. With all the new chips, lenses and software, it is hard to make a confident decision. Jorge put us through a fast- paced lecture hitting the broader issues in selecting digital cameras, including the buzz-words and their significance.

Engines. Jorge suggested we consider the digital camera a tool with two ‘engines’ - a camera engine, with the traditional controls for exposure/focus/focal length and a digital image processing engine consisting of the image sensor/CPU/software/memory/display. Briefly, the sensor replaces the film, the CPU & software translate the signals from the sensor and control the camera functions, the memory stores the images captured by the sensor, and the display is a viewing screen -- the instant feedback is a big factor in the popularity of digital technology.

Jorge began with a comparison between the consumer level ‘point & shoot’ cameras and the more expensive digital SLR (DSLR) cameras. The cameras in both categories have reached a point where the resolution (around 6 megapixels) exceeds the needs of most users. Emphasis is now usability and image quality.

Megapixels and Resolution. The digital camera's sensor determines the size of print you can make.

As a rule of thumb, a 1 megapixel camera can be used to make reasonable 4x6 inch prints, a 2 megapixel camera can print as large as 8x10 inches with reasonable detail while 3 megapixel cameras - almost the low end today - make good prints up to 10x12 inches.

Professional cameras with sensor resolution of 12 megapixel and greater are challenging traditional film for print detail and quality.

Left DSLR - Right P&S
Left DSLR - Right P&S

Noise & Dynamic Range. The better image quality of the DSLRs stems partly from the use of physically larger sensors. These result in a wider dynamic range with better shadow and highlight detail combined with less noise. The smaller sensors on consumer cameras can suffer blown-our highlights and blocked up shadows.

Digital circuitry is prone to random noise, especially in low light shots. This noise shows up in images as a 'coloured confetti' in large areas with an even tonal range. The more sensitive larger sensors have a better signal to noise ratio in low light situations resulting in an obviously better image as Jorge demonstrated with a number of comparison slides. Note: The examples shown here are not as clear as Jorge's presentation slides.

Dust. The consumer cameras with a permanently mounted lens are sealed from dust. The DSLRs offer interchangeable lens -- and direct access to the sensor by dust. The static charge created by the sensor acts like a 'dust magnet'. The dust can accumulate to the point were it interferes with image formation. Jorge recommended professional cleaning rather than risking damage to the delicate sensor by taking a DIY approach.

rays of traditional camera lensparallel rays lens for digital

Lens Compatibility. Can we expect the sensor to reach a standard size matching the size of a 35mm negative, allowing traditional 35mm lenses to be used on the DSLR?. This may not happen. The sensor is a three dimensional device -- picture an array of tiny deep cells collecting light 'particles'. If the light hitting the sensor is not a parallel beam, the cells at the edges will receive proportionally less light than the centre cells. To avoid this falloff, a digital camera lens is designed to produce light rays that are parallel to the lens axis. Film being flat, the usual spread of the rays in a traditional lens has little effect on the evenness of light hitting the film. Since the sensors are an integral part of the camera unlike film, there isn’t the same need for standardization. The current crop of DSLRs have a range of sensor sizes equal to or smaller than a 35mm negative, but larger that the sensor in a point & shoot.

Sensors. Sensors consist of an array of light sensitive cells. Most DSLRs use CCD or CMOS sensors that are a mosaic like the autochrome plates of a century ago with a tiny colour filter over each cell (pixel). Fuji uses hexagonal cells instead of the traditional rectangular pixels allowing an 50% greater effective resolution. Fuji have expanded on this with a design in which each hexagonal cell contains two light sensitive cells - one large and one small which they claim gives better shadow and highlight details. The overflow of highlight information is captured by the secondary cell and combined with the information from the main cell to create a wider dynamic range. Another unusual sensor is the Foveon CMOS which is like the modern day tri-pack colour film. The Foveon's sensor has three separate layers, each one sensitive to a different colour of light (RGB). To date, the Foveon has been adopted only by Sigma, a relatively minor player in the DSLR market. DALSA Corporation, makers of digital imaging technology, have a comparison of CCD vs. CMOS sensor technology on their site.

Dynamic Range is a measure of the brightest to darkest light values the sensor can capture. The figure of 2 (for prints), is a logrithmic value so the ratio is actually 100:1 (from brightest to darkest), 3 is 1,000:1, 4 is 10,000. In traditional photography this is the H&D curve.

The small-sensor consumer cameras have a lower dynamic range resulting in a tendency to block up shadows and blow out highlights when capturing a very contrasty scene.

Comparing the CCD Mosaic and the Foveon sensors

File Format. Traditionally digital cameras and scanners store images in JPEG or TIFF files. The JPEG, with lossy compression (some information is considered redundant and discarded during the compression process), creates a smaller file while the TIFF, with lossless compression, creates larger files but doesn’t lose image detail or create artifacts (noise). Both formats create 8 bit colour images – sufficient for human vision, but compromised by subsequent correction in Photoshop or other image processing software. Low end consumer cameras offer only JPEG files.

The image information from the sensor is recorded in a unique format in the camera before it is converted to a JPEG or TIFF file. This RAW format varies by manufacturer. The information has no preprocessing of any sort applied to it in the camera. It consists of uncompressed 12 bit colour. Recent DSLRs and higher end point & shoot cameras have offered the ability to output files in RAW format. These files are smaller than TIFF files, yet contain more image information. The human eye cannot differentiate between 8 and 12 bit colour images; however, the added bits allow image correction without loss of visible detail. The RAW files can be read and modified only by high end programs such as Photoshop. RAW format can be read but not saved by image software and is considered the same as a film negative – not easily manipulated for unlawful purposes.

The effect of image correction on the image detail was illustrated using the Levels palette in Photoshop on a low contrast image. The original image was mostly mid-tones with little shadow/highlight information. Using Photoshop, the dynamic range was expanded to give image content from deep shadows to bright highlights. The result improved the image but the adjusted Levels graph now shows a ‘comb’ effect with missing tones throughout the light levels. Jorge noted that the missing tones lead to posterization of portions of an image. Correcting an image in 16 bit mode, flattening the layers, and then saving it in 8 bit eliminates the comb effect and associated image degradation.

Olympus E-1Nikon D70Canon Digital Rebel

DSLR Cameras. A comparison of current DSLR cameras shows the efforts of various manufacturers to differentiate their products by lens design, sensor size and type, operation speed, ergonomics, body construction, colour space, light sensitivity, etc. The examples shown by Jorge had sensors rangeing from 4 to 6 megapixels in resolution. The high end professional cameras offer even larger sensors with greater sensitivity and higher resolution.

Film and Print Scanners. Jorge wrapped up with a brief discussion of scanners - high end flat bed models and the specialized negative scanners. He gave me the impression that flat-bed vs. negative scanners was much like consumer vs. DSLR cameras where the dynamic range, low image noise levels, and operational speed favour the negative scanner over the flat-bed transparency scanner.

Why Use Negatives? Negatives and slides can have a dynamic range of as much as 3+ to 4+ compared to prints with a range of 2. Thus, it's best to start with an original negative rather than a photographic print. Your scans will capture more image information and result in a better digital print.

Further Education. Anyone wishing to learn more about the various aspects of digital photography can sign up for evening or half-day courses at Henry’s School of Imaging. Visit Henry's school website, or visit the store and pick up a brochure.

All screen images were taken with a Nikon Coolpix 990 digital camera and adjusted in Photoshop CS. The images are © 2004 The Photographic Historical Society of Canada unless otherwise noted. The content of the presentation images is © 2004 Henry's.

This presentation gives a feeling of history in the making as once again the art of photography takes a major change in technological direction, one that after 165 years, moves the industry away from its roots in the effect of light on silver halides. Questions? Please contact me at

Robert Carter

Back to Past Programs

return to the home page
Main Index
Facelift & Design © 1999 Zero Cattle
Page © 2004 by The Photographic Historical Society of Canada
Webmaster: Bob Carter
-- See What's New for more details

Lost?   Find your way with our Site Map!