This guide was last revised 3 June 2009
Unless digital content is being created for one-time use and will be discarded, good practice should be to design and format your content for different environments and uses over time, even if they are not immediately known. The ability to re-format and re-purpose are key strengths of digital content given how easily it can be copied and distributed. The choices you make at the point of creation, in particular the format that your content is created in, greatly affects how useful and long-lived your content will be.
Digital technologies are sometimes said to be ‘disruptive’ technologies because of the way they challenge or alter long-standing practices. Rapid growth in affordable digital imaging, software and storage has made entry into publishing, photography, and audio and video production possible without any professional training or background. That same technology has for the first time made possible mass-digitisation projects driven by the likes of Google and Microsoft.
A consequence of this rapid change has been a disruption to the development of agreed practices and standards designed to ensure the outputs of these technologies would be long-lived. In fields such as text and digital audio, the practices and standards are relatively mature due to the relative maturity of the technologies used. In photography and image scanning, technology has only recently reached a point of maturity where practices and standards can be consistently applied. In digital video, the technology at present is evolving faster than consistent practice and standards, resulting in a mixture of formats and standards in common use, and ongoing issues with the huge storage requirements for high resolution video.
If your business or content creation activity depends on having content that can be used over a reasonable period of time, you need to anticipate that some technologies and standards will become obsolete. If you are looking for good or best practice, you also need to refer to guidance that is current and addresses recent developments. To achieve this, look for technology hardware and software that uses open standards and for guidance on best practice that is being maintained and has been updated within the last three years.
Good sources of best practice advice can often be found on websites of professional associations, or organisations that specialise in long-term management or storage of digital content (such as libraries and archives). Often this advice however is aimed at professionals using some of the highest quality hardware and software available. If these benchmarks for equipment, training and standards are beyond your reach, that does not mean there is no point in trying to follow them. You will find that the most useful guidance identifies minimum standards and practices, while also pointing to the best practice. Minimum standards and practices are a really good place to start.
Because of the different media and content types involved, we have split this guide into three main sections:
Each section provides an overview of the kind of technology involved and the approach to practice and minimum standards we recommend.
One of the most common forms of digital content creation is through use of a scanner or camera to take digital images. Digital photography is a rapidly maturing field, but has some importance differences from film-based photography, not the least of which is that pre-print images are usually viewed through a computer monitor. These differences can greatly affect hardware choices, workflow and formats in creating a usable digital image. This section provides an overview of digital still image technologies and some of the considerations that may affect your hardware and workflow choices.
Film and digital cameras differ significantly in the way they capture images. Print film is made of polyester coated in a light sensitive silver-halide emulsion. The size of the silver halide salts determines the light sensitivity and resolution (or grain) of the film, with superfine film allowing more fine detail but less light sensitivity. Colour film has at least three layers of emulsion, containing red, green and blue dyes. Exposure to light through a lens creates a negative image that is then used for printing (slide film is different in that it produces a positive image designed for direct viewing through a light projector). Different size film formats also affect the resolution of the image, with medium and large format film generally providing more detail than standard 35mm film.
A digital camera has an image sensor made up of millions of light sensitive pixels (picture elements), with the size and quality of the sensor affecting detail, colour and noise levels, while the number of pixels determines the native resolution. A chip within the camera processes the information captured by the sensor to create an image recorded to a storage device, such as a flash memory card. Higher-end digital cameras will allow the image to be stored unprocessed (raw) for later processing in software.
Film grain (left) and pixels compared
In both film and digital photography, the function and quality of the lens used generally has more impact on the clarity final image than the grain of the film or the number of pixels - a poor quality lens will not allow enough detail to reach the film or sensor for it to be visible in the final image. This makes the lens optics potentially the most important consideration when selecting a camera to use for accurate digitisation or detailed digital photography.
Until quite recently, storage capacity for digital cameras and computer hard drives was a significant factor in determining the quality of image coming from a digital camera or scanner. Digital camera manufacturers almost all used image compression to create smaller JPEG image files to reduce file storage costs. Scanners, while having capacity to capture uncompressed TIFF images, still often have the compressed JPEG image as their default setting.
Image file compression is a technology adopted from the computer industry. Image files such as an uncompressed TIFF or a raw digital camera image are generally large, which is a limitation when storage is expensive. It is also a limitation when files are being transferred over a network such as email or the internet.
There are two types of compression, lossless and lossy. A well-known example of lossless compression is the Windows Zip file compression, where a file can be restored to its identical state and size after unzipping, and no information is lost or altered. For an image, consecutive pixel values containing no information may be recalculated (e.g. “0000000” can be represented as “7x0”) and restored when decompressed. Lossy formats such as GIF and JPEG, on the other hand, were originally designed for easy viewing on computer screens and transfer over networks. Lossy compression averages, or interpolates, the difference between two or more bits of information, and discards those bits for one of average value. The discarded bits are lost, and can never be restored to the original, even if an attempt is made to resize the file. This is compounded whenever a lossy file is edited or saved, as each time more bits are averaged and discarded.
Example of detail loss through 'lossy' compression
In simple terms, lossless compression is fully reversible, while lossy compression is irreversible. Lossy compression should be avoided where high image quality and accurate reproduction is desired, such as the creation of reference or archival images, images that may be edited, retouched or cropped, or images that will have derivative copies made of them. Raw digital camera files and raw, uncompressed or lossless compressed scanner files will always offer the best output for image quality, accuracy and editing.
In relation to content creation and usage, arguably the biggest difference between a non-digital and a digital object is that the digital object requires a machine to view or use it (the same can be said of electronic content such as magnetic tape). As such the hardware used to create and view digital content becomes particularly important, as it can greatly affect our perception of what we see or hear.
Anyone who has seen a bank of flat screen televisions in an electronics store will have experienced the variations in image between different screens, including those with the same make and model. Those same variations occur between digital cameras, scanners and computer monitors, and will directly affect what different users see as well as what might be output to a printer. Calibration and image management techniques are used to help overcome these differences.
Understanding the basic differences between the way colour works with light and with print is essential if digital images are likely to ever be made into prints or used in a printed resource.
Cameras, scanners and monitors all create colour with light. Red, green and blue are the primary colours for these devices, and when overlaid produce white. With print, colour is created with pigments, the primary colours being cyan, magenta and yellow (black is added as a separate ink to improve contrast). Overlaying these colours create darker shades. Viewing digital images on a monitor or projector transmits coloured light to your eyes, while viewing printed materials involves reflecting light off the colours. The number of colours able to be transmitted by a monitor is significantly greater than the colours possible with pigment inks, which means that a printed digital image is likely to be less dynamic in its colours than the same image seen on a monitor.
One means of managing these differences is to assign a ‘colour space’ to work in (such as ProPhoto, Adobe RGB or sRGB), which identifies the colour origins of the image and allows it to be translated for different environments. Calibrating and profiling different cameras, scanners and monitors to aid accurate representations of images is also a simple but important part of image workflow to ensure that what you see on your screen can be seen by others whether on screen or in print. Colour or greyscale charts or targets included with a scanned image can also assist with recreating the original colour tones of an image when being printed.
There is a vast number of guides and advice available in bookshops and on the internet around choosing and using digital cameras for image capture. Perhaps the most useful thing to be aware of is to choose a camera that is fit for purpose.
Point and shoot digital cameras seldom have the kind of sensor or lens quality of Digital SLR (Single Lens Reflex) cameras, which is most important for image detail, even if the pixel count is the same or higher. Choice of lens will affect factors like image distortion, which is particularly noticeable at image corners when taking photographs close up or at full zoom.
Many cheaper digital cameras have difficulty capturing clear images under low light or indoors, while almost all indoor images can be improved with some attention to lighting and use of a stand alone camera flash instead of a built-in flash. Using a camera mount or tripod will improve the sharpness of images and allow longer exposures.
Only a few high end point and shoot cameras currently support raw image saving, while all Digital SLRs do. Saving images in raw format commits you to processing each image in software, but can make it easier to preserve the original settings an image was shot under. Ensuring camera time and date settings are correct will also help with later filing, while some cameras now are GPS enabled, which is useful identifying information if your photography is location-specific.
All digital scanners are effectively digital cameras, and are subject to the same kinds of features and limitations. The sensor quality, native resolution and lens quality all determine how capable a scanner is in reproducing an image – a cheap scanner will have a lower image quality in the same way a cheap digital camera does. As most scanners are controlled by a computer, the software settings used are also a key factor. In contrast to cameras, scanner lighting is always direct and artificial, making the quality and evenness of the lighting also important to image quality.
While flat bed scanners with film adaptors have improved greatly in the last couple of years in the quality of their optics, sensors and mounts, dedicated film and slide scanners will more readily provide a better and faster result from scanning film. Film scanners tend to have a larger sensor that can capture more image detail, while their dedicated lens makes a properly focused image a more likely result. Film scanners however are more expensive than flat bed scanners, so the volumes involved over time may need to be traded off against image accuracy. An important factor in choosing a scanner is knowing the quality of the film being scanned (superfine grain film and slide film will offer more detail to be captured if shot with a good lens) and what will be done with the scanned output (e.g. small prints or viewing on screen versus high quality prints for exhibition).
Drum scanners are still widely considered the highest quality for scanning reflective materials such as photographs and text. However drum scanners are becoming less favoured due to their continued high cost and relative improvements in flat-bed and dedicated book and document scanning technology.
Many reflective materials are not as demanding for scanning as film due to their lower information density, however other factors such as fragility and large size of the materials may pose practical difficulties. Guides on how to prevent damage to original materials should be consulted to ensure the hardware being used is suitable. An appropriately mounted camera may be more suitable for some tasks.
Most scanner software have settings that allow the user to manually adjust exposure and black and white points in the scanner hardware. These are important settings to check, as detail in image highlights and shadows can be obliterated by the default settings of a scanner, result in a poor quality image. Other settings such as colour correction, dust reduction and sharpening should be used sparingly. Images taken from printed documents may need a de-screening filter in order to produce a viewable image.
In the early days of digital cameras and scanners, pixel resolution was a major feature used by manufacturers to sell their equipment. The biggest claims were usually made using interpolated resolution (or digital zoom) rather than optical resolution. It is evident today however that increasing pixel resolution does little to improve image information and detail beyond a certain point, usually one limited by sensor and lens quality.
Megapixel ratings are the most common means of describing digital camera resolution. Megapixels are a spatial resolution – that is they are calculated by multiplying the horizontal and vertical numbers of pixels covering a digital sensor. Film cameras, being analogue, do not have an equivalent resolution as such, but various estimates place the density of information able to be captured on 35mm film at a maximum of about 25 megapixels, and more realistically at about 10 megapixels. A good Digital SLR at a lower megapixel rating can however beat an average film camera by having a better lens or good image processing.
Due to a legacy in the printing industry, DPI (dots per inch) is the most common term for describing scanner resolution. However PPI (pixels per inch) is a more accurate term, given sensors and images are made of pixels not dots. Fortunately the measurement used for both DPI and PPI are the same as most scanner software still refers to DPI.
There are two types of pixel resolutions used in relation to scanners – one is the resolution the scanner hardware itself is sampling at (e.g. 300ppi, 600ppi, 1200ppi, 2400ppi), and the second is the number of pixels on longest dimension of the resulting image. Most scanner software allows you to see both. Practice is divided over which setting takes precedence, but both are relevant.
Best practice often recommends that a scanner be set at its native resolution – being its highest optical resolution – or at a setting equally divides into it to minimise interpolation (see table below for settings for common scanner optical resolutions). Scanners should not be set above the optical resolution even if the software allows it, as all those settings will be poorly interpolated.
|
6400ppi optical scanner |
4800ppi optical scanner |
4000ppi optical scanner |
3600ppi optical scanner |
2400ppi optical scanner |
|
6400 |
4800 |
4000 |
3600 |
2400 |
|
3200 |
2400 |
2000 |
1800 |
1200 |
|
1600 |
1600 |
1000 |
1200 |
800 |
|
1280 |
1200 |
800 |
900 |
600 |
|
400 |
960 |
500 |
720 |
480 |
|
320 |
800 |
400 |
600 |
400 |
|
600 |
450 |
300 |
||
|
480 |
400 |
|||
|
400 |
360 |
|||
|
320 |
||||
|
300 |
Sampling settings that minimise scanner interpolation
The desired size of the output image can be determined by calculating the required pixels, and the scanner set to the closest higher sampling resolution that will achieve it. For photographic images we recommend a minimum setting based on a photographic print output of 300ppi – approximately the number of pixels required to create a smooth detailed image for a person viewing at a distance of 8-10 inches or 200-250 mm (note the dpi or lpi - lines per inch - for printer output is not relevant to this calculation). This requires you to do a little maths:
| 300ppi x length (desired print output on longest side) length (original image longest side excluding borders) |
Note that it is important to use a ruler to measure your image size – for example, 35mm film typically creates an image 36mm x 24mm, or 1.417” x 0.945”. Using the formula above we can arrive at a setting for creating an 8”x10” print off a 35mm film (for those not familiar with inches, the conversion from metric is 1” (inch) = 25.4mm):
 |
In this example, a film scanner with a 4000ppi optical resolution should be set to its highest resolution of 4000ppi. A 4800ppi scanner with a slide adaptor would be set to 2400ppi.
A recommended best practice setting will use 400ppi in the formula above. This will produce the equivalent to the archival standard used by the U.S. National Archives and Records Administration Technical Guidelines.
LCD and CRT computer monitors also use pixels in their resolution, and like cameras and scanners, they have native resolutions that provide the sharpest image. CRT monitors are generally no longer commercially manufactured outside of specialist high-end markets such as medical equipment, so if you are scanning or photographing for the screen, LCD resolutions are what you should be targeting. Unfortunately at present there are a wide range of screen resolutions in use. Some of the most common ones are shown in the table below, with the 16:10 aspect being the fastest growing in desktop and notebook hardware.
|
Desktop/Notebook widescreen (16:10 aspect) |
|
2560x1600 |
|
1920x1200 |
|
1440x900 |
|
1280x800 |
|
Desktop standard (5:4 aspect) |
|
1280x1024 |
|
Netbook (16:9 aspect) |
|
1024x600 |
|
Smart phone, photoframe (5:3 aspect) |
|
800x480 |
|
Mobile (4:3 aspect) |
|
480x320 |
|
320x240 |
Common LCD screen settings
One mistaken practice when creating digital images to be viewed on screen is to refer to the 72 or 96 DPI measure that is used by the operating software to scale icons and fonts. This measure has nothing to do with the resolution of your digital image file (or printer dots) and should be ignored in any image calculations, whether made for the screen or not.
Digital copying of analogue content almost always involves some compromise in terms of loss of information, and scanning of print film and slides is no exception. Unless badly damaged, scanning a negative can provide a far superior image to that obtainable from a print, due to the limits of picture information possible on paper. The greater density of information on film negatives and slides also means a much higher resolution scan is required than for paper to reveal detail than for scanning a print. Knowing something about the film stock being scanned can also greatly assist determining the appropriate scanner settings to use.
When copying is at the very basis of most digital technology today, and when hundreds of identical copies can be made with little effort, and thousands of photographs taken at little cost, what value does a digital original have?
Most image workflows used by professional photographers and by archivists make an early distinction between the working image designed for editing or access and the original master image it was made from.
The master image is never edited, and serves as the film equivalent of a negative – a reference version that future copies can be made from. Indeed, raw digital image files from digital cameras are often referred to as digital negatives for this very reason, as they contain all the original information captured. Reformatting these images, unless done with care, can risk losing image detail and descriptive information that can never be recovered. Master files are also often larger and more difficult to work with than derivative files sized and formatted for editing.
As a matter of good practice, a working copy of an original file should be made as early as possible, and renamed using a different naming convention from the original. Originals should ideally be filed and backed up separately from copies. The assumption here is that identical digital photographs can never be retaken and that the effort or cost required in rescanning images or film is greater than that required to make digital copies and back up originals.
In order to create digital objects that are accurate copies of an original and able to be re-purposed, good practice is to create a digital master, akin to a negative. Lower resolution copies for specific purposes can then be made from the master.
Scanning black and white images in full colour allows tinting, discolouration and any markings on the image to be more clearly visible, while improving the dynamic range of greys available for revealing detail.
|
Minimum (safe) |
Best practice |
|
Bit depth |
24-bit RGB (8-bit per channel) capture |
48-bit RGB (16-bit per channel) capture |
|
Capture Format |
Uncompressed TIFF |
Uncompressed TIFF or JPEG2000 |
|
Colour space |
sRGB |
Adobe 1998 (colour) |
|
Capture Resolution |
300ppi x output length original length |
400ppi x output length original length |
The output from a digital camera is dependent on its capabilities. Generally wherever possible the camera’s highest settings should be used to capture as much detail and colour information as possible. If a camera does not support raw image output, the highest detail setting for JPEG is the safest alternative for creating an image that can be re-purposed.
|
Minimum (safe) |
Best practice |
|
|
Bit depth
|
24-bit RGB (8-bit per channel) capture |
48-bit RGB (16-bit per channel) capture |
|
Capture Format
|
Full resolution JPEG (Fine or Superfine setting) |
Camera RAW or Adobe DNG |
|
Colour space
|
sRGB |
Adobe 1998 (colour) |
|
Capture Resolution
|
Minimum of 6 megapixels |
Minimum of 10 megapixels |
Practices and standards for digitising text are some of earliest developed, due to the legacy of microfilming and document scanning technologies.
This section is currently under development. If you have expertise on digitisation find out how to contribute.
In the meantime, please check our Digitisation Resources for advice, or post a question on our Ask a Question forum
While practices and standards for digital audio are relatively mature, digital video technology continues to develop rapidly, creating challenges for long-term digital formatting and storage.
This section is currently under development. If you have expertise on digitisation of sound or video find out how to contribute.
In the meantime, please check our Digitisation Resources for advice, or post a question on our Ask a Question forum
Except where otherwise noted, the Make it Digital Guides are licensed under a
Creative Commons Attribution-Noncommercial-Share Alike 3.0 New Zealand License
.
Comments