Last updated: 22 May 2006
Published in:
Digitising analogue media
Tags:
digitisation |
file formats |
file naming |
scanning |
workflow
The 35mm slide has long played a key role for delivering images in the teaching and learning environment. However, recent improvements in capture and delivery technology have seen digital images virtually replace analogue media. Many slide collections now have to be converted to a format which can more easily be incorporated into the modern workflow. This document explains the process of converting analogue slides into digital images and preparing them for delivery.
In a few years of collecting, it is easy to build up an impressive collection of 35mm slides for one’s own use. However, technology moves on: some camera manufacturers have ceased development of film cameras and in 2004 Kodak stopped making slide projectors. There therefore comes a time when we must think about converting our important personal slide collection into a form that we can more easily use on the computer and view in more technology centred teaching environments.
This is not a hard process but like many computer-based tasks there are a few snags waiting to catch the unwary and it can sometimes appear to be a daunting process if you have not done it before.
Like many other aspects of digital imaging, the first stage is always to research and consider how you want to use the images you are going to make. This will include considering whether any copyright clearance is needed before you digitise your slides. For further information on image copyright see JISC Digital Media’s information on Copyright and Digital Images.
The amount of ‘information’ held within a digital image is dependent upon the image-file’s size (or resolution). The higher the resolution at which we scan the slides, the more information we will collect and the higher ‘quality’ image-file we will have made. However, the bigger the file, the more storage we need to archive it and the harder (and slower) it is to work with the file. To create files any larger than we need would be a waste of time and money but to make image-files too small would mean that they are unable to fulfil our expectations of them. This means that in practice we tend to decide on what is the biggest image size that we are likely to need and then scan the slide to meet this need but no more.
So first of all we must decide to what use the digital images will be put. When we know how we are going to use the images we can work out how big the images will need to be for this use and can then create them to the correct size.
The following chart gives a selection of ‘rule of thumb’ sizes for possible uses from thumbnail to A4 print.
| Use | Size in pixels | File type | Compression (% of original file-size) | Approx file size |
|---|---|---|---|---|
| Monitor - thumbnail | 125 × 125 | JPEG | JPEG - 10% | 45 KB |
| Monitor - full screen | 800 × 600 | JPEG | JPEG - 10% | 140 KB |
| Monitor - ×4 screen size | 1600 × 1200 | JPEG | JPEG - 10% | 0.4 MB |
| A5 inkjet print @ 150 ppi (pixels per inch) |
1275 × 900 | JPEG TIFF |
JPEG - 35% TIFF - None |
JPEG - 1.1 MB TIFF - 3.3 MB |
| A4 inkjet print @ 150 ppi |
1800 × 1275 | JPEG TIFF |
JPEG - 35% TIFF - None |
JPEG - 2.3 MB TIFF - 6.6 MB |
| A4 commercial print @ 300 ppi |
3600 × 2400 | TIFF | None | 24.7 MB |
As photographic slides (35mm anyway) are so small, getting a useful amount of information out of them can be quite difficult and expensive. Being entirely pragmatic, if you have only a small collection of slides to scan then getting them done by a scanning bureau might well be the easiest and cheapest solution. However, if you have a large collection and foresee a steady need for the equipment then it will make sense to do the digitisation yourself. As you compare the available equipment it will become obvious that the quality of the equipment and its cost are closely linked.
There is a wide choice of scanners that can digitise your slides, from expensive and professional drum scanners to cheap consumer flat bed units. Here we provide an overview - more detailed information is available in JISC Digital Media’s advice document on Scanners.
When you choose your scanner, you should ensure that it is able to provide a high enough resolution to create the size of image that you intend to use. If these images are intended just for monitor use then this might be no more than about 1200 spi. However, for creating print-quality images, a resolution of up to 2800 spi could be useful. The ability of the scanner to hold information in both the shadows and highlights is measured by its density range. For the very best of scanners this might be as high as 3.8 and for our use we should hopefully be looking at devices with a density range of at least 3.2 - 3.4.
Some film scanners offer multiple pass scanning which rescans the same image a number of times to capture more detail and less noise in the darkest areas of the image, this will add considerably to the time needed for each scan. Dust, scratches and surface defects are a common problems and some of the better film scanners offer software solutions, which can identify and remove marks without altering original detail. Better quality scanners also offer the ability to scan and store images with an increased amount of colours (larger colour depth). This extra data is useful within the capture process and may be stored within your master archive file. Not all applications however currently support bit depths greater than 24-bit colour.
Flatbed scanners are designed to scan flat artwork and when fitted with a transparency adapter (or hood) can be used to scan slides. The range in cost of these units is vast and their quality ranges from completely professional to barely adequate for simple Web use.
Film scanners are designed purely for the purpose of scanning transparencies and negatives at high resolutions. Again quality tends to vary with cost but in general they are more expensive than the cheaper flatbeds and capable of giving higher quality results than all but the most professional flatbed scanners. Some scanners accept accessories such as uncut film holders and bulk slide feeders: while these devices will not speed up the scanning process they do free up staff for other less repetitive tasks.
Whatever scanner you use, you will find that working with images always puts a high strain on all the other computer equipment that you are using. Digital images tend to be large in size and will rapidly use up the available space on a hard disk. This will mean that you will require a fairly well specified computer with a fair amount of RAM memory and certainly lots of storage both on and offline to store your images.
The actual operation of the scanner will be largely dependent upon the capture software that controls the scanner.
However a generic capture workflow will normally follow these stages (see also our Generic Image Digitisation Workflow):
Once you have scanned the slide, you can replace it safely back into its storage. The image file should be given a unique name and be stored ready for optimisation.
Scanning your slides is only the first step in converting them into a digital form; next you will have to optimise them for their proposed use.
Sometimes when we scan a slide we know exactly how we are going to use the digital image and can scan specifically for that need. At other times we must create a file that we hope we will be able to apply to a variety of uses into the future. When we do this we call the file the Master Image and we then use this to create our required images at a later date.
This Master Image should be optimised before it is archived, surrogates taken from this may need further optimisation depending on their method of delivery.
It is important to realise that although we can use the Master Image file to make other smaller images it is pointless to make images any bigger than the original. While we can increase the number of pixels in an image this will only add to the picture’s file size, no more detail will be revealed.
The first step in any image processing is to make a backup copy of the data, then any loss or damage can be corrected by reloading the original image from the backup.
Most scanned images benefit from some basic image processing before they are stored in the Master Archive. These stages are quite generic and should be used for every scanned image. Some scanning software allows much of this processing to be undertaken at the time of digitisation. However, it is normally better to undertake it within your Image Processing software where you can take advantage of more control and additional functionality.
This optimised image file should be saved as an RGB TIFF within an appropriate Colour Space (sRGB for Monitor and Adobe RGB 1998 for print). The file should be saved with a unique name that shows it has been optimised. There is an argument for saving an additional version of this archive file in the proprietary file format used by the Image Processing software that you used to create it. This will allow the alterations to the file to be stored separately within the file so that they might be removed or edited at a later date. However, this does lead to a far larger sized image file and increased storage requirements.
Once the ‘Adjusted Master’ has been created it will become the source file for all future surrogate images. It is important that all surrogate or delivery images are made from an uncompressed ‘Master’ file and not from a JPEG compressed file, which can introduce an unacceptable level of image degradation. The size, file type and level of compression of these images will depend totally upon their proposed use.
Normally further optimisation will include the following stages:
It will have become apparent that after scanning your slides you can easily end up with many different versions of the same image file, every one made for a different use. In order to deal with this problem it would be useful to create a standardised naming convention that will always provide a unique name. This will enable you to recognise both what the image is, as well as what is was made for.
Deciding upon a reliable naming convention can be a hard task in itself and certainly some serious thought must be put into it before too many images are captured (and possibly lost!)
There is no established standard for this and you will need to consider a system that is appropriate for your use. A typical system could provide the following:
For further information on file naming see JISC Digital Media’s Choosing a File Name.
| Use of image | Possible file types | Level of compression | Rough compression |
|---|---|---|---|
| Archive image | TIFF | None (don’t use LZW) | 100% |
| PNG-24 | Dependent on image | 50 - 65% | |
| Print image | TIFF | None | 100% |
| JPEG | High quality - Larger file size | 25 - 30% | |
| Full colour image being used by image editing software | PSD (Photoshop) PSP (Paint Shop Pro) |
None - and additional Layers and Channels can lead to large increase in file size | File can expand, dependent upon extra work |
| Full colour photograph for use on Web site | JPEG | Small file size - Lower quality | 5 - 10% |
| Full colour graphic for use on Web site | PNG-8 | 256 indexed colour | 15 - 20% |
| Full high quality image for intranet monitor use | PNG-24 | Dependent on image | 50 - 65% |
| B&W bitmap image for use on Web site | PNG-1 | Dependent on image | 1 - 2% |
Last updated: 22 May 2006
Published in:
Digitising analogue media
Tags:
digitisation |
file formats |
file naming |
scanning |
workflow
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++