Starting a digitisation project completely from scratch has its challenges but it also has a great number of advantages, not least being given the opportunity to thoroughly understand the reasons for digitising the collection and being able to research and think through the workflow and methodology in advance to make sure that we get it right first time. So often, digitisation is undertaken piecemeal and without a clear plan, and what results is a collection of disparate images of varying quality which need to be revisited over time. The challenges of the project, namely a limited budget, tight time frame (two years of initial funding), reliance on a volunteer workforce and need to use pre-existing kit could have proved difficult but in fact, as we worked through our approach to this project, these limitations have actually helped to clearly inform decisions about what we can and cannot manage and how we should approach the material, and in their way have largely simplified the process.
Coming from a digitisation background, I was clear about the overall approach I wanted to take, but was unfamiliar with glass plate negatives as a photographic format. The collection we have is vast (around 10,000), and in varying states of repair. Some of the larger plates are in a scary looking state, and it was clear from the onset that even with careful handling there was no way that we could digitise a large number of these without conservation. A large number appeared to have *some* damage, by which I mean that they could probably be removed from envelopes and digitised with care, but the limited time scale of the project, the unknown quantity factor, and the fact that there are so many other plates in good condition to be tackled have meant that our selection of plates for digitisation has been very clear cut. If it’s damaged, we set it aside. At a point in the future we will audit all damaged plates for conservation and prioritise the most pressing cases. If time allows we will return to those which can be handled with care.
We are lucky among the Libraries and Archives community to have many colleagues willing to share expertise and so I leaned heavily on the professional experience of others in my quest to understand how best to tackle the plates. What is interesting is that while there are a number of clear guidelines with regard handling, storage, cleaning and conservation, the advice around digitisation was varied and contradictory. Should we scan or photograph the plates? This was actually non-negotiable as we already had a high-end scanner (Epson Expression 10000 XL with transparency hood) in place, but opinion varied widely. Should we scan our black and white negatives in colour or greyscale? Again, this was something I expected to be clear-cut, but was widely debated, and different approaches were taken by institutions across the board. Should we scan emulsion side up or down? Through acrylic or mylar sheets? In negative or positive? At what resolution and file size? Should we make adjustments in Photoshop or other software? Should we apply preset scanner settings to our scans?
Answers and advice were plentiful – to the point of becoming confusing – and in the end it boiled down to the specific needs of our own project, the limitations of our set-up and the compromises we were prepared to accept – and those we were not. So how and for what reasons have we approached this project?
The aim of this digitisation project is primarily to make safe and available, this wonderful collection of material which due to its extreme fragility and instability is unsuitable for handling but potentially contains a wealth of images of local people, places and events. After years getting off the ground, this really is a once in a lifetime opportunity to get this material recorded and it needs to be done correctly from the off so there is no need to revisit the work. Images need to be available online but will also be required for potential exhibition and promotional print purposes. The scans also need to be large enough to show small details such as cap badges, military insignia, jewellery and handwritten text. As a general rule, images need to be suitable for print at approximately A3 size, with some larger exceptions to be dealt with on an individual basis.
So: plates are removed from envelopes, assessed for damage, cleaned (using cotton wool and distilled water) on the glass side, and rehoused in conservation grade four flap folders by trained volunteers. All handling of the plates is done wearing powder free nitrile (rubber) gloves. This not only increases grip on the glass, but avoids leaving fingerprints, oils and other deposits on the glass or emulsion. It also helps to protect volunteers from the sharp edges of broken plates.
Plates then pass along to be scanned in small batches on one of two dedicated scanners. Plates are positioned, directly on to the scan bed, emulsion side up, inside precut cardboard masks which enable scan marquees to be pre-set and saved. Scans are taken in 24-bit RGB colour at hi-resolution (up to 1200dpi for the smallest plates), to enable a print output of around A3, and the files are saved as TIFs. These files become our archive masters. Duplicate files are created, to which minimal retouching and colour correction is undertaken. Typically this is just straightening, cropping (we do not crop into the image as we like to retain edge detail and annotation on the glass), adjustment of brightness and contrast using levels and resizing for whatever purpose they are required. In the meanwhile, the original envelopes are scanned (as JPG files) so we have a quick reference visual record, and information about the plates, and the scans, is recorded to a database.
This is not the method everyone has taken. One commonly offered piece of advice is to scan plates emulsion side down. The wisdom behind this is that by placing the plates matte (image) side down, you will be able to obtain the sharpest focus and also avoid potential distortion such as Newton Rings caused by scanning through the glass. Our issue with this was that placing the emulsion straight down on the glass increases the risk of damage to both plate and scanner. Working with multiple volunteers, some of whom would work with us only sporadically, we needed to avoid as much risk as we possibly could. Tests showed that there was no discernible issue with image focus when we scanned emulsion side up, despite the varying thickness of some of the plates. Adjusting the focus manually each time meant we were unable to scan multiple plates simultaneously, and also added an extra level of complication to a process which could seem dauntingly complicated to a volunteer who had limited experience or had never operated a scanner. In addition, the difference in results was negligible. Adding a sheet of mylar material between the plate and scan bed, was again a step we deemed unnecessary – another layer to keep clean and add potential distortion. Issues with distortion such as Newton Rings have not generally been a problem for us, and if they do occur, we deal with them on an individual scan by scan basis. A further advantage to scanning the plates this way means that the image does not need to be later flipped in Photoshop. A very small detail, but as we have discovered working with a team with little experience, every stage of the process offers another chance for error to be made. The smaller the number of steps required in the process, the better.
Many projects scan their black and white plates in greyscale. This includes many very renowned projects such as the Britain from Above project and the decision seems largely based on the huge difference in storage required when large numbers of scans are involved. Opinion on this point varied widely among the projects we spoke to and in our wider research. Best practice dictates that archival images should be captured in 48 bit RGB colour (the Epson Expression 10000XL scanner allows for capture of black and white film in RGB) to record maximum information and allow for fine colour adjustment before conversion down to 24 bit RGB for storage as a ‘working file’ but this method is both costly in terms of file-size (storage), and scan-time. Most projects we spoke to felt that a greyscale scan was more than sufficient for their needs. Given the time and budget limitations of the project and also given the expected future use to which the images will be put it was felt that we can manage without a 48 bit RGB archive scan, but by scanning initially in 24 bit RGB, tests showed that we could achieve a scan with a far greater tonal range than by capturing initially in greyscale. While it is more costly in terms of time and storage, our opinion was that the plates should be treated with as much care as we could allow. These plates are not just black and white. There is colour in the damage; in the oxidisation of the emulsion and the discolouration of the glass. There is colour and detail in the original retouching and masking undertaken by the photographer, and there is detail which is lost by reducing these images down to simple black and white scans.
The one most important piece of advice I would offer to any project undertaking similar work, is to take on board all the advice and then think specifically about YOUR project and YOUR material. There is no right or wrong way. I spent a lot of time reading online discussion threads between experienced photographers and technicians arguing for one way or the other but there are no definitive answers. The most important element of putting a project plan and workflow in place was to keep things simple, CONSISTENT, and to be methodical about all elements of the process. We have had to make changes and adaptations along the way of course but these have been tweaks rather than major changes of direction and of course we still regularly find issues and problems with our data – it is the nature of this kind of multi volunteer-led project, but being consistent in our methodologies means that usually, we can spot errors and resolve issues without too much difficulty.
We are lucky that the material is a constant pleasure to work with, and the small scale of the project means that we are able to see the plates through from initial re-housing to uncovering the stories behind the faces. It is a joy to be involved and as the database expands and our hard-drives fill with images week on week I feel very proud to call this my day job.
Abby Matthews, Project Officer, The Past on Glass.