We have built a contraption to digitise oversized documents!
Until now we were unable to digitise anything larger than A3, which excluded a lot of old newspapers and most posters.
A few years ago we dabbled with a cheap camera, sofa cushions, string and tape, photographing a couple of issues of Freedom from the 1920s, but the results were rather underwhelming and our time better spent focussing on smaller documents. However, the idea of photographing large documents from above stayed on our minds and recently we revisited the project more systematically.
We built a frame to hold a height adjustable document table (just shy of A1 size), topped by a clear Perspex lid to help levelling those pesky document folds. Along with two LED light panels, an aluminium rod, a lot of assorted clamps, chains, bolts, hinges, cables, some new software and an entry level DSLR camera this has now become a working digitisation station!
At this moment, everything is comfortably readable, although the images could be crisper when viewed at high magnifications. The OCR works very well, achieving well above 95 percent accuracy on most documents.
We still have a long way to go and a steep learning curve ahead as we work to improve the results, aiming to level them up from decent to awesome. Despite our lack of knowledge of cameras, illumination, etc. we are however already achieving satisfying results and hope that with experience, upgrades and tweaks to the setup (e.g. a more flexible and more precisely levelled camera mounting) we can soon improve on these early experiments.
We have jumped into this with much enthusiasm, but this is a true DIY project with a lot of scope for improvement. If anyone has any helpful suggestions, please feel free to get in touch. If you have questions, e.g. regarding our gear, settings, software etc., please get in touch.
We picked a few volumes of Freedom and an awkwardly sized issue of Black Flag as our test subjects. Two of those three vols of Freedom are already online in the Freedom Press Newspaper Archive, but we chose them as they are great documents to experiment with, not least as the font is ridiculously small in many places. We are happy to take suggestions what to focus on next, but will likely continue to work with Freedom for a few months, trying to help closing some of the remaining gaps in the online archive, but are also itching to start working on other documents, e.g. the many issues of Direct Action which could not be digitised with our old equipment.
In the meantime it is likely that we will take apart and reassemble the digitisation rig several times (current version 1.3ish, with new upgrades already planned), and before long we will hopefully get to a point where the balance of shutter speed, f/thingy and those other values is no longer the result of semi-random trial and error.
For now, have a look at what we have done so far and feel free to get in touch if you can help us improve the results!