The digital life of a Hebrew manuscript

Authors: Kristin A. Phelps, former Senior Imaging Technician, British Library and Dr Adi Keinan-Schoonbaert, Digital Curator (Polonsky Fellow), British Library.

Picture your favourite book. Is it just the printed words on the pages that make it special? What about your notes and doodles in the margins? And the dog-eared corners marking the important parts? It may be one of the first things you pack if you are going on holiday, or the first item you unpack when you move house. Let’s be honest, your e-reader is great but there’s something about the physical copy that makes it irreplaceable.

The unique value of books is rooted in their dual nature. They are vehicles of information and, at the same time, they are three-dimensional objects. Books and manuscripts have always had a secret life as objects; they interact with the senses of those who touch them and forever show the evidence of human interaction with them. They have been cherished objects people have carried in both life and in death. Libraries not only stand as testaments to the importance of knowledge, but also to the value of safe-keeping printed books and handwritten manuscripts.

In the Library’s Sir John Ritblat Treasures Gallery, the public is invited to see a selection of the Library’s books and collection items. While carefully selected pages and folios are visible in the Gallery, it is impossible to examine the entire book as a single object. Visitors and scholars may be interested in the unique bindings and spines of these treasures, but are not able to hold the object, feel it, and observe its physical properties to get a sense of the whole book. In fact, the visitors have only a small percentage of the whole ‘book experience’ – and for a good reason. These fragile treasures need to be preserved for future generations. This is the seemingly unsolvable problem for objects in our collection; how can a viewer have a total sense of any object in a complete way?

Enter 3D imaging. A novel solution which allows viewers to examine an entire book or manuscript as a whole object. The technology also provides a digital record which aids in the documentation and preservation of the item. Many museums are currently using 3D imaging and modelling for their collection items. These models provide website visitors (and visitors to galleries) a view of objects as a whole, giving a somewhat tactile feel to items which are generally untouchable. The same is true with manuscripts held in libraries. A good example is the significant collection of Hebrew manuscripts held at the British Library.

Generously funded by The Polonsky Foundation, the Library has digitised (in two dimensions) 1,300 manuscripts under the guidance of the Hebrew Manuscripts Digitisation Project (2013-16). Many of these newly digitised manuscripts are not readily available to the public due to their fragile condition. The project employs a digital curator (Polonsky Fellow) whose responsibility is to encourage the consultation of the project’s digital material for research and scholarship, and promote fresh uses for the new digital research items.

What better way to re-examine some of this material than in full 3D? Under the direction of digital curator Adi Keinan-Schoonbaert, a small project was conceived to create 3D models of three of the Hebrew manuscripts: Add MS 4709 (a 15th-century CE Pentateuch), Or 1087 (a 15th-century CE Book of Esther), and Add MS 11831 (a 17th-century CE Scroll of Esther).

9 x 6

9 x 6

FIG 1: A Pentateuch from central Italy, 1486 CE. The name of one of its past owners is inscribed on the binding with gold letters and ornamental design (Salomon da Costa, 1719 CE; Add MS 4709)


11.75 x 7

11.75 x 7

FIG 2: A 15th-century CE Book of Esther (Or 1087)

The method chosen for 3D modelling is called photogrammetry – or Structure-from-Motion (SfM). In simple terms, a three-dimensional structure can be created from a sequence of two-dimensional images. Special software creates the third dimension by following the ‘motion’ of the camera around the object. Unlike laser scanning, this method is affordable and simple even for non-specialists. All that’s needed is a digital camera – even just a smartphone camera.

In the case of the Library’s Hebrew manuscripts, we benefited from the Imaging Studio’s advanced photography and lighting equipment. The 3D modelling process began with taking photographs of each manuscript from different angles. The idea was to ensure that we had photographs of the manuscript covering its entire surface, with sufficient overlap. In order to do that, the manuscript was placed on a turntable and the camera was mounted on a tripod. We rotated the turntable at 10-15 degrees with a photo taken at each position. After completing a 360-degree circle the manuscript was turned to its reverse side and the process was repeated.

Once enough photos had been taken, the images were white balanced and then masked ready for the next stage. The masking process included making a copy of each image, outlining the object that we wanted to model in an image-editing software (such as Adobe Photoshop), then filling the selected object with white, and the background with black.

After uploading all the images and their masks to the software (we used Agisoft PhotoScan), it identifies which part of each image should be 3D modelled. The software then makes use of the overlap between the photos to ‘stitch’ the sets of images together, forming a three-dimensional shape of the manuscript. This is done with the help of manually selecting obvious markers visible in photos from both sets, to help the software recognise shared points.

1.35 x 0.21 meters

1.35 x 0.21 meters

FIG 3: A 17th-century CE Esther scroll in an ivory case (Add MS 11831)

Once the models were complete, they were published to Sketchfab – an online platform that hosts 3D content – and can be seen below. Sketchfab allows you to annotate models in order to give a structured narrative or journey through some of the important features. This can be seen in the model of the Scroll of Esther (Add MS 11831), on which annotations were written by the collection’s lead curator Ilana Tahan. Curatorial annotations add yet another dimension to a digitised manuscript, facilitating a more naturally flowing learning experience for researchers and the public alike. These exciting results allow users to have a more rounded view of the manuscripts and present new opportunities for engagement.



Why does this matter? The ownership section of each of our catalogue records reminds us all who has read it, and that these manuscripts have had impressive histories for hundreds of years. Books and manuscripts have most likely been bound and rebound several times and they have travelled the world in the hands of different owners. These owners from the past, like you, found their special book so important that they carried it with them. And now, these manuscripts have a new chapter in their history as objects – not just physical, but also digital.


Digitising the Christina Broom Glass Plate collection

Author: Gemma Cattell, Marketing Executive, Townsweb Archiving

The delicate glass plate negatives are checked before digitisation

The delicate glass plate negatives are checked before digitisation

Christina Broom, the UK’s first female press photographer initially started taking photographs in 1903 to support her family. But the following year a chance meeting resulted in a partnership with The Household Cavalry Regiment that lasted over 35 years. During this time she became The Regiment’s unofficial photographer, capturing a variety of pivotal moments in the history of The Regiment.

Now a unique archive of over 200 glass plate negatives featuring Christina Broom’s work for the unit are held at The Household Cavalry Museum’s archive in Windsor. They are part of a larger mixed media collection documenting the history of the two most senior regiments in The British Army – The Life Guards and The Blues and Royals.

In order to help preserve these fragile and historically important images, in 2015 The Household Cavalry Museum partnered with TownsWeb Archiving to digitise the Christina Broom collection.

The right time to Digitise

Having already worked with TownsWeb Archiving previously to digitally preserve the Museum’s WW1 and Battle of Waterloo photographic collections, The Regiment’s archivist decided to partner with the company again to digitise the Christina Broom glass plate collection.

Dating back to the early 20th Century, the plates in the collection are Silver Gelatine Dry Plate Negatives and are mostly A6 in size, with some up to A4.

The primary reason for digitising the glass plates was to preserve a digital copy of the unique images for posterity and to avoid any being lost as the physical plates degrade over time.

Considerations when digitising Glass Plates

Surface dirt and dust is carefully removed from the plate using a handheld bulb duster

Surface dirt and dust is carefully removed from the plate using a handheld bulb duster

Before digitisation, the glass plates were cleaned with a handheld bulb duster (air blower) to remove surface dirt and dust, which could have affected the quality and clarity of the final digitised images.

With some media it is possible to infrared clean during scanning – removing any ‘noise’ such as dirt and dust. However because of the emulsion used on the plates, this was not possible with the Christina Broom collection.

Due to the age and fragile nature of glass plates they can be one of the most difficult media to digitise. So throughout the project extreme care was needed, especially when handling broken or damaged plates.

Broken plates were a particular challenge. When digitising any broken plates technicians had to carefully place the fragments as close together as possible – to provide as complete a picture as the plate allowed.

In addition, on some of the older plates emulsion was peeling off. Technicians mitigated this by scanning them matte side down, so that the emulsion could lie flat and distortions were minimised on the digital images.

Scanning Equipment and Specification

  • Epson flatbed scanner (set on ‘Transparency’ mode)
  • Hand held bulb duster
  • Bespoke post – processing software
  • Approx. 200 glass plate negatives in the collection (up to A4)
  • Captured at 1200ppi to master TIFF format
  • Captured in Greyscale at 8 bit pixel depth
  • Average TIFF file size – 45Mb

A flatbed scanner set to Transparency mode was used to digitise the collection, mainly because it allowed for the greatest consistency across all images, in terms of colour profile and picture quality.

The high resolution offered by this type of scanner is also more suitable for capturing glass plates, which typically require a greater resolution than loose photographs or bound volumes (see below – File Requirements).

However, in many cases, when digitising glass plates a DSLR camera and light box capture workflow can be used instead.

Once the Museum’s collection had been captured, any blemishes or imperfections on the final images were fixed using bespoke post – processing software. Alternatively, some scanners have software that allows images to be automatically ‘corrected’ during the digitisation process.

Capturing the truest image

Throughout the digitisation, the Christina Broom glass plate collection was captured in 8 bit greyscale. This partially restored the plates to their former appearance and resulted in a greater clarity (due to the increased contrast) in the final digitised images.

Some archivists prefer to capture glass plate negatives in full colour as it shows a true snapshot of the plate at a particular point in time including how it’s aged, fading and any deterioration.

As mentioned, The Household Cavalry Museum’s main reason for digitising was for preservation purposes. For this reason greyscale proved the most suitable technique, as it resulted in the clearest and most useful images for the Museum.

One of the digitised images captured from the Christina Broom collection

One of the digitised images captured from the Christina Broom Collection © Household Cavalry Museum

Digital file requirements

In terms of file requirements, all of the 200 glass plates in the collection were captured at 1200ppi to master TIFF format. Scanning at a higher resolution meant as much tonal information as possible and all the detail from the glass plates was captured. As standard, The National Archives recommend that PPI should be considerably higher for any photographic material, including glass plate negatives.

The master TIFF files were then converted from negatives to positives in post – processing using a bespoke graphics programme.

The last stage in the digitisation process was to produce smaller, compressed JPEG versions from the master TIFF files. These smaller JPEG files make it easier to browse the digitised archive and can be published on the Museum’s website, if required, at a later date.

An historic collection preserved

The digital reproductions of the Christina Broom glass plate negatives are now held at the Museum’s archive in Windsor, safeguarded for future generations. A backup version is also stored elsewhere, as an extra precaution against potential unforeseen disasters.

If you would like to find out more about The Household Cavalry Museum and archive, check out their website. Or feel free to look at the TownsWeb Archiving website for further advice on digitising glass plates and information on glass plate digitisation.


Chinese Porcelain Photography

Author: Kevin Lovelock
Senior Photographer
Photography & Imaging Department
British Museum

Currently I am working on a rather large project photographing all manner of objects for a catalogue written by a curator in the Asia Department, Jessica Harrison-Hall for the National Museum of China. This is also in conjunction with the recent Memorandum of Understanding signed between the British and Chinese governments earlier this month providing over £4M to UK Museums and Galleries. The objects range from bone china, ceramic, bronze, paintings,hanging and hand scrolls, jade and are all various shapes and sizes.

One particular piece has caused me no end of problems. The curator required a general shot and a detail showing a detail of the dragon motif. It was the dragon motif that was the problem.

As you (photographers) may be aware, most of the work we carry out is essentially a problem solving exercise whereby you are confronted by an object and you photograph it using your equipment to hand and your expertise over many years of shooting items of antiquity. Whilst trying to convey a creative style and also being sympathetic to the piece with a mind to the attention to detail of the man/woman who made it over 600 years ago. Obviously over many years of object photography, you get the feeling for how to shoot various objects. You know what works and what doesn’t. When you’re working to a tight deadline this experience becomes more profound and you almost work on autopilot. There are times, objects come into the studio that kind of set you back and if you’re like me, put that back into the safe and get on with the next piece, whilst it’s still in the back of your mind “How am I going to light that piece?”.

Franks 1a

So please look at the first shot of Franks 1a (above). This piece came to the British Museum in 1876 shortly after being exhibited in Bethnal Green Museum. In the exhibition catalogue it stated that it was a pair but only one came to the BM. It was bequeathed by Sir Augustus Wollaston Franks. It is described on our object database as :

Porcelain bowl with incised and anhua decoration beneath a monochrome tianbai glaze. The sides of this delicately potted bowl flare out from a low foot ring. Six V-shaped indentations are made to the rim, suggesting the petals of a flower. Inside, beneath the yellow-tinged white glaze around the sides, is an anhua design of two prancing dragons chasing each other’s tails and separated by ruyi clouds. The five-clawed dragons are depicted in profile but with their heads twisted around to stare at the viewer rather than the dragon they pursue. Incised in the centre is a four-character Yongle reign mark written in seal-script characters. To achieve this thinness of body, the sides have been pared down before glazing. The porcelain is completely translucent and fingers can be seen through its walls when it is picked up.

It is Ming dynasty and was made between 1403 – 1424 in Jingdezhen, China.

The description seems to be quite straightforward, however what you don’t realise is that the dragon motifs are incredibly feint and they seem to be made with a white slip painted onto the same white/cream bowl. Even when viewed using transmitted light, to distinguish the motif from the background is incredibly subtle. Also when using direct transmitted light there is the problem of flare. I found that by careful positioning of the key light above and using a snoot produced the desired effect (See Franks 1b, below).

Franks 1b

You have to shoot on a dark background otherwise most detail is lost. It’s a similar problem when you’re shooting jade. I’ve found that wherever possible, not to light the object directly. The key light used to shoot this object does not fall onto the object but instead is a pool of light, actually lighting the mirror. The light is reflected into the object avoiding overspill of light otherwise this will cause flare (see Franks 1c, below).

Franks 1c

With careful attention to the depth of field owing to the shape of the object. See final shot (Franks 1d, below).

Franks 1d

Equipment List and Image Specifications

Elinchrom Micro 6000 AS power pack with a S 3000 N flash head.
Sinar P monorail camera with F5.6 180mm sinaron se lens.
Phase One P45 digital back fitted to a LightPhase FlexAdaptor for 5″x4″ camera.
Final shot was taken at f22 creating a 117.3 MB tiff file using Adobe 1998 RGB colour space.