Digitization with Transkribus ScanTent
Series A (Tyrolean Provincial Archives) and the time frame up to 1550 served as the basis for selecting the castle inventories. The digitisation was then carried out using a ScanTent from Transkribus. The scanning tent is lightweight and foldable and ‘provides optimal conditions for photographing handwritten and printed documents.’
Für den Scan wurde die Smartphone-App DocScan verwendet, die eigens für das Scannen historischer Dokumente entwickelt wurde. Die Scans werden direkt in die Tranksriptionsplattform Transkribus importiert und stehen so nach der automatischen Layout-Analyse für die Transkription bereit.
Transcription
A detailed transcription manual was drawn up for the transcription of the 130 inventories, based on the one published by Antenhofer (). The work procedures were specifically adapted to the objectives of the Inventaria project. The transcription work is carried out in accordance with historical criteria and is conducted in close consultation with and in collaboration with the project Hohensalzburg Digital.
The working window in Transkribus consists of an image of the original scan, as well as colour-coded structure types that can be assigned to sections of text. This is done partly automatically and partly manually. Errors caused by automatic layout segmentation (such as incorrectly recognised headings) are therefore corrected manually during transcription.
Directly below this is the automatic transcription suggestion generated by the HTR handwriting model. This means you do not need to type out the entire text; you simply need to correct any errors. At the same time, annotations (= underlining) have already been added at this stage.
The metadata for the digitised items is recorded separately in a table. The data recorded includes, for example, the original reference number in the Tyrolean Provincial Archives, a key theme of the inventory (if there are any particular features), dates, castle owners, inventory type, etc.
English 
Deutsch