Fair copies Gottfried Semper HTR dataset

This dataset contains gold standard transcriptions of manuscripts from ca. 1856-1863. The pages contain handwritten text in German Kurrent with additions in Latin and Greek scripts. The majority of the text is in German with occasional additions in Greek, Latin, French, and English. The handwritten text is in ink from a variety of hands who were making fair copies from draft book manuscripts of Gottried Semper (1803-1879), German and Swiss architect and art theorist. Additions in lead and ink are from the hand of Gottfried Semper himself. Identifiable hands include Gottlieb Baumann (1828-1900), Hans Semper (1845-1920), Elisabeth Semper (1836-1872), page numbers in lead by Wolfgang Herrmann (1899-?). These transcriptions were created for later processing and publication as part of the Semper Edition, www.semper-edition.ch. The transcription work was completed between October 2017 and August 2023. The text was transcribed with initially with the Semper HTR model and subsequently with the German_Kurrent_XIX_pylaia model in Transkribus. In both cases the automatic HTR was corrected by an expert human transcriber. The long German “s” (“ß”) has been transcribed as “ss”. Other orthographic specificities of the time of creation were transcribed verbatim. This work was undertaken as part of a project “Gottfried Semper. Style. Critical and commented edition” conducted over the period 2017 - 2028 jointly by the Università della Svizzera Italiana and the ETH Zürich and funded by the Swiss National Science Foundation. The documents transcribed here are preserved within the fonds of the gta Archives at ETH Zürich. Each folder represents a document as it is described in its archival context. In each folder you will find sub-folders containing page-xml and alto-xml transcriptions. The images of the transcribed documents are not included in this dataset but are available over IIIF. Each XML file contains a URL for the corresponding image. In the PAGE files this is found as part of the Metadata element and in ALTO files - as part of the sourceImageInformation element.