Luxembourgish language corpus platform
SproochHub holds two kinds of material: audio recordings of spoken Luxembourgish with transcripts, and written Luxembourgish texts. Every recording is transcribed and time-aligned, so that each sentence and word in the transcript is linked to the moment in the audio where it is spoken; researchers can listen to any search hit and study pronunciation, prosody and variation in spoken Luxembourgish. The texts are indexed word by word for concordance search and corpus statistics.
Corpora form a tree: a collection, its sub-collections, and the individual items, each an audio recording with its transcript or a text. Items carry metadata such as source, date, speakers and genre, and each inherits a licence from its collection. Searches, statistics and exports can be scoped to any node of the tree.
Each collection carries a licence stating who may read, listen to, download and publish its material. Much of it was provided by its owners for research use only and is available to ZLS staff and research partners with an account. The composition of the collection is documented for account holders inside the application. Material under open licences will be released through the Luxembourg open data portal.